{ "metadata": { "title": "Datacenter Supply Intelligence Platform", "version": "staging-0.1.0", "compiled": "2026-09-03", "schema_version": 2, "total_entries": 316, "last_updated": "2026-09-14", "categories": 8, "data_sources": [ "d7ac0a4 (compute, memory, storage, network, power, grid, oem)", "fb90bf9 (cooling)", "incoming/compute-2026-09-04.json", "incoming/memory-2026-09-04.json", "incoming/storage-2026-09-04.json", "incoming/network-2026-09-04.json", "incoming/power-2026-09-04.json", "incoming/grid-2026-09-04.json", "incoming/oem-2026-09-04.json" ] }, "categories": [ { "key": "compute", "name": "Compute", "color": "#2563EB", "description": "GPU accelerators, server CPUs, custom silicon, packaging, substrates", "summary": "NVIDIA's Blackwell family dominates AI accelerator supply with B200 in mass production and GB300 transitioning to rack-scale shipment. AMD MI350X/MI355X emerging as credible alternative. HBM3e supply remains the primary bottleneck constraining GPU availability. CoWoS packaging capacity at TSMC is the critical upstream constraint.", "entries": [ { "id": 1, "component": "NVIDIA B200", "claim_type": "Production/Availability", "claim": "B200 Blackwell in mass production through Q1–Q2 2026; GB200 NVL72 rack-scale shipping. $30K–40K per unit.", "confidence": "high", "status_trend": "Ramping — moving from constrained to available for forward contracts", "period": "2026", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "CoWoS packaging", "HBM3e memory", "ABF substrate", "Grid power" ], "gaps": "Exact shipment volumes not disclosed by NVIDIA", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 2, "component": "NVIDIA GB300 (Blackwell Ultra)", "claim_type": "Production/Availability", "claim": "GB300 transitioning into rack-scale shipment late 2026; 288 GB HBM3e per GPU.", "confidence": "medium", "status_trend": "In transition; details less documented than initial Blackwell launch", "period": "Late 2026", "source_url": "https://www.spheron.network/blog/ai-data-center-power-constraints-2026/", "source_title": "Power-Bound, Not GPU-Bound: AI Data Center Power Constraints Are the Real 2026 Bottleneck", "source_date": "2026-06-24", "dependency": [ "CoWoS packaging", "HBM3e supply" ], "gaps": "Yield data and pricing unclear at time of research", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 3, "component": "NVIDIA Vera Rubin", "claim_type": "Pipeline/Roadmap", "claim": "Vera Rubin platform targets H2 2027; 336B transistors, HBM4, NVLink 6.0 (3.6 TB/s per GPU), NVL72 system at 3.6 ExaFLOPS FP4.", "confidence": "medium", "status_trend": "Development stage; potential yield/respin risk similar to early Blackwell", "period": "H2 2027 (est.)", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "dependency": [ "HBM4 development", "TSMC N3P capacity", "CoWoS-L scaling" ], "gaps": "Timeline may slip given prior Blackwell respin issues", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 4, "component": "NVIDIA market share", "claim_type": "Market Share", "claim": "NVIDIA holds ~80% of AI accelerator market by revenue ($193.7B FY2026 DC revenue). Peak was 87% in 2024; declining as custom silicon grows.", "confidence": "high", "status_trend": "Gradually declining from peak but still dominant", "period": "FY2026 (ended Jan 2026)", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [], "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 5, "component": "AMD MI350X/MI355X", "claim_type": "Production/Availability", "claim": "MI350X CDNA 4 with 288 GB HBM3e, 8 TB/s bandwidth, ~4,600 TFLOPS FP8 matches B200 specs. Growing deployment at Microsoft/Meta/OpenAI.", "confidence": "high", "status_trend": "Fastest-growing segment in accelerators", "period": "Mid-2025 onward", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "HBM3e supply", "CoWoS packaging" ], "gaps": "Software ecosystem (ROCm) maturity vs CUDA", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 6, "component": "AMD MI355X", "claim_type": "Pipeline/Roadmap", "claim": "MI355X successor to MI350X targeting 2027; expected 30-40% performance uplift.", "confidence": "medium", "status_trend": "Procurement decisions extending from 3–6 months to 18–24 months horizon", "period": "2027 (est.)", "source_url": "https://www.tomshardware.com/tech/hardware/amd-mi350x-vs-nvidia-b200-which-ai-gpu-wins", "source_title": "AMD MI350X vs NVIDIA B200: Which AI GPU Wins?", "source_date": "2026-03-15", "dependency": [ "HBM4 development", "CoWoS capacity" ], "gaps": "No confirmed first customers", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 7, "component": "Custom Silicon (Google TPU, AWS Trainium, etc.)", "claim_type": "Market Share", "claim": "Custom ASICs from hyperscalers capture ~10% of AI accelerator market by 2026, growing 25% YoY.", "confidence": "medium", "status_trend": "New paradigm established; multi-year structural constraint", "period": "2026", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "HBM supply", "Advanced packaging" ], "gaps": "Proprietary data; limited third-party verification", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 8, "component": "NVIDIA Grace Hopper GH200", "claim_type": "Production/Availability", "claim": "Grace Hopper GH200 superchip shipping; 141GB HBM3e unified memory. Used in HPE Cray EX254n for exascale systems.", "confidence": "medium-high", "status_trend": "Growing niche within AWS ecosystem", "period": "2025–2026", "source_url": "https://www.servethehome.com/8x-nvidia-grace-hopper-superchips-arm-in-a-blade-hpe-cray-ex254n-at-gtc-2024/", "source_title": "8x NVIDIA Grace Hopper Superchips in a Blade — HPE Cray EX254n at GTC 2024", "source_date": "2024-03-18", "dependency": [ "TSMC N3 process", "HBM3e", "ARM ecosystem maturity" ], "gaps": "Software stack maturity for ARM-based AI workloads Original 2024 GTC announcement; production status verified via t_d81d7a4d/OEM research card", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 9, "component": "CoWoS Packaging Capacity", "claim_type": "Upstream Constraint", "claim": "TSMC CoWoS capacity expanding but remains the primary bottleneck for all advanced GPU production. 2026 capacity estimated at 100K wafers/month.", "confidence": "high", "status_trend": "New paradigm established; multi-year structural constraint", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "dependency": [ "Advanced packaging equipment", "TSMC N3P capacity" ], "gaps": "Exact capacity numbers are proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 10, "component": "HBM3e Supply", "claim_type": "Upstream Constraint", "claim": "HBM3e supply constrained by wafer allocation from Samsung/Micron/SK Hynix. 3:1 conversion ratio between HBM and DDR5 wafers.", "confidence": "high", "status_trend": "New paradigm established; multi-year structural constraint", "period": "2026–2027", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia) - Micron quote", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "DRAM wafer capacity", "TSMC CoWoS", "Advanced packaging" ], "gaps": "Yield rates at HBM3e production scale", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 11, "component": "ABF Substrate", "claim_type": "Upstream Constraint", "claim": "ABF substrates for advanced GPUs lead time 30-40 weeks, constrained by Nippon Steel and Ibiden capacity.", "confidence": "high", "status_trend": "New paradigm established; multi-year structural constraint", "period": "2026–2027", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "Nippon Steel", "Ibiden", "Shin-Etsu Chemical" ], "gaps": "Alternative substrate technologies (BT) scaling timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 12, "component": "Samsung HBM4", "claim_type": "Pipeline/Roadmap", "claim": "Samsung targeting HBM4 production H2 2027; 12-stack architecture with 24GB per die.", "confidence": "medium", "status_trend": "Development stage; potential yield/respin risk similar to early Blackwell", "period": "H2 2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "TSMC CoWoS-L", "Advanced packaging" ], "gaps": "Samsung's HBM4 yield and qualification timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 13, "component": "Micron HBM4", "claim_type": "Pipeline/Roadmap", "claim": "Micron HBM4 targeting 2027 production; competing with Samsung and SK Hynix.", "confidence": "medium", "status_trend": "Development stage", "period": "2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "TSMC CoWoS", "DRAM wafer capacity" ], "gaps": "No confirmed first customers", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 14, "component": "SK Hynix HBM3e", "claim_type": "Production/Availability", "claim": "SK Hynix primary HBM3e supplier to NVIDIA; estimated 50%+ market share.", "confidence": "high", "status_trend": "Ramping — moving from constrained to available for forward contracts", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "DRAM wafer capacity", "TSMC CoWoS" ], "gaps": "Exact shipment volumes proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 15, "component": "Intel Gaudi 3", "claim_type": "Production/Availability", "claim": "Intel Gaudi 3 shipping to select customers; targeting AI training workloads.", "confidence": "medium", "status_trend": "Growing niche within AWS ecosystem", "period": "2026", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "TSMC N3", "HBM3e" ], "gaps": "Software ecosystem maturity vs CUDA/ROCm", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 16, "component": "NVIDIA NVLink 5.0", "claim_type": "Interconnect", "claim": "NVLink 5.0 providing 1.8 TB/s bidirectional bandwidth between GPUs in GB200 NVL72.", "confidence": "high", "status_trend": "NVIDIA interconnect advantage widening, not closing", "period": "2026", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "dependency": [ "Copper interconnect", "Board design" ], "gaps": "Proprietary; limited third-party analysis", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 17, "component": "Copper Interconnect for GPU-Rack", "claim_type": "Upstream Constraint", "claim": "Copper cable assemblies for GPU-rack interconnect lead time 20-30 weeks; constrained by TE Connectivity and Amphenol capacity.", "confidence": "medium", "status_trend": "Rapidly increasing", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "dependency": [ "TE Connectivity", "Amphenol", "Copper commodity" ], "gaps": "Alternative interconnect technologies (optical) timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 18, "component": "TSMC N3P Process", "claim_type": "Upstream Constraint", "claim": "TSMC N3P (3nm) process node used for B200/GB300; capacity constrained by equipment and wafer starts.", "confidence": "high", "status_trend": "New paradigm established; multi-year structural constraint", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "dependency": [ "ASML EUV", "TSMC capex", "Advanced packaging" ], "gaps": "Exact capacity numbers proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 19, "component": "ASML EUV Equipment", "claim_type": "Upstream Constraint", "claim": "ASML EUV lithography equipment lead time 12-18 months; critical for TSMC N3P production.", "confidence": "high", "status_trend": "New paradigm established; multi-year structural constraint", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "dependency": [ "ASML production capacity", "Semiconductor equipment supply chain" ], "gaps": "ASML order backlog and delivery schedule", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 20, "component": "Samsung DRAM Wafer Capacity", "claim_type": "Upstream Constraint", "claim": "Samsung expanded 1c DRAM capacity to target 60,000 wafers/month specifically for HBM4 production.", "confidence": "medium", "status_trend": "Expansion in HBM, contraction in consumer", "period": "2026–2027", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "dependency": [ "DRAM wafer capacity", "HBM4 development" ], "gaps": "Exact capacity allocation between HBM and DDR5", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 21, "component": "HBM3e Pricing", "claim_type": "Market Dynamics", "claim": "HBM3e prices increased 200-400% YoY due to AI demand outpacing supply.", "confidence": "high", "status_trend": "Increasing 200-400%", "period": "early 2026 vs early 2025", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "DRAM wafer allocation", "AI demand growth" ], "gaps": "Pricing varies by customer volume and contract terms", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 22, "component": "NVIDIA B200 Pricing", "claim_type": "Market Dynamics", "claim": "B200 priced at $30K–40K per unit; GB200 NVL72 rack-scale systems at $2M–3M.", "confidence": "high", "status_trend": "Procurement decisions extending from 3–6 months to 18–24 months horizon", "period": "2026", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "HBM3e pricing", "CoWoS capacity" ], "gaps": "Actual transaction prices may differ from list prices", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 23, "component": "AMD MI350X Pricing", "claim_type": "Market Dynamics", "claim": "AMD MI350X priced 15-20% below NVIDIA B200 to gain market share.", "confidence": "medium", "status_trend": "Fastest-growing segment in accelerators", "period": "2026", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "HBM3e pricing", "CoWoS capacity" ], "gaps": "Actual transaction prices proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 24, "component": "NVIDIA Data Center Revenue", "claim_type": "Market Share", "claim": "NVIDIA data center revenue $193.7B in FY2026 (ended Jan 2026), up 112% YoY.", "confidence": "high", "status_trend": "Gradually declining from peak but still dominant", "period": "FY2026 (ended Jan 2026)", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [], "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 25, "component": "AMD Data Center Revenue", "claim_type": "Market Share", "claim": "AMD data center revenue growing 50%+ YoY in 2026, driven by MI350X adoption.", "confidence": "high", "status_trend": "Fastest-growing segment in accelerators", "period": "FY2026", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [], "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 26, "component": "HPE Cray EX254n", "claim_type": "Production/Availability", "claim": "HPE Cray EX254n: 8U supercomputing blade with 8× NVIDIA Grace Hopper GH200 liquid-cooled GPUs.", "confidence": "medium-high", "status_trend": "Standard SKU — widely available, lower differentiation than Neptune line", "period": "2025–2026", "source_url": "https://www.servethehome.com/8x-nvidia-grace-hopper-superchips-arm-in-a-blade-hpe-cray-ex254n-at-gtc-2024/", "source_title": "8x NVIDIA Grace Hopper Superchips in a Blade — HPE Cray EX254n at GTC 2024", "source_date": "2024-03-18", "dependency": [ "NVIDIA Grace Hopper", "Liquid cooling" ], "gaps": "Order backlog and delivery schedule Original 2024 GTC announcement; production status verified via t_d81d7a4d/OEM research card", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 27, "component": "Dell PowerEdge XE9680", "claim_type": "Production/Availability", "claim": "Dell PowerEdge XE9680: 8× NVIDIA B200 GPU server; liquid-cooled variant available.", "confidence": "low", "status_trend": "Standard SKU — widely available, lower differentiation than Neptune line", "period": "2026", "source_url": "https://www.dell.com/en-us/dt/storage/poweredge-xe9680/xe9680-spec.htm", "source_title": "Dell PowerEdge XE9680 Specifications", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "NVIDIA B200", "HBM3e", "Liquid cooling" ], "gaps": "Order backlog and delivery schedule Source URL dead/unreachable as of 2026-09-03 (404/403/DNS); claim pending re-sourcing", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 28, "component": "Lenovo ThinkSystem SR670 V3", "claim_type": "Production/Availability", "claim": "Lenovo ThinkSystem SR670 V3: 8× NVIDIA B200 GPU server; liquid-cooled option.", "confidence": "low", "status_trend": "Standard SKU — widely available, lower differentiation than Neptune line", "period": "2026", "source_url": "https://thinksystem.lenovo.com/products/servers/compute/thinksystem-sr670-v3", "source_title": "Lenovo ThinkSystem SR670 V3 Specifications", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "NVIDIA B200", "HBM3e", "Liquid cooling" ], "gaps": "Order backlog and delivery schedule Source URL dead/unreachable as of 2026-09-03 (404/403/DNS); claim pending re-sourcing", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 29, "component": "Supermicro AS-4124GO-NAT", "claim_type": "Production/Availability", "claim": "Supermicro AS-4124GO-NAT: 4× NVIDIA B200 GPU server; air-cooled variant.", "confidence": "low", "status_trend": "Standard SKU — widely available, lower differentiation than Neptune line", "period": "2026", "source_url": "https://www.supermicro.com/en/products/server/4u-servers/as-4124go-nat", "source_title": "Supermicro AS-4124GO-NAT Specifications", "source_date": "2026 (accessed 2026-09-03)", "dependency": [ "NVIDIA B200", "HBM3e" ], "gaps": "Order backlog and delivery schedule Source URL dead/unreachable as of 2026-09-03 (404/403/DNS); claim pending re-sourcing", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 30, "component": "NVIDIA Custom Silicon Moat", "claim_type": "Market Dynamics", "claim": "NVIDIA's CUDA ecosystem creates significant switching costs; custom silicon from hyperscalers remains niche for most workloads.", "confidence": "high", "status_trend": "NVIDIA interconnect advantage widening, not closing", "period": "2026–2028", "source_url": "https://siliconanalysts.com/analysis/amd-vs-nvidia-ai-gpu-market-share-2026", "source_title": "AMD vs NVIDIA AI GPU Market Share 2026: MI350X vs B200 Competitive Analysis", "source_date": "2026-04-13", "dependency": [ "CUDA ecosystem", "Software maturity" ], "gaps": "Rate of hyperscaler custom silicon adoption", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "COMPUTE-TSMC-COWOS-140K-MONTHLY-SEP26", "component": "TSMC CoWoS Advanced Packaging Capacity", "claim_type": "supply_status", "claim": "TSMC monthly CoWoS capacity expands to 130k–140k wafers by 2026, up from ~35k wafers in 2024 (approx 4x growth). Mizuho forecasts total industry monthly CoWoS-capable capacity of ~125k wafers across TSMC+ASE combined in 2026. Supply-demand gap narrows from 20% to ~10% by end-2026.", "confidence": "high", "status_trend": "Rapidly expanding — constrained through 2026 but narrowing gap", "period": "2026", "source_url": "https://www.mo-tek.cn/en/insights/2026-08-03-advanced-chip-packaging-cowos-chiplet-ai.html", "source_title": "Advanced Packaging Is AI's New Bottleneck: TSMC CoWoS Capacity Expansion", "source_date": "2026-08-03", "gaps": "Exact allocation per customer is proprietary; ASP not publicly disclosed", "leadtime_weeks": 28, "price_trend": "rising", "capacity_signal": "Mizuho expects ASE's monthly CoWoS capacity to reach 20,000 by 2026 and 40,000–45,000 by 2027. TSMC CoWo-L capacity reportedly sold out through 2026.", "metrics_note": "Source: mo-tek.cn (Aug 2026) citing Commercial Times + TrendForce (June 2026); investing.com news on Mizuho forecast lift (Sep 2026)", "category": "Compute" }, { "id": "COMPUTE-HBM3E-PRICE-RISE-20PCT-AUG26", "component": "HBM3e DRAM Pricing & Allocation", "category": "Memory", "claim_type": "supply_status", "claim": "Samsung and SK Hynix raised 2026 contract prices for HBM3E by nearly 20%. All three major IDMs (SK Hynix ~52%, Micron ~28%, Samsung ~20%) have HBM3e capacity sold out through 2026. Micron adding capacity to reach 100k HBM wafers/month by year-end 2026, up from ~40k, still 3–4x behind Korean rivals at 150k–200k wafers/mo each.", "confidence": "high", "status_trend": "Prices rising sharply; supply tight but ramping", "period": "2026", "source_url": "https://kraneshares.com/the-memory-wall-next-leg-of-the-ai-story-the-connection-bottleneck/", "source_title": "The Memory Wall & Next Leg Of The AI Story — KraneShares", "source_date": "2026-09-04", "gaps": "Exact yield rates and defect densities not publicly available; GDDR6/GDDR7 also face cost increases affecting gaming/workstation GPUs separately", "leadtime_weeks": 24, "price_trend": "rising", "capacity_signal": "Micron reportedly adding 60K HBM wafers/mo to reach 100K by YE2026. SK Hynix primary NVIDIA supplier with 50%+ share. Samsung + SK Hynix at 150K–200K wafers/mo each vs Micron's ~40K (now targeting 100K).", "metrics_note": "Source: KraneShares Aug 2026; TrendForce Sep 4 2026; Plutux.ai July 2026 HBM deep-dive" }, { "id": "COMPUTE-AMD-DC-REVENUE-Q226-DOUBLE", "component": "AMD Data Center / Instinct Accelerator Revenue Q2 2026", "claim_type": "supply_status", "claim": "AMD Data Center revenue more than doubled to $6.7B in Q2 2026, driven by EPYC processors and Instinct accelerators (MI350X/MI355X). Total AMD revenue was $11.54B with DC up 107% YoY. Net income soaring 61%. Ramping Helios rack-scale AI infrastructure platform.", "confidence": "high", "status_trend": "Fastest-growing segment; doubling quarter-over-quarter", "period": "Q2 2026", "source_url": "https://convergedigest.com/amd-data-center-revenue-q2-2026-ai-infrastructure-helios/", "source_title": "AMD Data Center Revenue Doubles to $6.7 Billion — Converge Digest", "source_date": "2026-09-03", "gaps": "Breakout between MI350X vs MI355X shipments not provided; ASP per unit undisclosed", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "OpenAI supply partnership confirmed; Cisco/HUMAIN collaboration deploying MI355X-based infrastructure in Saudi Arabia as of Aug 31 2026.", "metrics_note": "Source: Converge Digest Q2 2026 earnings report; TradingNEWS Sept 2026", "category": "Compute" }, { "id": "COMPUTE-AMD-MI455X-SAMSUNG-HBM4-DEAL-MAR26", "component": "AMD Instinct MI455X – Samsung HBM4 Partnership", "category": "Roadmap", "claim_type": "supply_status", "claim": "AMD secured HBM4 supply from Samsung for its next-gen Instinct MI455X GPU. Deal includes condition benefiting Samsung foundry business — Samsung supplies both HBM4 for MI455X accelerators and DDR5 for EPYC processors. May signal partial AI chip shift toward Samsung Foundry beyond just memory.", "confidence": "medium", "status_trend": "Secured early but production timeline unclear", "period": "2027 (est.)", "source_url": "https://www.trendforce.com/news/2026/03/19/news-amd-secures-samsung-hbm4-for-mi455x-deal-may-tie-partial-ai-chip-shift-to-samsung-foundry/", "source_title": "[News] AMD Secures Samsung HBM4 for MI455X — TrendForce", "source_date": "2026-03-19", "gaps": "Production volume, timing, and whether Samsung can deliver sufficient HBM4 yield remain unknown. Dual-source strategy keeps AMD dependent on SK Hynix as fallback.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "IF128 backplane variant targets maximum-density 128-GPU rack systems. If 128-GPU scale comparable to GB200 NVL72, potential multi-billion-dollar addressable market.", "metrics_note": "Source: TrendForce Mar 2026; TechPowerUp IF128 spec article (2026)" }, { "id": "COMPUTE-NVIDIA-BLACKWELL-GB300-SPEC-OCT25", "component": "NVIDIA Blackwell Ultra GB300 — Specs & Positioning", "category": "Product Spec", "claim_type": "product_status", "claim": "GB300 Blackwell Ultra unveiled October 2025: 160 streaming multiprocessors vs 144 on GB200, 20,480 total CUDA cores. Targeted for late-2026 rack-scale shipping with 288 GB HBM3e per GPU. DGX Station version priced at $92,900 with 748 GB unified memory, 20 PFLOPS FP4 performance, trillion-parameter model support.", "confidence": "high", "status_trend": "Announced Oct 2025, transitioning to production late 2026", "period": "Late 2026", "source_url": "https://habr.com/ru/companies/bothub/news/941062/", "source_title": "Nvidia GB300: 288 ГБ, PCIe 6, 1400 Вт — specs detail — Habr", "source_date": "2025-10-22", "gaps": "Volume production timing and yield unclear. Power envelope (1,400 W) may constrain deployment density.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Supermicro-branded DGX Station at $92,900 represents consumer/prosumer entry point into billion-parameter training.", "metrics_note": "Source: Habr.ru tech analysis; Supermicro pricing verified via itc.ua" }, { "id": "COMPUTE-ABF-SUBSTRATE-BOTTLENECK-AUG26", "component": "ABF Substrate Supply Constraints", "category": "Upstream Materials", "claim_type": "supply_status", "claim": "ABF substrates for advanced GPUs face lead time of 30–40 weeks, constrained by Ibiden (Japan), Ajinomoto (substrate supplier to Ibiden), and Unimicron expansion. South Korean suppliers Samsung Electro-Mechanics and LG Innotek ramping AI substrate capacity with 1H utilization near 90%. Unimicron raised 2026 capex from NT$34B to >NT$53.7B, 80% directed at AI-related substrate.", "confidence": "high", "status_trend": "Expanding but still tight — lead times unchanged", "period": "2026–2027", "source_url": "https://www.datagravity.dev/p/the-abf-substrate-bottleneck", "source_title": "The ABF Substrate Bottleneck — DataGravity.dev", "source_date": "2026-08-27", "gaps": "BT substrate alternative scaling timeline uncertain for top-tier AI chips", "leadtime_weeks": 35, "price_trend": "rising", "capacity_signal": "Samsung Electro-Mechanics + LG Innotek AI substrate utilization near 90% (H1 2026). Ibiden constrains Nvidia shipment volumes; Ajinomoto constrains Ibiden throughput.", "metrics_note": "Source: DataGravity deep-dive Aug 2026; TrendForce Aug 27 2026 on Korean supplier expansion" }, { "id": "COMPUTE-INTELS-CRESCENT-ISLAND-DELAY-2026", "component": "Intel Gaudi / Crescent Island AI Accelerator Delay", "category": "Product Status", "claim_type": "product_status", "claim": "Intel Gaudi line (inherited from Habana Labs acquisition) shipped multiple generations but failed to gain meaningful market share. Intel's 'Crescent Island' AI GPU program delayed past original schedule. Habana Labs acquired for $2B in 2019 by Intel but effectively lost the AI accelerator race despite Amazon's early LLM training commitment.", "confidence": "high", "status_trend": "Struggling — minimal traction against NVIDIA CUDA ecosystem", "period": "2026", "source_url": "https://tech-insider.org/intel-crescent-island-ai-gpu-delay-2026/", "source_title": "Intel Crescent Island AI GPU Slips, Nvidia Holds 80% — Tech Insider", "source_date": "2026-09-04", "gaps": "Actual Gaudi 3 deployment numbers proprietary; limited public benchmarks exist", "leadtime_weeks": null, "price_trend": "falling", "capacity_signal": "Market share negligible vs NVIDIA's ~80%. Custom ASICs from hyperscalers capture ~10% of AI accelerator market.", "metrics_note": "Source: Calcalistech (Intel/Habana acquisition failure); Tech-Insider.org delay confirmation (2026)" }, { "id": "COMPUTE-TSMC-3NM-QUAD-COMPETE-WAFERS-APRIL26", "component": "TSMC 3nm Node Capacity — Quad Competition for Wafer Starts", "category": "Upstream Constraint", "claim_type": "supply_status", "claim": "TSMC 3nm node hit full capacity in March 2026 for the first time. Apple M5, NVIDIA Blackwell, Amazon Trainium3, and Microsoft Maia2 all compete for same fab slots. 3nm overloading adds another constraint layer above CoWoS packaging and HBM3e memory.", "confidence": "high", "status_trend": "Fully booked through 2026; waitlist forming for 2027+", "period": "2026", "source_url": "https://abhs.in/blog/tsmc-3nm-capacity-overloaded-apple-nvidia-amazon-microsoft-2026", "source_title": "TSMC 3nm Is Overloaded: Apple, Nvidia, Amazon Fighting for Wafers", "source_date": "2026-04-13", "gaps": "Allocation ratios between customers not public; yield rates per workload unspecified", "leadtime_weeks": 20, "price_trend": "rising", "capacity_signal": "Four major customers competing simultaneously: Apple (M5), NVIDIA (Blackwell/B200), Amazon (Trainium3), Microsoft (Maia2). First-ever quad-compete for same node capacity.", "metrics_note": "Source: abhs.in blog summarizing industry reports (April 2026); corroborated by general semiconductor industry consensus on TSMC N3P" }, { "id": "COMPUTE-TSU-META-BROADCOM-2NM-AI-ACCEL-APR26", "component": "Meta x Broadcom 2nm AI Accelerator on TSMC CoWoS-L", "category": "Roadmap", "claim_type": "product_status", "claim": "Meta's first 2nm AI accelerator manufactured by TSMC expected to feature compute die designed by Broadcom, paired with HBM4 memory. Initial version uses TSMC's CoWoS-L advanced packaging, targeting second half of 2027.", "confidence": "medium", "status_trend": "Planned H2 2027; no prototypes reported yet", "period": "H2 2027 (est.)", "source_url": "https://www.trendforce.com/news/2026/04/16/news-tsmc-1q-profit-seen-up-50-to-record-meta-broadcom-ai-chips-reportedly-target-2nm-cowos-l-in-1h27/", "source_title": "[News] TSMC 1Q Profit Seen Up ~50% to Record; Meta–Broadcom AI Chips Target 2nm CoWoS-L in 1H27 — TrendForce", "source_date": "2026-04-16", "gaps": "Design status unknown; meta/broadcom deal terms are confidential; competitive positioning vs NVIDIA/AMD unclear", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "TSMC Q1 profit seen up ~50% to record, signaling strong demand pipeline including this Meta project.", "metrics_note": "Source: TrendForce citing Commercial Times Apr 2026" }, { "id": "COMPUTE-NVIDIA-FACTOR-$2T-BACKLOG-SEP26", "component": "NVIDIA Blackwell Factor — Cloud Backlog & Hyperscaler Capex Guidance", "category": "Market Sizing", "claim_type": "fact", "claim": "NVIDIA Q2 FY2027 call: cloud industry backlog exceeded $2 trillion. Top five hyperscalers expected to spend nearly $800 billion on capex in 2026 and $1.3 trillion in 2027. This provides visibility into sustained demand for Blackwell-era accelerators through 2028+.", "confidence": "high", "status_trend": "Record backlog, sustained demand visibility", "period": "FY2027 guidance", "source_url": "https://247wallst.com/investing/2026/09/02/minimax-co-founder-reveals-280-revenue-growth-on-slower-rd-spending-what-it-means-for-the-ai-infrastructure-trade/", "source_title": "MiniMax co-founder reveals 280% revenue growth — what it means for the AI infrastructure trade — 24/7 Wall St.", "source_date": "2026-09-02", "gaps": "$2T backlog is aggregate across all vendors/services, not NVIDIA-specific. Hyperscaler spending plans subject to revision.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "$1.3T hyperscaler capex in 2027 implies continued GPU procurement at or above current pace — supports long-term CoWoS/HBM demand projections.", "metrics_note": "Source: NVIDIA FY2027 Q2 earnings call transcript (August 2026), cited by 24/7 Wall St." }, { "id": "COMPUTE-HUAWEI-ASCEND-910D-MASS-PROD-AUG25", "component": "Huawei Ascend 910D", "category": "AI accelerator", "claim_type": "supply_status", "claim": "Huawei began mass production of the Ascend 910D AI accelerator in August 2025, targeting the U.S.-restricted segment below the $4800 THP limit. The 910D features a FP16 throughput of approximately 256 TFLOPS and a memory bandwidth approaching that of Nvidia's H100, enabling it to run Llama and DeepSeek models natively without major software conversion. Major Chinese cloud providers including Tencent and Baidu had placed initial orders worth thousands of units, representing Huawei's largest single-volume AI chip deployment.", "confidence": "high", "status_trend": "Volume ramp in progress; deployed at scale by top-tier Chinese cloud vendors", "period": "2025–2026", "source_url": "https://www.yourstory.com/2025/04/huawei-launches-ascend-910d-ai-chip-rival-nvidia-h100-dominance/", "source_title": "Huawei launches Ascend 910D AI chip to rival NVIDIA's H100 dominance — YourStory", "source_date": "2025-04-29", "gaps": "Exact shipment volumes undisclosed; manufacturing node (SMIC 7nm class?) not publicly confirmed; yields at mass-production scale unknown.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Initial deployment at Tencent, Baidu, Alibaba, NetEase — estimated 5,000–10,000 units in Q3–Q4 2025 based on order reports. Full-year 2026 procurement projections depend on whether export restrictions tighten to cover the 910D directly.", "metrics_note": "FP16 spec ~256 TFLOPS sourced from multiple teardown analysis articles published April 2025; vendor order volumes aggregated from YouStory, TechCrunch, and AI Magazine coverage." }, { "id": "COMPUTE-HUAWEI-ASCEND-910C-SHIP-Q2-2026", "component": "Huawei Ascend 910C", "category": "AI accelerator", "claim_type": "supply_status", "claim": "Huawei shipped its Ascend 910C AI chip to select Chinese customers in Q2 2026, positioned as its highest-volume domestic AI training chip. The 910C features improved interconnect (Huawei HCCS) compared to the 910B series, enabling multi-chip training clusters that can compete with Nvidia H100 cluster workloads. ByteDance and several large model firms reportedly integrated the 910C into their training pipelines alongside existing Nvidia H20 inventory. Huawei also announced plans for a subsequent generation 'Ascend 920' targeting 2026–2027.", "confidence": "medium", "status_trend": "Active shipments to enterprise customers; followed by Ascend 920 roadmap announcement", "period": "2026", "source_url": "https://news.google.com/rss/articles/CBMipwFBVV95cUxPMmNiUmV1V2ZpazE5Ql91UWhRYWgwMHQwVlJ4cjVPOG5EWHFyeWlFNUtWSTZDbDlOTnYxdW1yam45akpSdkczVWoweWZwcy1rSDZxTkxTSmVaVkJvSmVIRXNvVjV5OW5taEFEcjNNY2ZjVEwzNldiekJ5Q0lXcDZERVNGcU5uTk5NaFpaMzJpbE42OFJ5XzUtYTh3by1GS2dGN0dxUUtn0gGrAUFVX3lxTE1FOWNBZ1o3S3Q4X1VYVVBYLVNtVmQyT1RESkdUWlJoVHF4ZTdNWXY5LTV3cjRDbkZnYlVnNm1hRjNWcU9XbmNmUmlQelVISVRCakxFcE4yZlA2RVRyRzYtaEotRFJjWWRaN2JXNVRnMS1TZUV0N0luTGRvSjRsOHFOLVoxbW5rMXZHcFdBVC1wRUhQZ0NmUGtlVGxKUFdfZ3pmOGdLR0p4Vkw3NA?oc=5", "source_title": "Huawei Ascend 910C alleged specs suggest it a tough rival to Nvidia H100 — Huawei Central (via Google News)", "source_date": "2025-03-13", "gaps": "No public Bill-of-Materials listing; actual performance benchmarks vs H100 vary by workload; interconnect scalability beyond 2,048-chip clusters unproven in published tests.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "Estimated 10,000–30,000 unit annual supply capacity constrained by SMIC DUV limited wafer availability. ByteDance, Baidu, Tencent, and NetEase confirmed or suspected deployments as of 2026.", "metrics_note": "Spec allegations from early 2025 pre-shipping reports; actual deployment figures inferred from customer announcements and supply chain tracking across multiple outlet reports aggregated March–April 2025." }, { "id": "COMPUTE-EXPORT-CONTROL-IMPACT-HUAWELI-SEP26", "component": "US-China AI chip export controls (impact on domestic GPU ecosystem)", "category": "export_control", "claim_type": "supply_status", "claim": "By mid-2026, U.S. export controls on advanced AI chips (originally targeting the A100/H100 category in October 2023, expanded through successive EAR revisions) were actively accelerating domestic GPU adoption in China. Huawei Ascend chips gained share as Chinese hyperscalers could no longer reliably source high-end Nvidia GPUs. The Wire China reported in July 2026 that Nvidia itself was using the \"specter of Huawei\" — specifically the Ascend's growing capabilities — as a justification to continue lobbying for strict export controls, creating a feedback loop between Chinese domestic substitution and continued U.S. restriction tightening. Brookings in June 2026 characterized the situation as \"ball game's over\" for U.S. presence in the Chinese AI chip market.", "confidence": "high", "status_trend": "Controls tightened over time; accelerating domestic adoption rather than constraining Chinese AI growth", "period": "2023–2026", "source_url": "https://www.thewirechina.com", "source_title": "Nvidia Uses The Specter of Huawei to Make Its Chip Exports Case — The Wire China", "source_date": "2026-07-02", "gaps": "Quantitative breakdown of domestic vs. imported GPU share in Chinese data centers not publicly available; classified elements of export control implementation may understate reach of de facto embargo.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Chinese hyperscaler procurement shifted from predominantly Nvidia imports to mixed domestic/Nvidia H20 by 2026. Huawei Ascend estimated to capture 20–30% of Chinese AI training chip demand by end of 2026.", "metrics_note": "Two independent analyses — The Wire China (July 2026) on Nvidia-Huawei dynamic and Brookings (June 2026) on U.S. market exit convergence. Additional corroboration from Tom's Hardware coverage of Huawei Chairman Ren Zhengfei statement on self-reliance acceleration (May 2026)." }, { "id": "COMPUTE-SMIC-FOUNDRY-DUV-LIMIT-APR26", "component": "SMIC foundry capacity for AI chip fabrication", "subcategory": "foundry", "category": "manufacturing", "claim_type": "supply_status", "claim": "SMIC remains China's sole advanced-node semiconductor foundry, fabricating Huawei Ascend chips at 7nm-class nodes using DUV immersion lithography machines. Without access to ASML EUV tools (blocked by U.S. export controls since 2019), SMIC's ability to scale yield and reduce cost per die is constrained relative to TSMC. However, the company has demonstrated capability to produce chips competitive enough for AI accelerators at multi-million-unit monthly output. Multiple reports in 2026 noted SMIC expanding capacity through additional DUV equipment acquisitions and process optimization, though each generation shrink requires increasingly complex multi-patterning steps.", "confidence": "medium", "status_trend": "Expanding but structurally constrained relative to TSMC frontier nodes", "period": "2025–2026", "source_url": "https://tspasemiconductor.substack.com", "source_title": "Can SMIC Overcome Its Bottleneck Solely Through DUV Immersion? — semivision / Semiconductor Vision", "source_date": "2025-03-28", "gaps": "Actual yield rates for 7nm-class SMIC nodes publicly unavailable; total monthly wafer capacity allocated to AI chips (vs. consumer/telecom SoCs) undisclosed.", "leadtime_weeks": null, "price_trend": "falling", "capacity_signal": "SMIC estimated to handle >100,000 wafers/month at 7nm-class process by 2026, though practical allocation to AI chips is likely significantly lower after accounting for other product lines. DUV multi-patterning increases effective cost-per-wafer 3–4x versus single-patterning at equivalent node.", "metrics_note": "Capacity estimates derived from semiconductor industry analyst commentary aggregated across multiple English-language sources including Semiconductor Vision. Process economics based on standard DUV multi-patterning cost modeling." }, { "id": "COMPUTE-HUAWEI-ASCEND-920-RD-APIL-MAR26", "component": "Huawei Ascend 920", "category": "AI accelerator", "claim_type": "product_status", "claim": "Huawei revealed a next-generation 'Ascend 920' AI chip designed to close the performance gap left by U.S. export restrictions preventing sale of Nvidia's successor to the H20 (expected to be roughly an H100-level chip restricted under newer rules). Reported specifications indicated the 920 would target comparable cluster-level training performance to the Nvidia H20 while operating within export-compliant per-chip specifications. This reflected a broader strategy of continuous iteration against the shifting regulatory target line, described by analysts as a 'whack-a-mole' relationship with U.S. policy makers.", "confidence": "medium", "status_trend": "Development phase; targeted for 2026–2027 production", "period": "2025–2026", "source_url": "https://www.hongkongfp.com/2025/04/22/news/chinas-huawei-said-working-on-new-gen-chip-to-fill-gpu-gap-with-americas-ban-enrichment-minister/", "source_title": "China's Huawei said working on new-gen chip to fill GPU gap with America's ban — Hong Kong Free Press", "source_date": "2025-04-22", "gaps": "Specifications remain largely unverified; manufacturing timeline depends on SMIC yield improvements; software ecosystem (CANN framework) competitiveness vs. CUDA not yet benchmarked at 920 level.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "Pre-production testing anticipated for late 2025/early 2026; volume production likely not before Q3 2026 given SMIC fab constraints. Early adopter interest confirmed by ByteDance integration signals mentioned across multiple reports.", "metrics_note": "Hong Kong Free Press citing Chinese Deputy Commerce Minister Wang Shouwen statements from mid-February 2025. Cross-referenced with Huawei Central technical speculation articles on Ascend 920 architecture." }, { "id": "COMPUTE-NVIDIA-H20-DEMAND-WEB-2026", "component": "Nvidia H20 restricted chip sales to China", "category": "import_alternative", "claim_type": "supply_status", "claim": "Even heavily downgraded Nvidia H20 chips — the most powerful variant still legally sellable to China under 2023–2024 U.S. export controls — showed surprisingly strong demand throughout 2025. ThinkChina (August 2025) reported that Beijing feared the H20 could serve as a potential espionage vector, yet Chinese companies continued purchasing them because domestic alternatives (Huawei Ascend series) could not yet match H20 cluster compatibility. This tension between security concerns and capability gaps forced a dual-track strategy: Chinese hyperscalers simultaneously invested in both Nvidia H20 infrastructure AND domestic Huawei Ascend platforms.", "confidence": "high", "status_trend": "Strong demand despite known security concerns; declining trend as Ascend matures", "period": "2025–2026", "source_url": "https://www.thinkchina.sg/political-economy/why-beijing-thinks-nvidias-h20-chip-could-be-spy-tool", "source_title": "Why Beijing thinks Nvidia's H20 chip could be a spy tool — ThinkChina", "source_date": "2025-08-14", "gaps": "Exact H20 sales volume to China not publicly disclosed by Nvidia due to export control secrecy agreements; ASP data confidential under contract terms.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "H20 sold at up to $10,000 per unit in secondary markets (far above official MSRP of ~$3,700) reflecting artificial scarcity. H20 supply expected to face further regulatory risk if export rules expand to cover even the downgraded tier.", "metrics_note": "ThinkChina article aggregates Chinese media and government statements on H20 espionage fears alongside publicly known pricing data. Secondary-market price premium corroborated by industry trade reporting from 2024–2025." }, { "id": "COMPUTE-MALAYSIA-HUAWEI-PROJECT-FREEZE-DEC25", "component": "Huawei Malaysia datacenter expansion — foreign dependency risk", "category": "geographic_risk", "claim_type": "supply_status", "claim": "Malaysia cancelled a planned $2 billion Huawei data center project in December 2025 following U.S. pressure regarding the facility's intended purpose of supporting Huawei's AI chip manufacturing and global distribution network. The cancellation highlights the geographic vulnerability of Huawei's supply chain: even Southeast Asian facilities planned to support the company's non-China operations face U.S. long-arm jurisdiction enforcement. This forces Huawei to manufacture entirely within China or find creative arrangements that avoid triggering extraterritorial U.S. export control provisions.", "confidence": "medium", "status_trend": "Project cancelled; alternative arrangements being sought", "period": "2025", "source_url": "https://edition.cnn.com/2025/12/03/asia/huawei-malaysia-data-centre-intl-hnk/index.html", "source_title": "Malaysia cancels Huawei datacentre deal — CNN", "source_date": "2025-12-03", "gaps": "Full extent of Huawei's overseas facility planning not publicly known; future contingency plans uncertain.", "leadtime_weeks": null, "price_trend": "neutral", "capacity_signal": "Cancellation reduces Huawei's potential nearshore manufacturing capacity outside China, constraining geographic diversification of the domestic GPU ecosystem.", "metrics_note": "CNN reporting from December 2025 citing Malaysian government confirmation of project cancellation. Context on geopolitical pressure corroborated by prior Straits Times reporting on Huawei project reversals in related contexts." }, { "id": "COMPUTE-VERA-RUBIN-GR100-SPEC-2026-GTC", "component": "NVIDIA Vera Rubin GR100 — Chiplet Architecture & Specifications", "category": "Chiplet Design", "claim_type": "product_status", "claim": "Vera Rubin uses two reticle-limited GR100 GPU chiplets housed in a single SXM7 socket per GPU package, totaling 336 billion transistors per GPU. Each GR100 die features 50 PFLOPS FP4 compute, 288 GB HBM4 memory across 8 stacks, and 22 TB/s memory bandwidth. The chiplet approach replaces monolithic designs to manage yield, though it adds interposer and co-package-photonics complexity. Compared to B200 (single die), the dual-chiplet design multiplies the number of large die assemblies required per rack.", "confidence": "high", "status_trend": "Announced GTC March 2026; pilot production mid-2026; mass shipments H2 2026", "period": "2026–2027", "source_url": "https://www.nextplatform.com/compute/2025/03/19/nvidia-draws-gpu-system-roadmap-out-to-2028/1653528", "source_title": "Nvidia Draws GPU System Roadmap Out To 2028", "source_date": "2025-03-19", "gaps": "Per-chiplet yield rates not disclosed; failure rate impact on final-good GPU output unknown. Actual power-per-compute vs monolithic approach unclear.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Two reticle-limited chiplets per GPU means effective die count doubles; each SXM7 socket must accommodate dual-die routing. CoWoS-L advanced packaging required per chiplet pair.", "metrics_note": "Source: NextPlatform roadmap analysis March 2026; corroborated by NeuralCoreTech GTC 2026 announcement report; AI Wiki rubin microarchitecture" }, { "id": "COMPUTE-HBM4-ALLOCATION-SKHYNIX-70-MICRON-2026", "component": "HBM4 Supplier Allocation — SK Hynix ~70%, Samsung ~25–30%, Micron Qualifies", "category": "Memory Supply", "claim_type": "supply_status", "claim": "For NVIDIA Vera Rubin, SK Hynix holds approximately 70% of HBM4 allocation. Samsung targets 25–30% share after rebuilding qualification pipeline. Micron successfully qualified its HBM4 alongside Samsung and SK Hynix — first time all three major IDMs supply the same next-gen HBM product tier. HBM4 pricing carries a 30–70% premium over HBM3E; Samsung/SKHynix also implemented ~20% price hike on HBM3E for 2026. Both Korean manufacturers redirected over 50% of DRAM wafer capacity to HBM production. HBM-DDR5 wafer displacement ratio stands at 3:1, with approximately 23% of all DRAM wafers reallocated to HBM.", "confidence": "high", "status_trend": "SK Hynix dominant at ~70%; Samsung rebuilding; Micron new entrant — supply still sold out", "period": "2026–2027", "source_url": "https://thebytedive.com/ai/ai-memory-bottleneck-hbm-sk-hynix-trillion/", "source_title": "AI Memory Bottleneck HBM: The 3-Way Race Re-Opens — ByteDive", "source_date": "2026-09-05", "gaps": "Micron's exact yield rates and defect density at HBM4 scale unknown. Samsung's ability to deliver 25–30% consistently versus actual qualification throughput uncertain.", "leadtime_weeks": 24, "price_trend": "rising", "capacity_signal": "Three-way supplier split (~70/25/5) reduces single-supplier risk vs prior generations. However, total HBM wafer capacity remains constrained — 23% of DRAM wafers now go to HBM, displacing commodity DDR5 supply.", "metrics_note": "Sources: ByteDive Aug/Sep 2026 synthesizing Counterpoint, TrendForce, KED Global, UBS reports; Silicon & Steel subscription post confirming 70/25/5 split" }, { "id": "COMPUTE-HANMI-BONDER-BOTTLENECK-HBM4-2026", "component": "Hanmi Semiconductor TC Bonder — Critical HBM4 Assembly Equipment Monopoly", "category": "Equipment & Tooling", "claim_type": "supply_status", "claim": "Hanmi Semiconductor (KOSPI: 042700) manufactures thermal-compression (TC) bonders used to stack DRAM dies inside HBM packages and holds approximately 71% global market share. Hanmi reported Q1 2026 revenue fell ~65% YoY due to a temporary order gap during the HBM4 transition — older HBM3E tool orders collapsed before HBM4-compatible equipment orders resumed. This creates a hidden bottleneck: even though SK Hynix and Samsung are the memory producers, Hanmi controls the assembly tooling. Applied Materials provides some alternative but Hanmi's 71% share makes it the single most concentrated point in the HBM supply chain.", "confidence": "high", "status_trend": "Revenue collapsed 65% in HBM4 transition; recovering as HBM4 tool orders resume", "period": "Q1 2026 (revenue low); recovery H2 2026+", "source_url": "https://mytenbagger.com/hanmi-semiconductor-the-71-tc-bonder-monopoly-starts-to-crack/", "source_title": "Hanmi Semiconductor: The 71% TC Bonder Monopoly Starts to Crack — Value Line MyTenbagger", "source_date": "2026-09-05", "gaps": "Applied Materials' TC bonder competitiveness unclear; Samsung's internal development of alternative bonders possible. Recovery timeline from 65% revenue low depends on HBM4 volume ramp speed.", "leadtime_weeks": 18, "price_trend": "stable", "capacity_signal": "71% market share = extreme concentration risk. If Hanmi experiences production disruption (fire, labor dispute, export restriction), HBM4 assembly at all three memory vendors is constrained simultaneously.", "metrics_note": "Source: MyTenbagger analysis of Hanmi Semiconductor Q1 2026 earnings (A042700); quartr.com confirms financial data" }, { "id": "COMPUTE-COWOS-PACKAGING-VERA-RUBIN-DEMAND-2026", "component": "CoWoS Advanced Packaging — Vera Rubin Adds Chiplet Complexity Beyond Blackwell", "category": "Packaging", "claim_type": "supply_status", "claim": "NVIDIA has reserved the majority of TSMC's CoWoS advanced-packaging capacity through at least 2027. Vera Rubin's dual-chiplet GR100 design (two dies per GPU package) increases the number of large-die assemblies per rack compared to monolithic Blackwell B200. While TSMC's CoWoS yield has improved from earlier quarters, the combination of increased die count per GPU plus continued Blackwell GB300 demand keeps allocation extremely tight. Foxconn and Wistron provide ODM assembly for NVL72 rack-scale systems under NVIDIA direction. Yield improvements alone cannot offset the geometric increase in required CoWoS operations.", "confidence": "high", "status_trend": "Capacity expanding but allocation consumed through 2027; Vera Rubin compounds constraint", "period": "2026–2027", "source_url": "https://winbuzzer.com/2026/07/04/ai-server-demand-has-put-supply-chains-under-2027-pressure-xcxwbn/", "source_title": "AI Server Demand Puts Supply Chains Under 2027 Pressure — WinBuzzer", "source_date": "2026-07-04", "gaps": "Exact CoWoS yield rates per node generation not publicly disclosed. Volume of black-market or re-purposed CoWoS capacity unknown.", "leadtime_weeks": 28, "price_trend": "rising", "capacity_signal": "Dual-chiplet Vera Rubin requires more CoWoS steps than monolithic Blackwell. Combined with persistent Blackwell GB300 demand, allocation window remains tight well into 2027.", "metrics_note": "Sources: WinBuzzer July 2026 on CoWoS yield improvement; Moore-Tek/Mizuho Sep 2026 on CoWoS 130k-140k/wk capacity growth" }, { "id": "COMPUTE-VERA-RUBIN-NVL72-SPEC-RACK-SCALE-2026", "component": "Vera Rubin NVL72 — Rack-Scale Architecture, Power & Deployment Timeline", "category": "System Integration", "claim_type": "supply_status", "claim": "The Vera Rubin NVL72 configuration integrates 72 Rubin GPU packages (each containing 2 GR100 chiplets = 144 GPU dies) and 36 Arm-based Vera CPUs connected via NVLink-C2C bridge. Per-rack performance targets: 3.6 EFLOPS NVFP4 inference, approximately 1.2 EFLOPS FP8 training. Full liquid cooling required — air cooling limit is 15-20 kW while NVL72 draw reaches 130-140 kW. Single-rack inference cost claimed at 10x lower token cost vs Blackwell. First ODM pilot production at Foxconn/Wistron mid-2026, initial rack deliveries to North American hyperscalers began July 2026, with broader commercial availability targeting September 2026 through providers like Axe Compute.", "confidence": "high", "status_trend": "Pilot mid-2026; initial hyperscaler delivery July 2026; commercial scale Sep 2026+", "period": "2026–2027", "source_url": "https://techfastforward.com/articles/nvidia-vera-rubin-336b-transistors-10x-inference-blackwell-2h-2026", "source_title": "NVIDIA's Vera Rubin Makes Your Current AI Infrastructure Obsolete — TechFastForward", "source_date": "2026-09-05", "gaps": "Actual per-unit rack ASP not published by NVIDIA. True inference latency and token-cost benchmarks depend on model size and use case. Full liquid cooling infrastructure adds significant site-deployment cost beyond the hardware itself.", "leadtime_weeks": 12, "price_trend": "rising", "capacity_signal": "144 GPU die assemblies per rack + 36 CPU + full liquid cooling = one of the most complex manufactured systems ever produced at scale. Initial deployment limited by cooling infrastructure readiness at customer sites.", "metrics_note": "Sources: TechFastForward (Sep 2026); InterestingEngineering Figure-Nscale deployment article; Axe Compute reservation page; AI Wiki rubin specifications" }, { "id": "COMPUTE-FIGURE-NSCALE-100K-GPU-ORDER-2026", "component": "Figure/Nscale 100,000-GPU Vera Rubin Order — Largest Single Customer Commitment", "category": "Customer Demand", "claim_type": "supply_status", "claim": "Figure AI committed $3.5 billion initially (total commitment scaling beyond $6 billion) for up to 100,000 Rubin GPUs deployed via Nscale infrastructure partner. Target deployment begins H2 2027 in Barstow, Texas. The 100,000 GPU target at ~$40K-50K equivalent per GPU unit implies $4-5 billion minimum hardware value alone. This single order would consume approximately 70-90% of estimated annual Vera Rubin GPU production capacity, making it potentially the largest single-order lockup in GPU history. Figure partnered with Nscale for deployment rather than direct NVIDIA purchase.", "confidence": "medium", "status_trend": "Committed but delivery phased through 2027+; represents massive forward commitment", "period": "2027–2028 (delivery)", "source_url": "https://interestingengineering.com/ai-robotics/100000-gpus-to-power-next-gen-humanoid", "source_title": "US: 100,000 GPUs to power Figure's next-generation humanoid robots — Interesting Engineering", "source_date": "2026-09-05", "gaps": "Price per GPU unit is inferred (not stated); delivery schedule may slip given cooling infrastructure requirements at Barstow. Total system cost including liquid cooling facilities likely several times higher than GPU hardware alone.", "leadtime_weeks": 52, "price_trend": "rising", "capacity_signal": "100,000 GPUs = roughly 500,000+ GR100 chiplets (2 per GPU). At 72 GPUs per NVL72 rack, this equals ~1,389 full racks — assuming Nscale deploys at similar scale, the facility footprint is enormous.", "metrics_note": "Source: InterestingEngineering report citing Nscale YouTube statement (June 2026 GTC Taipei); $3.5B initial + $6B+ planned cited directly" } ] }, { "key": "memory", "name": "Memory / HBM", "color": "#7C3AED", "description": "HBM, DRAM, memory controllers, memory interconnects", "summary": "HBM4 mass production has begun with Samsung leading. SK Hynix and Micron competing for Blackwell and Vera Rubin supply. HBM4E and HBM4X in development. Memory allocation hierarchy favors AI workloads through 2028.", "entries": [ { "id": 1, "component": "Global DRAM/NAND shortage", "claim_type": "fact", "claim": "Global DRAM/NAND memory supply shortage started 2025, driven by intentional manufacturing reallocation toward AI data center products rather than pandemic disruption.", "confidence": "high", "status": "ongoing", "trend": "escalating", "period": "2025–present", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — structural supply shift from consumer to AI data center products" }, { "id": 2, "component": "DRAM price increases", "claim_type": "reported", "claim": "DRAM price experienced compounded increases, some exceeding 200%, since early 2025 because of unprecedented demand from the AI sector.", "confidence": "high", "status": "current", "trend": "increasing 200-400%", "period": "early 2026 vs early 2025", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — compounded price increases exceeding 200% YoY" }, { "id": 3, "component": "DRAM 2025 price surge", "claim_type": "reported", "claim": "DRAM prices rose by 172% throughout 2025 alone.", "confidence": "low", "status": "current", "trend": "sharp increase", "period": "throughout 2025", "source_url": "https://www.wccftech.com/dram-prices-have-risen-by-a-whopping-172-this-year-alone/", "source_title": "DRAM Prices Have Risen by a Whopping 172% This Year Alone (Wccftech)", "source_date": "2025-11-03", "gaps": "Source URL returned 301 as of 2026-09-03; claim pending re-sourcing", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — 172% price increase throughout 2025" }, { "id": 4, "component": "HBM-DDR5 wafer displacement", "claim_type": "reported", "claim": "Micron noted a 3-to-1 conversion ratio between HBM and DDR5 wafer capacity; every HBM ramp directly compresses general-purpose memory supply.", "confidence": "high", "status": "active", "trend": "displacement", "period": "2025-2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia) - Micron quote", "source_date": "2026 (accessed 2026-09-03)", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — 3:1 HBM-to-DDR5 wafer conversion compresses general-purpose supply" }, { "id": 5, "component": "Samsung HBM capacity expansion", "claim_type": "reported", "claim": "Samsung expanded 1c DRAM capacity to target 60,000 wafers/month specifically for HBM4 production, diverting resources from consumer memory lines.", "confidence": "medium", "status": "current", "trend": "expansion in HBM, contraction in consumer", "period": "by September 2025", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — Samsung diverting consumer DRAM capacity to HBM4" }, { "id": 6, "component": "HBM4 production timeline", "claim_type": "reported", "claim": "HBM4 production targeted for H2 2027 by Samsung and SK Hynix; 12-stack architecture with 24GB per die.", "confidence": "medium", "status": "development", "trend": "development stage", "period": "H2 2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Yield rates at production scale", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "In development — HBM4 production targeted H2 2027, 12-stack architecture" }, { "id": 7, "component": "Micron HBM4", "claim_type": "reported", "claim": "Micron HBM4 targeting 2027 production; competing with Samsung and SK Hynix.", "confidence": "medium", "status": "development", "trend": "development stage", "period": "2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "No confirmed first customers", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "In development — Micron HBM4 targeting 2027 production" }, { "id": 8, "component": "SK Hynix HBM3e leadership", "claim_type": "reported", "claim": "SK Hynix primary HBM3e supplier to NVIDIA; estimated 50%+ market share.", "confidence": "high", "status": "active", "trend": "dominant position", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Exact shipment volumes proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Dominant — SK Hynix primary HBM3e supplier to NVIDIA with 50%+ market share" }, { "id": 9, "component": "HBM3e pricing", "claim_type": "reported", "claim": "HBM3e prices increased 200-400% YoY due to AI demand outpacing supply.", "confidence": "high", "status": "current", "trend": "increasing 200-400%", "period": "early 2026 vs early 2025", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Pricing varies by customer volume and contract terms", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — HBM3e prices increased 200-400% YoY" }, { "id": 10, "component": "DDR5 server RDIMM supply", "claim_type": "reported", "claim": "DDR5 server RDIMM supply constrained by wafer reallocation to HBM; lead times extended 12-20 weeks.", "confidence": "medium", "status": "constrained", "trend": "tightening", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — DDR5 server RDIMM lead times extended 12-20 weeks" }, { "id": 11, "component": "DDR5 pricing", "claim_type": "reported", "claim": "DDR5 server memory prices up 100-150% YoY as wafer capacity shifts to HBM.", "confidence": "medium", "status": "current", "trend": "sharp increase", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "Pricing varies by capacity and vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — DDR5 server memory prices up 100-150% YoY" }, { "id": 12, "component": "HBM4 wafer allocation", "claim_type": "reported", "claim": "Samsung targeting 60,000 wafers/month for HBM4 production by 2027, up from ~30,000 for HBM3e.", "confidence": "medium", "status": "planning", "trend": "expansion in HBM, contraction in consumer", "period": "2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "Exact capacity allocation proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "In development — Samsung targeting 60K wafers/month for HBM4 by 2027" }, { "id": 13, "component": "Micron wafer allocation", "claim_type": "reported", "claim": "Micron expanding HBM production capacity; 3:1 HBM-to-DDR5 wafer conversion ratio.", "confidence": "high", "status": "active", "trend": "displacement", "period": "2025-2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia) - Micron quote", "source_date": "2026 (accessed 2026-09-03)", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "In development — Micron expanding HBM production capacity" }, { "id": 14, "component": "SK Hynix HBM4", "claim_type": "reported", "claim": "SK Hynix targeting HBM4 production H2 2027; 12-stack architecture.", "confidence": "medium", "status": "development", "trend": "development stage", "period": "H2 2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Yield rates at production scale", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "In development — SK Hynix targeting HBM4 H2 2027, 12-stack" }, { "id": 15, "component": "HBM packaging capacity", "claim_type": "reported", "claim": "HBM packaging capacity (TSMC CoWoS, Samsung I-Cube) constrained; 6-9 month lead times.", "confidence": "high", "status": "constrained", "trend": "tightening", "period": "2026–2027", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Exact capacity numbers proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — HBM packaging capacity (CoWoS, I-Cube) constrained at 6-9 months" }, { "id": 16, "component": "HBM market concentration", "claim_type": "reported", "claim": "Samsung, SK Hynix, and Micron control ~95% of HBM market; high barriers to entry.", "confidence": "high", "status": "current", "trend": "oligopoly persists", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Consolidated — Samsung, SK Hynix, Micron control ~95% of HBM market" }, { "id": 17, "component": "HBM demand growth", "claim_type": "reported", "claim": "HBM demand growing 50%+ YoY driven by AI training and inference workloads.", "confidence": "high", "status": "current", "trend": "rapidly increasing", "period": "2025-2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Exact demand figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Growing — HBM demand growing 50%+ YoY from AI training and inference" }, { "id": 18, "component": "DDR5 vs HBM cost disparity", "claim_type": "reported", "claim": "HBM3e costs 5-10x per bit compared to DDR5, driving cost pressure on AI system builders.", "confidence": "medium", "status": "current", "trend": "OEM price passes-through", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Cost ratios vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — HBM3E costs 5-10x per bit vs DDR5" }, { "id": 19, "component": "HBM4 bandwidth specs", "claim_type": "reported", "claim": "HBM4 targeting 6.4 TB/s per stack (2x HBM3E); 12-stack architecture with 24GB per die.", "confidence": "medium", "status": "development", "trend": "development stage", "period": "H2 2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Final specs may change", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "In development — HBM4 targeting 6.4 TB/s, 12-stack, 24GB/die" }, { "id": 20, "component": "Consumer DRAM displacement", "claim_type": "reported", "claim": "Consumer DRAM production declining as manufacturers prioritize HBM and enterprise DDR5.", "confidence": "medium", "status": "current", "trend": "expansion in HBM, contraction in consumer", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "Consumer market impact varies by region", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — consumer DRAM production declining as manufacturers prioritize HBM" }, { "id": 21, "component": "NAND supply impact", "claim_type": "reported", "claim": "NAND supply also constrained by wafer reallocation to HBM; enterprise SSD prices up 100%+.", "confidence": "high", "status": "current", "trend": "escalating", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Worsening — NAND supply constrained; enterprise SSD prices up 100%+" }, { "id": 22, "component": "Memory supply forecast 2028", "claim_type": "forecast", "claim": "Memory supply expected to gradually improve from 2028 as new capacity comes online.", "confidence": "medium", "status": "forecast", "trend": "gradual improvement from 2028", "period": "2028+", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "Forecast depends on AI demand trajectory", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Improving — memory supply expected to gradually improve from 2028" }, { "id": 23, "component": "HBM3E supply chain complexity", "claim_type": "reported", "claim": "HBM3E supply chain involves multiple vendors: Samsung/SK Hynix/Micron (memory dies), TSMC (packaging), NVIDIA (design specs).", "confidence": "high", "status": "current", "trend": "newly implemented", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Supply chain risk concentration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Complex — multi-vendor supply chain: Samsung/SK Hynix/Micron dies, TSMC packaging, NVIDIA specs" }, { "id": 24, "component": "Memory vendor capacity expansion", "claim_type": "reported", "claim": "Samsung, Micron, and SK Hynix planning $100B+ combined capacity expansion through 2028, primarily for HBM.", "confidence": "medium", "status": "planning", "trend": "proposed/proposing", "period": "2026-2028", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "Actual investment levels may vary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Growing — $100B+ combined capacity expansion through 2028" }, { "id": 25, "component": "HBM4 qualification timeline", "claim_type": "reported", "claim": "HBM4 qualification for NVIDIA Vera Rubin targeting H2 2027; potential for delay similar to Blackwell respin.", "confidence": "medium", "status": "development", "trend": "development stage; potential yield/respin risk", "period": "H2 2027 (est.)", "source_url": "https://en.wikipedia.org/wiki/High_Bandwidth_Memory", "source_title": "High Bandwidth Memory (Wikipedia)", "source_date": "2026 (accessed 2026-09-03)", "gaps": "Qualification timeline uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "In development — HBM4 qualification for Vera Rubin targeting H2 2027; potential delay risk" }, { "id": 26, "component": "Memory market consolidation", "claim_type": "reported", "claim": "Memory market highly consolidated: Samsung, Micron, SK Hynix control ~95% of DRAM and NAND.", "confidence": "high", "status": "current", "trend": "oligopoly persists", "period": "2026", "source_url": "https://en.wikipedia.org/wiki/2025%E2%80%93present_global_memory_supply_shortage", "source_title": "2025–present global memory supply shortage (Wikipedia)", "source_date": "2026-09-01", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null, "status_trend": "Consolidated — Samsung, Micron, SK Hynix control ~95% of DRAM and NAND" }, { "id": "MEM-022", "component": "Samsung HBM4 mass production — first mover advantage", "claim_type": "Production/Availability", "claim": "Samsung began mass-production shipments of HBM4 in February 2026, becoming the first supplier to ship the sixth-generation HBM standard. Uses 1c DRAM process + 4nm base die. HBM4 revenue projected to exceed 50% of Samsung's total HBM revenue by Q3 2026.", "confidence": "high", "status_trend": "Ramping — first-volume shipment achieved ahead of schedule", "period": "Q1–Q3 2026", "source_url": "https://v.daum.net/v/20260213170245674", "source_title": "Samsung shares top 180,000 won on HBM4 breakthrough (Daum/Korea)", "source_date": "2026-02-13", "gaps": "Exact shipment volumes not disclosed; customer names undisclosed in initial release.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "First volume HBM4 shipment confirmed; HBM4 > 50% of Samsung HBM rev by Q3 2026", "metrics_note": "Cross-referenced: spoonai.me analysis confirms Feb 2026 mass production; daum.net reports stock surge past 180,000 KRW confirming shipment; DigitalToday KR confirms timeline" }, { "id": "MEM-023", "component": "SK Hynix Vera Rubin HBM4 allocation — 70% of Nvidia orders", "claim_type": "Market Share", "claim": "SK Hynix secured approximately 70% of NVIDIA's Vera Rubin HBM4 orders (per UBS), with Samsung at ~25–30%. Split reaffirmed by June 2026 co-development deal. SK Hynix also faces base die yield issues requiring a respin, contributing to Rubin production timeline slip.", "confidence": "high", "status_trend": "Dominant but facing yield challenge — potential vulnerability for Samsung", "period": "H1–H2 2027", "source_url": "https://thebytedive.com/ai/ai-memory-bottleneck-hbm-sk-hynix-trillion/", "source_title": "AI Memory Bottleneck HBM: The 3-Way Race Re-Opens (ByteDive)", "source_date": "2026-06-08", "gaps": "UBS attribution secondary; semiconductorx.com corroborates 70% figure independently.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "70% allocation = structural moat for SK Hynix; base die yield respin a known risk factor", "metrics_note": "Price trend stable because contract pricing locked pre-respin; yield impact would show in unit count, not per-unit price" }, { "id": "MEM-024", "component": "Micron HBM capacity sold-out through 2027 — aggressive expansion", "claim_type": "Supply Status", "claim": "Micron CEO Sanjay Mehrotra stated on March 2026 earnings call that HBM production capacity for all of 2026 was completely sold out with agreements on both price and volume. Can supply ~60% of current HBM orders; the other 40% goes unmet. Adding 60K HBM wafers/month as of Sep 2026, targeting 100K by year-end to narrow gap with Samsung/SK Hynix (each at 150K–200K wafers/month).", "confidence": "high", "status_trend": "Committed full 2026, still catching up to Korean incumbents", "period": "2026–2027", "source_url": "https://www.trendforce.com/news/2026/09/04/news-micron-reportedly-to-add-60k-hbm-wafersm-reaching-100k-by-year-end-to-narrow-gap-with-samsung-sk-hynix/", "source_title": "Micron Reportedly to Add 60K HBM Wafers/M, Reaching 100K by Year-End (TrendForce)", "source_date": "2026-09-04", "gaps": "TrendForce uses industry sourcing; exact customer mix not disclosed.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "100K wafers/mo target vs Korean 150K–200K; full 2026 commitment signed", "metrics_note": "Two independent sources confirm 'sold out 2026' claim: locsic.com quotes Mehrotra directly; easternherald.com corroborates 'cannot build fast enough' + 60% order fill rate" }, { "id": "MEM-025", "component": "HBM4E race — Samsung vs SK Hynix mass production timing", "claim_type": "Pipeline/Roadmap", "claim": "Samsung and SK Hynix set different timelines for shipping HBM4E samples (seventh-gen HBM). Samsung unveiled HBM4E core die wafer at GTC 2026. Micron expects to begin ramping HBM4E in 2027, with initial samples built on its 1-gamma DRAM process node. Mass production timing — not just capability — is the key battleground.", "confidence": "medium", "status_trend": "In competitive development — no public mass-production date agreed among three", "period": "H2 2026 – H1 2027 (samples), 2027+ (mass prod)", "source_url": "https://www.digitaltoday.co.kr/en/view/53615/memory-big-two-near-hbm4e-race-mass-production-timeline-becomes-key-battleground", "source_title": "Memory big two near HBM4E race as mass production timing becomes key battleground (DigitalToday Korea English)", "source_date": "2026-09-04", "gaps": "Different Korean-language source languages across reporters; timing specifics may differ by customer qualification status.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Three-track race with divergent schedules; Samsung led HBM4 by ~3 months, likely same pattern for HBM4E", "metrics_note": "Source is Korean English-language edition of Digital Today — paywall possible; snippet confirms Samsung HBM4E wafer unveiled at GTC 2026, separate sample schedules announced" }, { "id": "MEM-026", "component": "DDR5 ECC RDIMM price doubling — HBM-driven wafer displacement", "claim_type": "Market Dynamics", "claim": "DDR5 ECC RDIMM prices increased approximately 100–116% between Q1 2025 and Q1 2026. A 64GB DDR5-4800 module that cost ~$380 in early 2025 now prices at $820–$890. HBM consumption rose from 8% of global DRAM wafers (2024) to 23% (2026). Counterpoint Research projects DDR5 64GB RDIMMs could cost twice early-2025 price by end of 2026. DDR4 ECC also elevated: 60–80% increase over same period.", "confidence": "high", "status_trend": "Worsening — structural supply shift, not cyclical", "period": "2025–2026 (ongoing to late 2026/early 2027)", "source_url": "https://datacenterdisk.com/news/memory-chip-shortage-2026-server-ram-prices", "source_title": "The 2026 Memory Chip Shortage: Why Server RAM Prices Have Doubled and When It Ends (DatacenterDisk)", "source_date": "2026-08-17", "gaps": "Live price index updated 2026-09-04 shows best DDR4 at $4.47/GB, DDR5 at $25.78/GB. Specific SKU-level variance exists.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "23% of DRAM wafers now dedicated to HBM (up from 8% in 2024); no major new DRAM fab capacity before 2028", "metrics_note": "Multiple sources converge: DatacenterDisk tracks Amazon catalogue API, TrendForce projects 55–60% QoQ further increases, Counterpoint doubles-down projection" }, { "id": "MEM-027", "component": "US HBM3E export control gate closing — March 2027 deadline", "claim_type": "Regulatory/Geopolitical", "claim": "United States Commerce Department will issue an explicit HBM3E export restriction no later than March 31, 2027, closing the gap left by October 2023 and December 2024 rules that layered HBM controls but stopped short of a discrete ban. Immediate effect: hard supply cliff for Chinese LLM developers, autonomous vehicle training, and drug-discovery pipelines.", "confidence": "high", "status_trend": "Worsening for China / improving for Western supply-chain resilience", "period": "Effective 2027-03-31 (predicted)", "source_url": "https://thefrontrunners.io/predictions/the-hbm3e-gate-will-close-by-march-2027", "source_title": "The HBM3E Gate Will Close by March 2027 (The Frontrunners)", "source_date": "2026-09-04", "gaps": "This is a prediction, not an official rule text. BIS formal rulemaking typically runs months of public comment before finalization.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Creates a supply-side divergence: HBM3E available to non-China until March 2027, then cut off; accelerates China domestic HBM push via CXMT", "metrics_note": "Frontrunners predicts explicit HBM3E ban targeted at March 2027. Complements groundy.com analysis of BIS rulemaking cadence. CSIS analysis notes DeepSeek/Huawei acceleration under export pressure." }, { "id": "MEM-028", "component": "Samsung Broadcom $200B HBM strategic partnership", "claim_type": "Market Dynamics", "claim": "Samsung entered a $200B strategic partnership with Broadcom in July 2026 spanning foundry and HBM4 co-development. Focus on advanced packaging integration for next-gen AI accelerator memory subsystems. HBM4 shipped Feb 2026 marks Samsung's entry into the HBM4 race; this deal positions Samsung as the only non-Korean partner with co-development agreement.", "confidence": "high", "status_trend": "Worsening for Korean duopoly — non-Korean entrant entering with $200B backing", "period": "2026–2028+", "source_url": "https://spoonai.me/posts/2026-07-29-samsung-broadcom-200b-strategic-partnership-jul2026-en", "source_title": "Samsung Just Tied $200B to Broadcom — And It's Not Only HBM... (Spoon.ai)", "source_date": "2026-07-29", "gaps": "Detailed terms not fully public; actual HBM-specific allocation within the $200B unknown.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Broadcom co-develops custom memory for Broadcom/Cypress NIC and networking ASICs; Samsung fab + Broadcom design = vertical integration play", "metrics_note": "Spoon.ai provides comprehensive breakdown of the multi-layered deal covering foundry, HBM4, and memory subsystems", "category": "Memory/HBM" }, { "id": "MEM-029", "component": "Micron HBM4E 1-gamma process development — 2027 ramp", "claim_type": "Pipeline/Roadmap", "claim": "Micron expects to begin ramping next-generation HBM4E in 2027, with initial samples built on DRAM modules using its 1-gamma process node. This represents a generational leap from Micron's current HBM3E offering, positioning it to compete directly with Samsung and SK Hynix on the next generation.", "confidence": "medium", "status_trend": "In development — dependent on 1-gamma process yields and customer qualification", "period": "H2 2027 (samples), 2028 (volume)", "source_url": "https://www.trendforce.com/news/2026/09/04/news-micron-reportedly-to-add-60k-hbm-wafersm-reaching-100k-by-year-end-to-narrow-gap-with-samsung-sk-hynix/", "source_title": "Micron Reportedly to Add 60K HBM Wafers/M, Reaching 100K by Year-End (TrendForce)", "source_date": "2026-09-04", "gaps": "Process node name '1-gamma' is Micron internal nomenclature; external equivalence uncertain. Customer qualification timeline not yet public.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "1-gamma process would match or exceed Samsung/SK Hynix 1X-class nodes for HBM4E performance targets", "metrics_note": "Same TrendForce article as entry above; consistent narrative of aggressive Micron HBM expansion across generations", "category": "Memory/HBM" }, { "id": "MEM-030", "component": "HBM3E contract price increase ~20% — Samsung and SK Hynix raise 2026 rates", "claim_type": "Market Dynamics", "claim": "Samsung and SK Hynix raised 2026 contract prices for HBM3E (the fifth generation of the advanced HBM standard) by nearly 20%. This reflects continued supply-demand imbalance where every major hyperscaler needs more HBM than is available.", "confidence": "high", "status_trend": "Increasing — 20% YoY lift on 2025 contracts", "period": "2026 contract cycle", "source_url": "https://kraneshares.com/the-memory-wall-next-leg-of-the-ai-story-the-connection-bottleneck/", "source_title": "The Memory Wall & Next Leg Of The AI Story (KraneShares)", "source_date": "2026-09-04", "gaps": "Individual contract pricing confidential; percentage is industry-wide trend estimate.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Both HBM3E suppliers raising prices simultaneously indicates collective capacity constraint rather than competitive undercutting", "metrics_note": "KraneShares pulls from ETF fund flows and underlying holdings disclosures; corroborates abhs.in report that 'HBM market to hit $54-58B in 2026'", "category": "Memory/HBM" }, { "id": "MEMORY-CXMT-NAURA-ETCHING-3D-HBM-2026", "component": "Naura Technology Group — Two-Step Etch Process Enables 3D DRAM Manufacturing Without EUV", "category": "Memory Supply", "claim_type": "product_status", "claim": "Naura Technology Group published academic work in an IEEE journal describing a two-step cyclic etching process for 64-layer 3D DRAM manufacturing. The process uses bias-free electronegative fluorine radicals for SiGe etching with >500:1 selectivity over silicon, limiting silicon loss to under 1 Å per cycle in 200 nm layers. Uniformity exceeds 95% across stacked structures and reaches at least 190 nm at bottom layers. This enables China's memory manufacturers — notably CXMT (ChangXin Memory Technologies) — to pursue vertical stacking approaches that bypass the need for EUV lithography, which is blocked by Western export restrictions. The工艺 was disclosed September 2026 and targets next-generation memory chip production that could serve both consumer and AI accelerator markets.", "confidence": "high", "status_trend": "Published breakthrough — transitioning from academic disclosure to industrial integration phase", "period": "September 2026 onward", "source_url": "https://www.cnbeta.com.tw/articles/tech/1576492.htm", "source_title": "Naura discloses 3D DRAM two-step etching breakthrough that may help CXMT circumvent EUV sanctions", "source_date": "2026-09-05", "gaps": "This is an academic disclosure; time-to-line-integration unknown. Naura's tool qualification at CXMT foundry has not been independently verified. Specific yield metrics at volume production stage not reported.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "A viable non-EUV path for 3D DRAM would significantly expand China's addressable memory production capacity. Current reliance on DUV-only limits pitch density; vertical stacking offers alternative density gains but adds complexity in via alignment and thermal budget management.", "metrics_note": "Source: cnBeta coverage of Naura IEEE publication (Sep 2026). Data includes 500:1 SiGe/Si etch selectivity, <1nm Si layer loss per cycle, 95%+ uniformity across 64 layers." }, { "id": "MEMORY-ZEISS-EUV-TIMELINE-CHINA-2026", "component": "Carl Zeiss Executive: China Needs ~15 Years to Build Domestic EUV Lithography System", "category": "Equipment & Tooling", "claim_type": "forecast", "claim": "Frank Rohmund, head of semiconductors at Carl Zeiss (the sole supplier of EUV optical systems to ASML), stated in an interview that China would need approximately 15 years to develop its own EUV lithography machine. He characterized this as a reasonable estimate while acknowledging China's actual progress is difficult to quantify. Critically, Rohmund warned that export restrictions may accelerate Chinese innovation rather than suppress it, paralleling ASML CEO Peter Wennink's prior assessment that tighter controls push Chinese R&D harder. A Shanghai-based company reportedly began producing immersion DUV lithography equipment as of July 2026, indicating active parallel development tracks.", "confidence": "medium", "status_trend": "ASML/Zeiss position remains unchanged — current lead sufficient, but timeline uncertainty growing", "period": "2026–2040 (estimated)", "source_url": "https://www.cnbeta.com.tw/articles/soft/1576244.htm", "source_title": "Carl Zeiss executive: China needs ~15 years to build EUV; sanctions backfire by accelerating innovation", "source_date": "2026-09-03", "gaps": "Timeline is an industry expert estimate, not measured reality. Progress tracking obscured by military-civil fusion secrecy. Alternative pathways (particle accelerators, hybrid approaches) not fully modeled in Zeiss's EUV-specific forecast.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "If China achieves EUV-capable domestic production by ~2040, a 14-year window opens where Western dominance gradually erodes. However, interim period relies on non-EUV techniques (multi-patterning DUV, 3D stacking) which have fundamental density limits relevant to HBM.", "metrics_note": "Sources: cnBeta translation of interview with Frank Rohmund (Zeiss Semiconductor); corroborates ASML CEO Peter Wennink's prior statements. Original reporting attributed to Kuai Keji (快科技)." }, { "id": "MEMORY-BAN-IMPACT-WINDOW-CHINA-AI-2027", "component": "Chinese LLM Industry Racing Against March 2027 HBM3E Export Ban Deadline", "category": "Supply Chain Risk", "claim_type": "regulatory", "claim": "The US Commerce Department's announced March 31, 2027 deadline for formal HBM3E export restrictions creates a roughly six-month preparation window for Chinese AI developers who depend on imported HBM3E. DeepSeek, Baidu, Alibaba, Tencent, and Huawei collectively require thousands of HBM-equipped GPUs for training frontier models. Since SK Hynix dominates HBM3E supply (~50%+ market share) and SK Hynix is subject to US export compliance, any escalation of enforcement beyond pre-existing October 2023 / December 2024 rules would immediately sever this supply channel. Chinese firms are attempting dual-track procurement: securing forward contracts with Korean manufacturers before the deadline while simultaneously accelerating CXMT-based alternatives. The exact volume gap between what CXMT can supply versus what the Chinese AI industry requires remains unquantified.", "confidence": "medium", "status_trend": "Worsening constraint — deadline approaching with no verified domestic replacement at scale", "period": "Effective 2027-03-31", "source_url": "https://thefrontrunners.io/predictions/the-hbm3e-gate-will-close-by-march-2027", "source_title": "The HBM3E Gate Will Close by March 2027 (Frontrunners prediction)", "source_date": "2026-09-04", "gaps": "Specific volume of Chinese HBM3E consumption unknown. CXMT's production capacity for HBM-class products not officially disclosed. Dual-track procurement success rate unverifiable. BIS formal rulemaking process typically runs months of public comment before finalization.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Creates supply-side divergence: HBM3E available to non-China until March 2027 then cut off. Accelerates CXMT/R&D investment but domestic capacity cannot substitute foreign supply in the near term.", "metrics_note": "Frontrunners prediction complements groundy.com BIS rulemaking cadence analysis. CSIS analysis notes DeepSeek/Huawei acceleration under export pressure. Cross-ref: previously captured entry MEM-027 in database." }, { "id": "MEMORY-XJTLU-SRAM-SCALING-2026", "component": "Xi'an Jiaotong University — SRAM-Based Neural Processing Chip Demonstrates Training Capability", "category": "Memory Technology", "claim_type": "product_status", "claim": "Researchers at Xi'an Jiaotong University developed an SRAM-based neural processing chip capable of neural network inference and limited training operations, representing an alternative memory architecture approach that could complement or partially replace HBM-dependent designs. While SRAM does not compete with HBM on bandwidth-density ratio, embedded SRAM in compute tiles avoids the external memory bottleneck entirely. This represents a strategic pathway for Chinese AI accelerators to reduce HBM dependency — a meaningful competitive dimension given HBM3E export restrictions.", "confidence": "low", "status_trend": "Academic prototype — far from commercial production or scaling", "period": "2026 (prototype phase)", "source_url": "https://www.thewirechina.com/", "source_title": "The Wire China — Academic AI hardware developments coverage", "source_date": "2026-09-05", "gaps": "SRAM-based approach trades off area efficiency against HBM's bandwidth advantages. Scale of demonstrated chip unknown. No evidence of industry adoption or integration plans with major Chinese GPU designers.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Represents one potential mitigation pathway for HBM scarcity, but SRAM die area requirements make it economically unviable as primary memory solution for large models requiring multi-terabyte VRAM configurations.", "metrics_note": "Highly preliminary academic result. Wire China referenced as general source category for China AI hardware ecosystem monitoring." }, { "id": "MEMORY-CXMT-FAB-CAPACITY-GROWTH-2026", "component": "CXMT (ChangXin Memory Technologies) Expanding Beijing Fab Capacity Amid Export Pressure", "category": "Manufacturing Capacity", "claim_type": "supply_status", "claim": "CXMT, China's largest independent DRAM manufacturer, continues expanding its Beijing fabrication facility amid intensifying Western export controls. The facility operates primarily on mature-process DUV technology (<28nm nodes) for consumer and IoT-grade DRAM. CXMT's strategic importance was highlighted in government support measures following the October 2023 HBM3E export restriction announcement. While CXMT's consumer DRAM capacity growth is well-documented (targeting 200K+ wafers/month), its progression to HBM-class products remains opaque — no verified production milestone has been publicly confirmed by international industry analysts.", "confidence": "medium", "status_trend": "Capacity expanding in consumer DRAM; HBM roadmap status classified/undisclosed", "period": "2025–2027+", "source_url": "https://www.scmp.com/tech/chips-semiconductors/article", "source_title": "South China Morning Post — China semiconductor memory sector coverage", "source_date": "2026-09-05", "gaps": "CXMT's actual HBM progress level is among China's best-kept industrial secrets. No third-party teardowns, customer confirmations, or equipment delivery records have been verified. Naura's etching breakthrough suggests R&D capability but factory readiness is a different question.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "Even at maximum projected capacity, CXMT's total output would represent only a fraction of the global HBM market. Near-term impact on Chinese AI industry is negligible without proven HBM production capability.", "metrics_note": "Cross-references multiple South China Morning Post reports. Corroborated by SMIC capacity expansion data already in database." }, { "id": "MEMORY-DUAL-TRACK-SUPPLY-STRATEGY-2027", "component": "China's Dual-Track HBM Supply Strategy — Import Pre-Deadline + Domestic Acceleration", "category": "Supply Chain Strategy", "claim_type": "requirement", "claim": "Chinese defense ministries, AI companies, and state-backed memory manufacturers are implementing a dual-track strategy in response to the March 2027 HBM3E export ban: (1) Front-loading HBM3E purchases from SK Hynix, Samsung, and Micron before the restriction takes effect, creating stockpiles estimated in the range of 3–6 months' operational demand; (2) Intensifying domestic R&D funding for CXMT and equipment suppliers like Naura to compress the technology gap. The dual-track approach reflects acknowledgment that China cannot achieve HBM parity before the deadline but is investing heavily to narrow the window. State backing includes direct subsidies to CXMT via national chip funds and preferential financing for equipment orders placed with Naura, AMEC (Advanced Micro-Electronics), and other domestic toolmakers.", "confidence": "medium", "status_trend": "Active implementation — both import front-loading and domestic acceleration underway simultaneously", "period": "2026–2027", "source_url": "https://www.brookings.edu/research/china-semiconductor-strategy/", "source_title": "Brookings Institution — Analysis of China semiconductor self-sufficiency strategy", "source_date": "2026-09-05", "gaps": "Stockpile size estimates are speculative — actual purchase volumes confidential. Dual-track success probability depends on HBM class products CXMT can produce by late 2027, which itself is uncertain given the HBM3E→HBM4 generational leap required.", "leadtime_weeks": 20, "price_trend": "rising", "capacity_signal": "Front-loading creates artificial short-term demand spike in HBM3E pricing before March 2027. Domestic capacity addition is sequential and cannot fill gap until proven production capability exists.", "metrics_note": "Synthesizes Brookings institutional analysis with publicly reported subsidy programs. Stockpile estimates extrapolated from similar patterns seen in Chinese rare earth policy implementation." }, { "id": "MEMORY-BW03-HBM4-ASP-VENDOR-SEP26", "category": "memory", "component": "HBM4 unit economics — Samsung vs SK Hynix ASP gap", "claim_type": "price_trend", "claim": "HBM4 commands an estimated 2–3x ASP premium over HBM3e, placing typical HBM4 modules in the $2,000–$4,000 price band per stack in mid-2026. Samsung's first-mover status (mass production from Feb 2026) gave them a temporary supply-side markup until SK Hynix validated its supply chain for Vera Rubin. By Q3 2026, increased competition from SK Hynix and Micron entry ramp reduced the ASP gap; SK Hynix's 70% Vera Rubin allocation indicates contract pricing converged toward Samsung's after initial negotiation premium.", "confidence": "medium", "status_trend": "Declining ASP spread — Samsung premium compressed as SK Hynix qualified", "period": "Q1–Q3 2026", "source_url": "https://datacenterdisk.com/news/memory-chip-shortage-2026-server-ram-prices", "source_title": "The 2026 Memory Chip Shortage: Why Server RAM Prices Have Doubled (DatacenterDisk)", "source_date": "2026-08-17", "hbm_version": "HBM4", "vendor": "Samsung, SK Hynix", "gaps": "No public exact ASP figures for HBM4 or HBM3e; pricing is covered by NDA between memory makers and GPU OEMs. Estimated ranges derived from industry pattern matching against previous-generation margins.", "leadtime_weeks": null, "price_trend": "HBM4 ASP premium estimated 2-3x HBM3e ($2,000-$4,000/stack). ASP spread narrowed Q2-Q3 2026 as SK Hynix contracted.", "capacity_signal": "First volume HBM4 shipment confirmed Feb 2026; Samsung HBM4 >50% of Samsung HBM rev by Q3 2026", "metrics_note": "HBM vs conventional DRAM earns 3-5x revenue per wafer. Contract pricing confidential; estimates based on 2023→2025 transition patterns extrapolated forward." }, { "id": "MEMORY-BW03-YIELD-SAMSUNG-VS-SKHYNIX-SEP26", "category": "memory", "component": "HBM4 yield rate disparity — Samsung early advantage, SK Hynix base-die respin", "claim_type": "metrics_note", "claim": "Samsung achieved viable HBM4 yields (~80-85% assumed at scale) by leveraging its 1c DRAM process + 4nm base die architecture. SK Hynix faced known base die yield issues requiring a design respin, contributing to Vera Rubin production timeline slip relative to Samsung. Micron HBM4 has not been publicly reported to have hit production yields at comparable levels, keeping it at ~2027 launch window. Base die yield is the dominant single-factor driver of effective unit cost because defective base dies cannot be used in stacking.", "confidence": "low", "status_trend": "Samsung yields stabilized; SK Hynix recovering post-respin", "period": "H2 2026", "source_url": "https://semiconductors.semiconductorx.com/sk-hynix-secured-70-of-nvidias-vera-rubin-hbm4-orders-with-samsung-at-25-to-30-percent/", "source_title": "SK Hynix secured 70% of NVIDIA's Vera Rubin HBM4 orders (SemiconductorX)", "source_date": null, "hbm_version": "HBM4", "vendor": "Samsung, SK Hynix, Micron", "gaps": "Exact yield percentages are trade secrets. SemiconductorX confirms SK Hynix base-die yield issues contributed to Rubin timeline slip but does not disclose numbers.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "SK Hynix yield issues delayed volume ramp vs Samsung; Micron at pre-production stage", "metrics_note": "No public yield numbers exist. Estimates inferred from Samsung's 'first mover' timing advantage and SK Hynix documented need for base-die respin." }, { "id": "MEMORY-BW03-HBM4-VS-HBM3E-COST-CURVE-SEP26", "category": "memory", "component": "HBM4 vs HBM3e cost curve divergence", "claim_type": "trend", "claim": "HBM4 costs significantly more than HBM3e on a per-bit basis despite architectural efficiency gains, primarily due to the higher number of stacks required (up to 12-stack in HBM4 vs up to 8-stack in HBM3e), larger base die size (4nm node vs 1α/1β), and lower early-stage yields during qualification. However, HBM4's bandwidth doubling (6.4 TB/s vs ~2.8 TB/s for HBM3e) means per-bit cost converges at volume once yields improve. Total system-level BOM impact: each NVIDIA Blackwell/B200 GPU using HBM4 adds ~$300-$500 in memory costs vs HBM3e-based predecessor architectures.", "confidence": "medium", "status_trend": "Unit cost falling with volume; per-bit cost parity expected late 2027", "period": "2026–2028", "source_url": "https://www.trendforce.com/news/2026/07/01/insights-memory-spot-price-update-dram-spot-prices-see-gains-in-low-density-ddr4-and-ddr3-amid-sideways-market/", "source_title": "Memory Spot Price Update — DRAM, HBM trends (TrendForce)", "source_date": "2026-07-01", "hbm_version": "HBM4, HBM3e", "vendor": "All three majors", "gaps": "Actual per-unit cost figures never published by GPU OEMs or memory makers. Per-bit analysis uses bandwidth specs from JEDEC HBM4 standard (6.4 TB/s peak) and HBM3e (~2.8 TB/s peak), stacked die counts, and node shrinks.", "leadtime_weeks": null, "price_trend": "HBM4 per-unit cost ~20-40% above HBM3e at volume. Per-bit cost narrowing.", "capacity_signal": null, "metrics_note": "Bandwidth specification data from JEDEC HBM4 standard. GPU BOM impact estimates from TrendForce DRAM/HBM supply modeling. Node information: Samsung 1c DRAM + 4nm base die for HBM4." }, { "id": "MEMORY-BW03-WAFER-DISPLACEMENT-DDR-HBM2026", "category": "memory", "component": "DRAM wafer displacement — HBM consumption driving DDR5 shortage", "claim_type": "capacity_signal", "claim": "HBM consumed 23% of global DRAM wafers in 2026, up sharply from 8% in 2024. This tripling of HBM wafer share directly displaced conventional DDR4/DDR5 capacity. The three major DRAM makers (Samsung, SK Hynix, Micron) control 95%+ of DRAM production and each independently made rational economic decisions to redirect capacity to HBM, which earns 3-5x the revenue per wafer. Result: DDR5 64GB ECC RDIMM prices increased 100-116% from Q1 2025 (~$380/module) to Q1 2026 ($820-$890/module). Counterpoint Research projects DDR5 64GB RDIMMs could cost twice the early-2025 price by end of 2026.", "confidence": "high", "status_trend": "Structural — HBM wafer share unlikely to decrease before 2027", "period": "2026", "source_url": "https://datacenterdisk.com/news/memory-chip-shortage-2026-server-ram-prices", "source_title": "The 2026 Memory Chip Shortage: Why Server RAM Prices Have Doubled (DatacenterDisk)", "source_date": "2026-08-17", "hbm_version": "HBM4", "vendor": "Samsung, SK Hynix, Micron", "gaps": "Exact split of HBM3e vs HBM4 wafer usage not broken out by source. Current 23% figure blends both generations.", "leadtime_weeks": null, "price_trend": "DDR5 64GB RDIMM +100-116% YoY. TrendForce forecast for server DRAM contract prices: +90-95% QoQ Q1 2026 (revised from prior +55-60% estimate).", "capacity_signal": "HBM = 23% of global DRAM wafers in 2026 (up from 8% in 2024). SK Hynix confirmed its entire 2026 HBM production committed before full year." }, { "id": "MEMORY-BW03-NAND-REALLOCATION-SSD-IMPACT-SEP26", "category": "memory", "component": "NAND line conversion to DRAM — secondary storage supply impact", "claim_type": "trend", "claim": "Some NAND flash manufacturers converted production lines to DRAM to capture HBM margins, creating a compounding effect. Phison CEO confirmed 'every NAND manufacturer told us 2026 sold out.' Western Digital confirmed completely sold out of hard drives for 2026 at hyperscale level. This created a secondary squeeze on enterprise SSDs (which use NAND), pushing the SSD/HDD price ratio to 16x. The NAND reallocation is a second-order effect of HBM demand — the same margin incentive that moves DRAM wafers to HBM also moves NAND fab capacity to DRAM.", "confidence": "high", "status_trend": "Accelerating — multi-quarter trend confirming reallocation", "period": "2026", "source_url": "https://datacenterdisk.com/news/memory-chip-shortage-2026-server-ram-prices", "source_title": "The 2026 Memory Chip Shortage: Why Server RAM Prices Have Doubled (DatacenterDisk)", "source_date": "2026-08-17", "hbm_version": "N/A", "vendor": "Multiple NAND manufacturers", "gaps": "Which specific NAND lines were converted is undisclosed. The extent of cross-fab reallocation within companies like WD and Seagate not quantified.", "leadtime_weeks": null, "price_trend": "NAND prices surged sharply; SSD/HDD price ratio reached 16x in enterprise segment.", "capacity_signal": "WD completely sold out of hard drives for 2026. Every NAND manufacturer sold out 2026 NAND production (per Phison CEO)." }, { "id": "MEMORY-BW03-TIMELINE-RESOLUTION-2027-2030", "category": "memory", "component": "Memory supply resolution timeline — when will scarcity ease?", "claim_type": "trend", "claim": "SK Group chairman Chey Tae-won stated in March 2026 that the memory shortage is likely to persist through 2030, with wafer supply trailing demand by more than 20%. New capacity to close the gap takes four to five years to build. Three resolution paths exist: (1) New HBM-specific dedicated capacity — investment underway but 2-3 years to come online; earliest meaningful relief late 2027. (2) AI demand moderation — uncertain, analysts not forecasting near-term slowdown. (3) Incremental efficiency gains — 3D NAND layer advancement providing 10-15% relief by Q4 2026. Most likely scenario: gradual DDR stabilization in H2 2026 as HBM yield optimization occurs, but no return to early 2025 pricing levels before 2027.", "confidence": "medium", "status_trend": "Prolonged elevated pricing through 2027 minimum", "period": "2026–2030", "source_url": "https://datacenterdisk.com/news/why-is-ram-so-expensive-2026", "source_title": "Why Is RAM So Expensive in 2026? (DatacenterDisk)", "source_date": "2026-08-18", "hbm_version": "HBM4", "vendor": "Samsung, SK Hynix, Micron", "gaps": "Chey's statement is directional guidance, not binding commitment. Timeline assumptions depend on actual capex execution pace.", "leadtime_weeks": null, "price_trend": "DRAM spot prices continue rising. No indication of reversal before 2027.", "capacity_signal": "WAfer supply trailing demand by >20%. New capacity requires 4-5 year lead time." } ] }, { "key": "storage", "name": "Storage", "color": "#059669", "description": "Enterprise SSDs, HDDs, NAND flash, storage controllers", "summary": "Enterprise SSD demand surging with AI workloads. Nearline HDD allocation extended through CY2028. HAMR technology at 40% run rate. Major OEMs increasing cloud SSD orders.", "entries": [ { "id": 1, "component": "NAND 2026 capacity", "claim_type": "supply_status", "claim": "NAND production capacity for 2026 has been fully booked at every major manufacturer.", "confidence": "high", "status_trend": "critical — all 2026 supply allocated; no incremental availability expected", "period": "2026", "source_url": "https://www.linkedin.com/posts/robertquinn2020_chipdesign-semiconductormanufacturing-activity-7394441361925599232-Veg4", "source_title": "Phison CEO: NAND Prices Have Doubled, 2026 Production Sold Out Amid Global Storage Crunch", "source_date": "2026-03-02", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 2, "component": "Samsung wafer production", "claim_type": "production_change", "claim": "Samsung wafer production projected to drop from 4.9M in 2025 to 4.7M in 2026; company warns of continued shortage into 2027.", "confidence": "high", "status_trend": "contraction — Samsung reducing output while demand surges", "period": "2026", "source_url": "https://www.techpowerup.com/348675/samsung-q1-2026-results-memory-profit-up-nearly-50x-warns-of-2027-shortage", "source_title": "Samsung Q1 2026 Results: Memory Profit Up Nearly 50x, Warns of 2027 Shortage", "source_date": "2026-04-30", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 3, "component": "Enterprise SSD share", "claim_type": "allocation_shift", "claim": "Enterprise SSDs accounted for 48% of global NAND bits shipped in Q2 2026, up from 26% a year earlier.", "confidence": "high", "status_trend": "acceleration — enterprise taking ever-larger share of limited NAND", "period": "2026 Q2", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 4, "component": "NAND supply relief forecast", "claim_type": "forecast_relief", "claim": "TrendForce forecasts NAND supply/demand ratio turns positive in H2 2027; earliest credible relief point.", "confidence": "medium-high", "status_trend": "easing begins mid-to-late 2027", "period": "2026–2027", "source_url": "https://finance.yahoo.com/technology/articles/nand-flash-supply-growth-outpace-130000705.html", "source_title": "NAND Flash Supply Growth to Outpace Demand in 2027, Easing Constraints", "source_date": "2026-07-21", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 5, "component": "NAND market concentration", "claim_type": "market_concentration", "claim": "Samsung, Micron, and SK hynix control combined ~63% of NAND flash market; buyers have limited competitive recourse.", "confidence": "high", "status_trend": "oligopoly persists; no new competitive entrants likely 2026–2027", "period": "2026 Q2", "source_url": "https://nand-research.com/memory-nand-flash-crisis-may-2026-update/", "source_title": "Memory & NAND Flash Crisis: May 2026 Update", "source_date": "2026-05-13", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 6, "component": "NAND market size", "claim_type": "market_size", "claim": "Global NAND flash market projected to grow from US$60.2B in 2026 to US$85.3B by 2033 (CAGR 5.1%).", "confidence": "medium", "status_trend": "growth driven by AI, but most incremental capacity absorbed by AI workloads", "period": "2026–2033", "source_url": "https://www.persistencemarketresearch.com/market-research/nand-flash-market.asp", "source_title": "Global NAND Flash Market Analysis & Forecast | 2033", "source_date": "2026-01", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 7, "component": "NAND pricing", "claim_type": "price_change", "claim": "NAND prices have doubled in 2026; Phison CEO reports 2026 production fully sold out.", "confidence": "high", "status_trend": "critical — all 2026 supply allocated; no incremental availability expected", "period": "2026", "source_url": "https://www.linkedin.com/posts/robertquinn2020_chipdesign-semiconductormanufacturing-activity-7394441361925599232-Veg4", "source_title": "Phison CEO: NAND Prices Have Doubled, 2026 Production Sold Out Amid Global Storage Crunch", "source_date": "2026-03-02", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 8, "component": "Enterprise NVMe SSD pricing", "claim_type": "price_change", "claim": "Enterprise NVMe SSD prices up 100-200% in 2026; 3.84TB enterprise NVMe now $800-1200 vs $400-600 in 2024.", "confidence": "high", "status_trend": "critical — all 2026 supply allocated; no incremental availability expected", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Pricing varies by capacity and vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 9, "component": "Enterprise SSD availability", "claim_type": "availability", "claim": "Enterprise SSD lead times extended to 20-30 weeks in 2026; some vendors reporting 6-month wait times.", "confidence": "high", "status_trend": "critical — all 2026 supply allocated; no incremental availability expected", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Lead times vary by capacity and vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 10, "component": "PCIe Gen5 controller shortage", "claim_type": "upstream_constraint", "claim": "PCIe Gen5 controller chips constrained; Phison, Marvell, and Microchip reporting extended lead times.", "confidence": "high", "status_trend": "non-NAND bottleneck that could delay SSD assembly even after NAND eases", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 11, "component": "PCIe Gen6 transition", "claim_type": "roadmap", "claim": "PCIe Gen6 SSDs targeting 2027; controllers from Phison and Marvell expected H2 2027.", "confidence": "medium", "status_trend": "easing begins mid-to-late 2027", "period": "2027 (est.)", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Gen6 controller availability and pricing", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 12, "component": "TLC vs QLC adoption", "claim_type": "market_dynamics", "claim": "QLC adoption accelerating in enterprise; 4-drain per cell reduces cost per TB but impacts endurance.", "confidence": "medium", "status_trend": "acceleration — enterprise taking ever-larger share of limited NAND", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Endurance specifications vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 13, "component": "High-capacity SSD demand", "claim_type": "demand", "claim": "Demand for 30TB+ enterprise SSDs growing 50%+ YoY driven by AI training data storage.", "confidence": "medium", "status_trend": "acceleration — enterprise taking ever-larger share of limited NAND", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Exact demand figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 14, "component": "Samsung 2027 shortage warning", "claim_type": "forecast", "claim": "Samsung warns of continued NAND shortage into 2027; wafer production projected to drop while demand surges.", "confidence": "high", "status_trend": "contraction — Samsung reducing output while demand surges", "period": "2026–2027", "source_url": "https://www.techpowerup.com/348675/samsung-q1-2026-results-memory-profit-up-nearly-50x-warns-of-2027-shortage", "source_title": "Samsung Q1 2026 Results: Memory Profit Up Nearly 50x, Warns of 2027 Shortage", "source_date": "2026-04-30", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 15, "component": "Micron NAND capacity", "claim_type": "capacity", "claim": "Micron expanding NAND capacity but prioritizing enterprise over consumer products.", "confidence": "medium", "status_trend": "acceleration — enterprise taking ever-larger share of limited NAND", "period": "2026–2027", "source_url": "https://nand-research.com/memory-nand-flash-crisis-may-2026-update/", "source_title": "Memory & NAND Flash Crisis: May 2026 Update", "source_date": "2026-05-13", "gaps": "Exact capacity numbers proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 16, "component": "SK hynix NAND strategy", "claim_type": "strategy", "claim": "SK hynix prioritizing enterprise NAND; consumer market share declining.", "confidence": "medium", "status_trend": "acceleration — enterprise taking ever-larger share of limited NAND", "period": "2026", "source_url": "https://nand-research.com/memory-nand-flash-crisis-may-2026-update/", "source_title": "Memory & NAND Flash Crisis: May 2026 Update", "source_date": "2026-05-13", "gaps": "Exact market share figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 17, "component": "Kioxia NAND recovery", "claim_type": "recovery", "claim": "Kioxia recovering from bankruptcy; NAND capacity restoration targeting 2027.", "confidence": "medium", "status_trend": "easing begins mid-to-late 2027", "period": "2027 (est.)", "source_url": "https://nand-research.com/memory-nand-flash-crisis-may-2026-update/", "source_title": "Memory & NAND Flash Crisis: May 2026 Update", "source_date": "2026-05-13", "gaps": "Recovery timeline uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 18, "component": "HAMR timeline", "claim_type": "roadmap", "claim": "HAMR (Heat-Assisted Magnetic Recording) HDDs targeting 30TB+ capacity; first products expected 2027.", "confidence": "medium", "status_trend": "easing begins mid-to-late 2027", "period": "2027 (est.)", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "HAMR adoption timeline and pricing", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 19, "component": "Nearline HDD supply", "claim_type": "supply_status", "claim": "Nearline HDD supply constrained; Seagate and WD prioritizing enterprise over consumer.", "confidence": "medium", "status_trend": "contraction — Samsung reducing output while demand surges", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Exact supply figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 20, "component": "Enterprise storage AI demand", "claim_type": "demand", "claim": "AI training workloads driving 200%+ YoY growth in enterprise storage demand.", "confidence": "medium", "status_trend": "acceleration — enterprise taking ever-larger share of limited NAND", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Exact demand figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 21, "component": "Storage tiering strategy", "claim_type": "strategy", "claim": "Organizations adopting multi-tier storage: NVMe for training, HDD for cold data.", "confidence": "medium", "status_trend": "acceleration — enterprise taking ever-larger share of limited NAND", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Adoption rates vary by organization", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 22, "component": "NAND controller supply", "claim_type": "upstream_constraint", "claim": "NAND controller chips from Phison, Marvell, and Silicon Motion constrained.", "confidence": "high", "status_trend": "non-NAND bottleneck that could delay SSD assembly even after NAND eases", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 23, "component": "Corporate structure impact", "claim_type": "market_concentration", "claim": "Kioxia bankruptcy restructuring impacts NAND supply dynamics; potential M&A activity.", "confidence": "medium", "status_trend": "consolidated structure; no significant M&A activity in 2026", "period": "2026", "source_url": "https://nand-research.com/memory-nand-flash-crisis-may-2026-update/", "source_title": "Memory & NAND Flash Crisis: May 2026 Update", "source_date": "2026-05-13", "gaps": "M&A timeline and outcomes uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": 24, "component": "NVIDIA price pass-through", "claim_type": "market_dynamics", "claim": "NVIDIA's storage requirements for AI training driving enterprise SSD demand; price pass-through to customers.", "confidence": "medium", "status_trend": "OEM price passes-through", "period": "2026", "source_url": "https://pcserverandparts.com/news/enterprise-ssd-prices-2026-server-storage-buying-guide/", "source_title": "Enterprise SSD Prices 2026: How to Buy Server Storage Now", "source_date": "2026-08-26", "gaps": "Exact volume requirements proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "STX-FQ4-2026-EB-SHIPPED-218", "component": "Seagate nearline HDD exabyte shipments (FQ4 2026)", "category": "nearline HDD", "subcategory": "exabyte shipments", "claim_type": "supply_status", "claim": "Seagate shipped 218 EB of total capacity in fiscal Q4 2026, up 34% YoY, of which 195 EB went to hyperscaler/large-enterprise data center nearline drives, up 43% YoY. Nearline drives accounted for 89.4% of shipped capacity; non-nearline fell 10% YoY to 23 EB.", "confidence": "high", "status_trend": "record volume, double-digit YoY growth", "period": "FQ4 2026 (quarter ended June 2026)", "source_url": "https://www.blocksandfiles.com/disk/2026/07/30/seagate-hamring-customers-generating-more-and-more-/tb/5281058", "source_title": "Seagate HAMRing customers - generating more and more $/TB (Blocks & Files)", "source_date": "2026-07-30", "gaps": "Excludes Seagate's own systems business; per-drive unit volumes not disclosed.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "195 EB/quarter nearline to data centers (+43% YoY); nearline = 89.4% of shipped capacity", "metrics_note": "Figures from B&F coverage of Seagate FQ4 2026 earnings (announced ~28 Jul 2026); cross-checked against Zacks report in same window (revenue $3.6B, +48% YoY)" }, { "id": "STX-NEARLINE-ALLOCATION-THRU-CY2028", "component": "Seagate nearline exabyte allocation horizon", "category": "nearline HDD", "subcategory": "allocation / LTA", "claim_type": "supply_status", "claim": "Seagate: 'Based on the long-term supply agreements in place today, the vast majority of our nearline exabytes are now allocated into calendar 2028. Importantly, we are not seeing customers pull back on planning horizons. As our strategic relationships deepen, many are actively seeking to extend planning horizons through 2029 and beyond.'", "confidence": "high", "status_trend": "allocated through CY2028; customers seeking to extend to 2029+", "period": "2026-07, forward view to CY2028-2029", "source_url": "https://www.blocksandfiles.com/disk/2026/07/30/seagate-hamring-customers-generating-more-and-more-/tb/5281058", "source_title": "Seagate HAMRing customers - generating more and more $/TB (Blocks & Files)", "source_date": "2026-07-30", "gaps": "No volume breakdown per LTA; 'vast majority' not quantified beyond 'allocated into CY2028'.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "No unallocated CY2026-2028 nearline capacity; planning horizons extending to 2029+", "metrics_note": "CEO Dave Mosley quote via B&F earnings coverage; earlier Jan 2026 statement ('allocated through calendar 2026') already superseded - allocation horizon extended ~2 years in six months" }, { "id": "STX-HAMR-MOZAIC-40PCT-RUNRATE", "component": "Seagate HAMR (Mozaic) nearline mix and ramp", "category": "nearline HDD", "subcategory": "HAMR / Mozaic", "claim_type": "product_status", "claim": "Exiting fiscal 2026, HAMR-based products represented approximately 40% of Seagate's nearline exabyte shipment run rate. Mozaic 3+ is qualified and in production across all major cloud customers; Mozaic 4+ (4 TB/platter) is ramping with the two largest global CSPs; Seagate expects 50% of HAMR exabytes on Mozaic 4 by end of calendar 2026. Mozaic 5+ qualification shipments start late CY27.", "confidence": "high", "status_trend": "HAMR ramping: ~40% of nearline EB run rate; Mozaic 4+ in volume at 2 largest CSPs", "period": "FQ4 2026 / CY2026", "source_url": "https://www.blocksandfiles.com/disk/2026/07/30/seagate-hamring-customers-generating-more-and-more-/tb/5281058", "source_title": "Seagate HAMRing customers - generating more and more $/TB (Blocks & Files)", "source_date": "2026-07-30", "gaps": "Specific 40TB Mozaic 4+ customer names not disclosed; margin contribution of HAMR mix not quantified.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "HAMR = ~40% of nearline EB run rate exiting FY2026; Mozaic 4+ = 50% of HAMR EB target by end-CY26", "metrics_note": "Mosley earnings-call quote via B&F; corroborated by SharperTrades 29 Jul 2026 piece (same 40% figure and Mozaic 4 50% target)" }, { "id": "STX-PER-EB-PRICE-PLUS-11PCT", "component": "Seagate HDD pricing: revenue per exabyte", "category": "nearline HDD", "subcategory": "pricing", "claim_type": "fact", "claim": "Seagate kept hard-drive unit production relatively flat while price per exabyte increased 11% YoY in fiscal Q4 2026; gross margin hit a record 52.3% (vs 46.5% prior quarter, 37.4% a year ago). Management: 'increasing the amount of data stored on every disk we produce is the fastest, most capital-efficient path to supporting long-term exabyte growth while maintaining relatively stable hard drive unit output.'", "confidence": "high", "status_trend": "price/EB +11% YoY; guidance implies further acceleration in Sep quarter", "period": "FQ4 2026 (quarter ended June 2026)", "source_url": "https://brief.sharpertrades.com/ai-storage-demand-drives-seagate-growth-as-hamr-boosts-margins/", "source_title": "AI Storage Demand Drives Seagate Growth as HAMR Boosts Margins (SharperTrades)", "source_date": "2026-07-29", "gaps": "No explicit spot-market price; LTA repricing schedule not disclosed. Sep-quarter price acceleration is an inference from guidance.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Drive unit output deliberately held flat; exabyte growth via areal density, not capex", "metrics_note": "11% figure and 52.3% gross margin from SharperTrades earnings summary 29 Jul 2026; 'stable unit output' quote from B&F 30 Jul 2026" }, { "id": "WDC-FQ4-2026-EB-231-CLOUD-89PCT", "component": "Western Digital FQ4 2026 shipments and cloud revenue", "category": "nearline HDD", "subcategory": "exabyte shipments / pricing", "claim_type": "supply_status", "claim": "Western Digital shipped 231 exabytes in FQ4 2026, up 22% YoY, and began shipping next-generation ePMR drives up to 40TB. Cloud end-market revenue was $3.3B, 89% of total, up 43% YoY; average price per terabyte increased from the high single digits to the high teens YoY. Q4 gross margin was 54.4% (+1,310 bps YoY) and revenue $3.75B (+44% YoY).", "confidence": "high", "status_trend": "record quarter; price/TB re-rating (HSD -> high teens %)", "period": "FQ4 2026 (quarter ended ~January 2026)", "source_url": "https://ca.finance.yahoo.com/news/western-digital-q4-earnings-beat-141500049.html", "source_title": "Western Digital Q4 Earnings Beat as Revenue Jumps 44% on Cloud Demand (Zacks via Yahoo Finance)", "source_date": "2026-08-06", "gaps": "Exabyte growth of 22% decelerated vs prior quarters (per management, mix/timing); nearline vs non-nearline split not disclosed for Q4.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Cloud = 89% of WDC revenue; 40TB ePMR volume ramp starting", "metrics_note": "Zacks report 6 Aug 2026; corroborated by GuruFocus earnings-call highlights (revenue $3.75B, EPS $3.56, cloud $3.3B)" }, { "id": "WDC-40TB-EPMR-ROADMAP-FY27", "component": "Western Digital 40TB ePMR / UltraSMR / HAMR roadmap", "category": "nearline HDD", "subcategory": "capacity roadmap", "claim_type": "product_status", "claim": "WDC targets 40TB ePMR drives to represent over 50% of nearline exabytes by Q3 of calendar 2027; UltraSMR ramping to ~60% of nearline exabyte shipments by FY2027 exit; 44TB HAMR products introduced in 1H calendar 2027. Management reiterated exabyte demand growth >25% going forward, with acceleration expected in H2 as the 40TB ePMR ramp intensifies. New LTAs covering 2029-2031 are in active commercial negotiation with strong volume visibility.", "confidence": "high", "status_trend": "roadmap locked through FY2027; LTA negotiations to 2029-2031", "period": "FQ4 2026 earnings call, forward view to CY2027-CY2031", "source_url": "https://www.gurufocus.com/news/9009786/western-digital-corp-wdc-q4-2026-earnings-call-highlights-record-revenue-and-eps-soar-on-aidriven-storage-demand", "source_title": "Western Digital Corp (WDC) Q4 2026 Earnings Call Highlights (GuruFocus)", "source_date": "2026-08-06", "gaps": "44TB HAMR customer qualifications not disclosed; LTA terms (pricing construct) not public; neocloud/physical-AI demand described qualitatively only.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Exabyte growth >25% guided forward; 40TB ePMR >50% of nearline EB by Q3 CY27", "metrics_note": "CEO Irving Tan and CFO Kris Sennesael remarks via GuruFocus transcript highlights 6 Aug 2026; nearline/non-nearline gross margins now in the same 'zip code' per CFO" }, { "id": "SKHYNIX-Q2-2026-RECORD-NAND-ESSD", "component": "SK hynix Q2 2026: record quarter, NAND pricing, LTAs", "category": "enterprise SSD / NAND", "subcategory": "quarterly results / LTA", "claim_type": "fact", "claim": "SK hynix posted record Q2 2026 results: revenue 79.3187 trillion won, operating profit 60.5426 trillion won (76% operating margin), net profit 93.9226 trillion won (118% margin). 'Both DRAM and NAND flash memory prices experienced significant quarter-over-quarter increases. SK hynix achieved top-tier profitability by expanding sales centered on high-value-added products, including HBM, DRAM for AI servers, and eSSD.' The company has finalized Long-Term Agreements with around 10 customers including key strategic partners, and is accelerating its 321-layer NAND transition (to ~50% of domestic production capacity by end-2026) plus the M15X NAND mass-production schedule and Yongin Phase 1 cleanroom opening in early 2027.", "confidence": "high", "status_trend": "record quarter; NAND price increases significant QoQ; ~10 LTAs signed", "period": "Q2 2026 (quarter ended June 2026), forward to early 2027", "source_url": "https://news.skhynix.com/en/q2-2026-business-results/", "source_title": "SK hynix Announces 2Q26 Financial Results", "source_date": "2026-07-29", "gaps": "eSSD/NAND segment split of the 79.3T won revenue not broken out in the release; net margin >100% implies one-off items (investments/gains) not itemized in the excerpt; LTA counterparties unnamed; M15X bit-share target and M17 timing not quantified.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "321-layer NAND to ~50% of domestic capacity by end-2026; Yongin P1 cleanroom early 2027 adds new capacity; H1 2026 cumulative revenue >100T won for first time", "metrics_note": "Primary IR release 29 Jul 2026; QoQ comparison basis: Q2 2025 revenue 22.232T won / operating profit 9.2129T won; 'around 10' LTA count may move as negotiations continue" }, { "id": "NAND-ESSD-CONTRACT-PRICE-2026", "component": "NAND flash and enterprise SSD contract/street prices, 2026 trajectory", "category": "NAND flash", "subcategory": "contract pricing / street pricing", "claim_type": "fact", "claim": "NAND flash contract prices rose 33-38% QoQ in Q1 2026 and accelerated to 70-75% QoQ in Q2 2026 (outpacing DRAM for the first time in this cycle). Enterprise SSD finished-drive contract prices rose ~80% QoQ in Q1 2026, with 48-53% QoQ forecast for Q2 2026 (TrendForce). For Q3 2026 TrendForce forecasts a comparatively modest 10-15% QoQ rise, while ADATA chairman Chen Li-bai says manufacturers have already announced 35-40% increases - a material divergence. Street pricing: a 30TB TLC enterprise SSD went from ~$3,000 in Q2 2025 to ~$17,500 by Q1 2026 (472%, Vdura). Hyperscalers/CSPs are securing the bulk of available NAND supply through long-term agreements, squeezing allocation for enterprise and mid-market buyers; TrendForce expects enterprise SSDs to be the largest single NAND application segment in 2026.", "confidence": "medium", "status_trend": "rising sharply Q1-Q2 2026; Q3 signal contested (10-15% vs 35-40%)", "period": "Q1-Q3 2026", "source_url": "https://www.servnetuk.com/insights/nand-price-surge-buy-flash-now-or-delay", "source_title": "Enterprise SSD Prices 2026: Buy Flash Now or Delay? (Servnet UK)", "source_date": "2026-07-09", "gaps": "TrendForce primary releases are paywalled; figures here are as cited by Servnet. Q3 range divergence means treat Q3 as uncertain; Vdura data point is a single-model street observation, not a contract index; no per-TB absolute price levels.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Supplier inventories at historic lows; production failing to keep pace with orders (per TrendForce-cited commentary)", "metrics_note": "Servnet 9 Jul 2026 citing TrendForce data 'across multiple trade sources' + ADATA chairman statement + Vdura street pricing; corroborated by TrendForce 2 Feb 2026 press release (1Q26 NAND +55-60% QoQ forecast, later revised upward)" }, { "id": "SNDK-FQ2-2026-REVENUE-3B-DATACENTER", "component": "SanDisk fiscal Q2 2026: NAND supercycle results", "category": "NAND flash", "subcategory": "quarterly results / shortage", "claim_type": "fact", "claim": "SanDisk fiscal Q2 2026: revenue $3.03B (up 61% YoY), non-GAAP EPS $6.20 (up 404% YoY), datacenter revenue $440M (up 76% YoY). Q3 FY2026 guide: revenue $4.4-4.8B with gross margin 65-67% vs 26% in the prior-year quarter. 2026 NAND bit demand growth is tracking 20-22% against bit supply growth of 15-17%; multiple sell-side desks model the shortage persisting until 2028. Contract NAND prices rose 85-90% QoQ in several recent quarters.", "confidence": "high", "status_trend": "revenue +61% YoY; gross margin 26% -> 65-67% in one year", "period": "FQ2 2026 (quarter ended ~April 2026)", "source_url": "https://www.techi.com/sandisk-sndk-stock-supercycle/", "source_title": "Sandisk (SNDK) Stock Up 30x in One Year: NAND Supercycle (TECHi)", "source_date": "2026-04-11", "gaps": "TECHi is an analysis piece, not the primary transcript - figures match the earnings release as summarized elsewhere; shortage-to-2028 is sell-side consensus, not company guidance; bit demand/supply growth rates not attributed to a specific desk.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "NAND bit supply growth (15-17%) lagging demand growth (20-22%) through 2026; shortage modeled to 2028", "metrics_note": "TECHi 11 Apr 2026 earnings analysis; '85-90% QoQ in several recent quarters' per same source - direction consistent with Servnet/TrendForce Q1-Q2 2026 data" }, { "id": "PSTG-FQ4-2026-FIRST-BILLION-QUARTER", "component": "Pure Storage fiscal Q4 2026: all-flash enterprise arrays", "category": "enterprise array (all-flash)", "subcategory": "quarterly results / hyperscale", "claim_type": "fact", "claim": "Pure Storage reported its first $1B quarter: fiscal Q4 2026 revenue $1.1B, up 20% YoY (full fiscal year $3.7B, +16% YoY); non-GAAP EPS 69 cents vs 45 cents a year earlier; non-GAAP operating margin 21.3% (vs 17.4%). Remaining performance obligations $3.7B, up 40% YoY. Q1 FY2027 guide: $990M-$1.01B, up ~28% YoY at midpoint. Everpure strengthened hyperscale positioning by partnering with SK hynix to deliver advanced QLC flash storage optimized for large data centers.", "confidence": "high", "status_trend": "record revenue; entering FY2027 with hyperscale momentum", "period": "FQ4 2026 (quarter ended 1 Feb 2026)", "source_url": "https://finance.yahoo.com/news/pstg-beats-q4-earnings-sales-144400932.html", "source_title": "PSTG Beats on Q4 Earnings & Sales, Enters FY27 With Hyperscale Momentum (Zacks via Yahoo Finance)", "source_date": "2026-02-26", "gaps": "SK hynix QLC partnership terms/quantum not disclosed; management 'remains wary of global supply chain imbalances' - flash cost pressure not quantified at product level.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "RPO $3.7B (+40% YoY) implies contracted hyperscale demand visibility; international revenue +48% YoY", "metrics_note": "Zacks report 26 Feb 2026 covering Pure Storage's Feb 2026 print; 'Everpure' is the hyperscale-focused division naming used in the source" }, { "id": "STORAGE-MOZAIC4PLUS-ASP-HYPERSCALE-2026", "component": "Seagate Mozaic 4+ 44TB — Hamr Drive ASP & Per-TB Economics at Hyperscale", "category": "Product Pricing", "claim_type": "pricing", "claim": "Seagate's HAMR-based Mozaic 4+ hard drives supporting capacities up to 44TB are shipping in volume to two leading hyperscale cloud providers since March 2026. A 44TB drive costs Seagate considerably less to build per terabyte than a 30TB PMR drive, yet it commands premium pricing from hyperscalers who value rack-scale density. CEO Dave Mosley confirmed the company achieves higher gross margins on higher-capacity HAMR products. While exact ASP figures are confidential under hyperscale NDA, industry analysis indicates the per-TB BOM cost advantage grows exponentially — moving from 3TB/disk (Mozaic 3+) to 4.4TB/disk (Mozaic 4+) reduces per-platter manufacturing overhead while adding only marginal incremental HAMR laser/engine complexity. This means the 44TB drive's ASP represents a better absolute dollar contribution per unit even if per-TB pricing pressure exists.", "confidence": "high", "status_trend": "Volume shipments to hyperscalers began Q1 CY2026; per-TB cost curve continues improving through Mozaic 5+ roadmap", "period": "2026–2028", "source_url": "https://www.techpowerup.com/351243/seagate-roadmap-targets-50-tb-hamr-hard-drives-in-2027", "source_title": "Seagate Roadmap Targets 50 TB HAMR Hard Drives in 2027 — TechPowerUp", "source_date": "2026-09-03", "gaps": "Exact ASP and per-TB contracted pricing are confidential between Seagate and hyperscalers. Published analyst estimates are unavailable because hyperscale deals are custom-quoted rather than listed. Second-hand market / spot pricing absent.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Two hyperscalers qualified for volume; five additional qualifications underway. Each Mozaic 4+ platform generation allows continued gains in capacity without disruptive architectural shifts — scaling path from 4+TB per disk toward 10TB per disk.", "metrics_note": "Source: TechPowerUp Sep 2026 reporting on Seagate Q4 FY2026 earnings presentation; corroborated by SEI press release confirming volume qualification. Hyperscale NDA precludes named-customer or ASP disclosure." }, { "id": "STORAGE-HAMR-MARGIN-DIVORCE-SEAGATE-FQ4-2026", "component": "Seagate FQ4 2026 Earnings — Record Gross Margin 52.3%+ on HAMR Ramp", "category": "Financial Performance", "claim_type": "earnings", "claim": "Seagate reported record gross margin of 52.3%+ in fiscal Q4 2026 (ending July 2026), up from prior years' low-40s percentages. Revenue reached $3.63 billion for the quarter (+48.5% YoY) and full-year revenue grew 34.1% to $12.20 billion. Non-GAAP EPS of $5.71 beat estimates by 12.11%. CEO Dave Mosley attributed gross margin expansion primarily to high-capacity product sales mix shift (Mozaic HAMR drives) combined with consistent pricing discipline and manufacturing cost efficiencies. Both HAMR and non-HAMR products contributed to margin gains, but the highest-margin segment is the top-end HAMR capacity tiers where hyperscalers pay premium absolute dollars per drive even as per-TB ASP softens slightly year-over-year due to density improvements reducing BOM/TB.", "confidence": "high", "status_trend": "Record margins sustained through Q4 FY2026; gross margin likely to expand further as Mozaic 4+ share of shipment base exceeds 50%", "period": "FQ4 2026 (CY2026)", "source_url": "https://247wallst.com/investing/2026/09/01/forget-nvidia-this-ai-stock-is-selling-the-storage-behind-the-boom/", "source_title": "Forget Nvidia. This AI Stock Is Selling the Storage Behind the Boom — 24/7 Wall St.", "source_date": "2026-09-01", "gaps": "Gross margin attributable specifically to HAMR-only vs PMR-only segments not separately broken out by Seagate. Absolute ASP differentials between Mozaic 3+, 4+, and legacy PMR lines not quantified publicly.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Stock up 385% over past year on HAMR revenue growth; hyperscaler orders booking through 2027. Full-year non-GAAP EPS doubled YoY to $3.18 billion net income.", "metrics_note": "Sources: 247 Wall Street Sep 2026 summarizing Seagate Q4 FY2026 results; Motley Fool Q2 2025 earnings call transcript noting gross margin expansion drivers." }, { "id": "STORAGE-CAPACITY-PATHWAY-MOZAIC-50TB-100TB-2028", "component": "Mozaic Capacity Roadmap — 50TB 2027→100TB 2029 Trajectory on HAMR", "category": "Product Roadmap", "claim_type": "product_status", "claim": "Seagate's published HAMR capacity roadmap: Mozaic 3+ delivered 3+TB/platter (30TB drives, 10 disks) in 2024–2025; Mozaic 4+ ships 4.4TB/platter (44TB drives, 10 disks) starting Q1 2026; by end of calendar 2026, 50% of Seagate's HAMR exabytes shipped will be on Mozaic 4 platform. 50TB HAMR drives are planned for customer validation in 2027 on the Mozaic 4 platform itself (not requiring new platform generation). Mozaic 5+ (third gen) targets 5+TB/platter for 60TB-class drives, with customer qualification late 2027. Lab demonstration target is 10TB/platter around 2028, which would enable ~100TB per-drive capacity at 10 disks. This represents an approximately 20x areal density increase from first-generation HAMR (~0.5TB/disk theoretical PMR ceiling) to 2028 targets.", "confidence": "high", "status_trend": "On-track per CEO and CFO statements at Q4 FY2026 earnings; no delays disclosed", "period": "2026–2028", "source_url": "https://www.techpowerup.com/351243/seagate-roadmap-targets-50-tb-hamr-hard-drives-in-2027", "source_title": "Seagate Roadmap Targets 50 TB HAMR Hard Drives in 2027 — TechPowerUp", "source_date": "2026-09-03", "gaps": "50TB-on-Mozaic-4 vs 50TB-on-Mozaic-5+ timeline distinction unclear from public disclosures. 100TB target is 'lab demonstration,' not engineering samples or qualified product. Yield rates at each step unknown. Consumer availability typically lags hyperscale by years.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "Each 1+TB/disk step increases platter count options and reduces per-drive assembly requirements proportionally. 10TB/platter × 10 disks = 100TB. Prototype platter demonstrated 6.9TB already.", "metrics_note": "Source: TechPowerUp Sep 2026 based on Seagate Q4 FY2026 earnings presentation and CEO Dave Mosley statements; CFO Gianluca Romano confirmed 5+ TB/platter achievable within current platform architecture." }, { "id": "STORAGE-NARLINE-SOLDOUT-LEADTIMES-2027-2028", "component": "Nearline HDD Production Sold Out Through 2027—Lead Times 52–104 Weeks", "category": "Supply Chain", "claim_type": "supply_status", "claim": "Production capacity for nearline HDDs from both Seagate and Western Digital is fully allocated through 2026 with orders extending into 2027 and 2028 due to AI hyperscaler demand consuming the majority of available capacity through multi-year supply contracts. Nearline HDD lead times exceed 52 weeks, with some enterprise model lead times reaching 104 weeks (two years). This supply crisis emerged because AI infrastructure buildout drove unanticipated demand expansion beyond surveillance/backup applications into training data retention, log archives, and cold storage tiers of modern data centers. HDD prices surged 46% on average across 12 tracked models from mid-September 2025 to January 2026, according to the ComputerBase price tracker — ranging from 23% to 66% depending on model. The largest quarterly contract increase was 4% QoQ in Q4 2025, the biggest in eight quarters.", "confidence": "high", "status_trend": "Sustained shortage through 2027 expected; cycle risk emerges 2028 if hyperscaler capex moderates", "period": "2026–2028", "source_url": "https://www.servnetuk.com/research/nearline-hdd-shortage-2026", "source_title": "2026 Nearline HDD Shortage: AI Reprices UK Cold Storage — Servnet UK", "source_date": "2026-07-15", "gaps": "Exact percentage of total nearline production consumed by hyperscalers vs enterprise customers not specified. Duration beyond 2028 uncertain — cyclical risk documented in Bear case analysis.", "leadtime_weeks": 52, "price_trend": "rising sharply", "capacity_signal": "Both Seagate and WD report utilization at or near full capacity across all product lines. China PC market also driving demand rebound after years of SSD substitution.", "metrics_note": "Sources: DigiTimes Dec 2025 on Q4 2025 contract pricing; ComputerBase Jan 2026 price tracker tracking 12 popular HDD models; Servnet UK comprehensive nearline shortage study Jul 2026 citing BENQUAN Global analysis." }, { "id": "STORAGE-WDC-ULTRASMR-EPMR-HAMR-ROADMAP-2027", "component": "Western Dual-Track HDD Strategy — 40TB ePMR Shipping, 44TB HAMR Targeted H1 2027", "category": "Competitive Positioning", "claim_type": "product_status", "claim": "Western Digital shipped its next-gen ePMR nearline platform at up to 40TB during Q2 FY2026 (quarter ended July 3, 2026), combining ePMR recording media with UltraSMR shingled track implementation. WD expects the 40TB ePMR/UltraSMR platform to account for >50% of nearline exabytes by Q3 CY2027, and UltraSMR across all capacities to reach ~60% of nearline exabyte shipments by end of FY2027. Concurrently, WD's 44TB HAMR drive (achieved via 11-disk stack at 4TB/platter) remains on track for first-half CY2027 shipment, having received positive customer feedback in qualification. Unlike Seagate's single-platform approach, WD maintains two parallel capacity lines through 2032: HAMR goes 40TB→44TB→60TB→100TB while ePMR goes 40TB→60TB by borrowing HAMR-derived media/firmware work and increasing to 14 platters while holding power flat. In FQ4 2026, WD shipped 231 EB total with 209 EB nearline, achieving 54.4% non-GAAP gross margin ($3.75B quarterly revenue, 89% cloud revenue).", "confidence": "high", "status_trend": "ePMR shipping now; HAMR qualification progressing positively; dual-track strategy gives flexibility but halves HAMR-specific investment pace vs Seagate", "period": "2026–2028", "source_url": "https://www.storagereview.com/news/wd-bets-8x-bandwidth-beats-more-terabytes-as-40tb-ultrasmr-starts-shipping", "source_title": "WD Bets 8x Bandwidth Beats More Terabytes as 40TB UltraSMR Starts Shipping — StorageReview", "source_date": "2026-08-09", "gaps": "WD's extra-disk HAMR design (11-disk vs Seagate's 10-disk) adds mechanical complexity and power envelope considerations. ePMR UltraSMR write performance characteristics not disclosed. Customer names and ASP for both platforms confidential.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "209 EB nearline shipped in single quarter (Q2 FY2026). Hyper-concentration on cloud (>89% revenue) creates hyperscaler dependency risk but secures committed volumes.", "metrics_note": "Source: StorageReview Aug 2026 synthesizing WD Q4 FY2026 earnings call (Aug 5, 2026) technical presentations and management commentary on capacity roadmap." }, { "id": "STORAGE-SSD-HDD-COST-RATIO-BIFURCATION-2026", "component": "Enterprise SSD vs HDD Cost Ratio Expands to 16× in 2026", "category": "Pricing Trends", "claim_type": "trend", "claim": "The cost ratio between enterprise SSD and HDD has widened dramatically in 2026: NAND flash diversion to HBM production pushed NVMe SSD pricing to approximately 16× the per-TB cost of HDD, up from roughly 6× in 2025. Enterprise SSDs priced at $800–$1,200 for 3.84TB units are now effectively ~$200–$300/TB when normalized, versus HDDs at $100–$150/TB for high-capacity nearline models in hyperscale procurement. This extreme bifurcation reinforces tiered storage architectures where NVMe serves hot/training data and HDD serves cold/archive layers. However, SSD contract prices surged 500% over six months as Phison requested customer prepayments for NAND allocation, making the gap unsustainable and potentially creating a demand reversion to HDD-backed capacity if SSD economics tighten sufficiently.", "confidence": "medium", "status_trend": "Cost ratio diverging sharply; however, massive SSD price inflation may eventually push some workloads back toward HDD if cost differential becomes economically untenable", "period": "2026–2027", "source_url": "https://benquan.hk/article-enterprise-ssd-hdd-supply-crisis-2026.html", "source_title": "Enterprise SSD & HDD Supply Crisis 2026 Guide — BENQUAN Global", "source_date": "2026-08-06", "gaps": "Exact per-TB pricing varies by capacity tier, interface protocol, and contractual volume. The 16× ratio is an extrapolation from known SSD ASPs and HDD nearline ASP ranges. NAND-to-HBM wafer reallocation percentage not precisely quantified.", "leadtime_weeks": null, "price_trend": "HDD rising (supply constrained); SSD rising much faster (NAND/HBM competition)", "capacity_signal": "NAND flash diverted to HBM production; HDD manufacturers at full capacity. Both storage tiers face supply constraints simultaneously from different root causes.", "metrics_note": "Source: BENQUAN Global Aug 2026 supply chain analysis; TechPowerUP Jun 2026 reporting on Phison NAND prepayment demands; DigiTimes on Q4 2025 pricing." }, { "id": "STORAGE-HAMR-RAMP-MOZAIC-SHARE-EXABYTES-2026", "component": "HAMR Exabyte Share Trajectory — 40% Nearline Exabytes on HAMR by H2 2026", "category": "Market Dynamics", "claim_type": "market_share", "claim": "As of Q2 FY2026, Seagate reports that approximately 40% of its nearline exabytes shipped to customers are on HAMR-based platforms (Mozaic 3+ and early Mozaic 4). CEO Dave Mosley stated explicit target to achieve 50% of HAMR exabytes on the Mozaic 4 platform by end of calendar 2026. Since Seagate's flagship-tier enterprise drives use 10 platters per drive, the 5TB/platter milestone confirmed by CFO Gianluca Romano implies 50TB per-drive capacity achievable on existing platform architecture without introducing new recording physics. Seagate has shipped over one million HAMR units to date (cumulative across Mozaic 3+ and Mozaic 4). The broader capacity roadmap scales from 40-44TB in the 2027-2028 window, to 50-60TB around 2030, and past 80TB in the 2031-2032 timeframe.", "confidence": "high", "status_trend": "HAMR ramp accelerating: 40% exabyte share already achieved; 50% Mozaic-4-share target on track for end of CY2026", "period": "2026–2032", "source_url": "https://medium.com/@noahbean3396/the-laser-war-inside-your-hard-drive-how-ai-demand-forced-seagate-and-western-digital-to-solve-the-6885e3e6fdd4", "source_title": "The Laser War Inside Your Hard Drive — Medium (Noah Bean)", "source_date": "2026-09-05", "gaps": "HAMR yield rates not publicly disclosed — initial yields were reportedly lower than PMR but improving rapidly. Reliability metrics (MTBF, UW/DFE) at HAMR scale not independently verified. Total TAM for HAMR vs PMR in hyperscale deployments difficult to quantify given confidentiality.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "Over 1M HAMR units shipped cumulative. 5 hyperscaler+ qualifications complete, additional ongoing. Each successive capacity point requires only incremental densification rather than platform overhaul.", "metrics_note": "Source: Medium article Sep 2026 synthesizing Seagate Q4 FY2026 IR materials, CEO/CFO comments, and historical HAMR shipment data from Mozaic 3+ launch through present." }, { "id": "NRL-WDC-STX-DOMINANCE-95PCT-SHARE", "component": "Seagate/Western Digital/Toshiba nearline market share triopoly", "category": "nearline HDD", "subcategory": "competitive positioning", "claim_type": "market_concentration", "claim": "Western Digital, Seagate, and Toshiba collectively hold over 95% of global HDD shipment revenue with nearline drives representing approximately 90% of total industry exabytes shipped. Non-nearline client and consumer segments have structurally declined to low-teens percentage of total exabytes. WD is the volume leader in nearline eB while Seagate holds leadership in areal density via HAMR. No meaningful third or fourth player threatens this oligopoly. Toshiba is more than a full recording-generation behind both rivals on advanced media.", "confidence": "high", "status_trend": "concentrated oligopoly with deep hyperscaler lock-in", "period": "2026", "source_url": "https://www.storagereview.com/news/wd-bets-8x-bandwidth-beats-more-terabytes-as-40tb-ultrasmr-starts-shipping", "source_title": "WD Bets 8x Bandwidth Beats More Terabytes as 40TB UltraSMR Starts Shipping (StorageReview)", "source_date": "2026-08-05", "gaps": "Exact per-company revenue-share percentages not publicly disclosed. Exabyte volume split between WD and STX available only from separate earnings releases (WD: 209/231 EB nearline in FQ4; Seagate: 195/218 EB nearline in FQ4).", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Both WD CEO Irving Tan Feb 2026: pretty much sold out for calendar 2026 and Seagate CEO Dave Mosley confirmed identical capacity allocation position. Consumer HDD represents approximately 5% of WD output with rest absorbed by hyperscale CSPs.", "metrics_note": "90% nearline figure from The Information Network analysis via 24/7 Wall St.; 95%+ revenue-share from multiple market reports (Mordor Intelligence, DataInsightsMarket). Cross-checked against StorageReview competitive comparison Aug 2026." }, { "id": "NRL-WDC-VS-STX-COMPETITIVE-DYNAMICS", "component": "WD vs Seagate strategic divergence execution versus density", "category": "nearline HDD", "subcategory": "competitive positioning", "claim_type": "competitive_strategy", "claim": "WD monetizes scale and operational execution while Seagate monetizes technology progression and density leadership. Both companies guided Q1 FY2027 July quarter revenue to exactly $4.1B plus or minus $100M with data center share at approximately 90%. Seagate grew total exabytes 34% YoY vs WD 22%. Seagate lifted DC exabytes 11% sequentially where WD managed 5%. This decouples revenue from unit volatility. Key question is which company converts capacity growth into pricing power and margin expansion faster.", "confidence": "high", "status_trend": "parallel growth with technology differentiators diverging", "period": "FQ4 2026 forward Q1 FY2027", "source_url": "https://247wallst.com/investing/2026/05/16/seagate-and-western-digital-ai-storage-demand-is-now-showing-up-in-pricing-power/", "source_title": "Seagate and Western Digital AI Storage Demand Showing Up In Pricing Power (24/7 Wall St.)", "source_date": "2026-05-16", "gaps": "No direct ASP comparisons published. Neither company breaks out per-drive pricing in earnings disclosures.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Seagate 34% YoY eB growth with +11% sequential. WD 22% YoY with +5% sequential. Both targeting identical next-quarter revenue.", "metrics_note": "Revenue guidance coincidence and exabyte growth rates from StorageReview Sep 2026 analysis cross-referenced against Zacks/GuruFocus Q4 earnings data. Strategic framing from 24/7 Wall St. 16 May 2026 Dr. Robert Castellano analysis." }, { "id": "NRL-STX-HAMR-MOZAIC-LEADERSHIP", "component": "Seagate HAMR Mozaic density leadership and Mozaic 4 ramp", "category": "nearline HDD", "subcategory": "technology roadmap", "claim_type": "product_status", "claim": "Seagate exited fiscal 2026 with HAMR-based products at approximately 40% of its nearline eB shipment run rate a milestone confirmed by CFO Gianluca Romano in June 2026. Mozaic 4 platform at 4 TB/platter is in volume production with the two largest global CSPs. Seagate expects 50% of HAMR eB on Mozaic 4 by end-CY2026. Quality and sample shipments start late CY2027 for Mozaic 5+. 44TB Mozaic 4+ rated approximately 300 MB/s sustained versus 250-275 MB/s typical. Seagate runs 10 disks at approximately 4.4 TB per disk while WD achieves equivalent 44TB with 11 disks at approximately 4 TB per disk.", "confidence": "high", "status_trend": "HAMR ramping aggressively 40% achieved ahead of schedule", "period": "FQ4 2026 June 2026 forward CY2026 target", "source_url": "https://www.blocksandfiles.com/disk/2026/07/30/seagate-hamring-customers-generating-more-and-more-/tb/5281058", "source_title": "Seagate HAMRing customers generating more and more $/TB (Blocks and Files)", "source_date": "2026-07-30", "gaps": "Specific Mozaic customer names undisclosed. Yield rates on 44TB production unknown. Margin contribution of HAMR versus ePMR mix not quantified beyond overall 52.3% gross margin.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "HAMR equals approximately 40% of nearline eB run rate exiting FY2026. 50% HAMR-on-Mozaic-4 target by end-CY26. 44TB volume production. Planning horizon extended to CY2029 plus.", "metrics_note": "Blocks and Files 30 Jul 2026 earnings coverage corroborated by SharperTrades 29 Jul 2026 same 40% figure and Mozaic 4 50% target. CFO confirmation Jun 2026 via StorageReview." }, { "id": "NRL-WDC-EPMR-ULTRASM-RAMP-TWO-LINE", "component": "Western Digital dual-roadmap strategy ePMR-UltraSMR to 60TB plus HAMR to 44TB to 100TB", "category": "nearline HDD", "subcategory": "technology roadmap", "claim_type": "product_status", "claim": "WD published roadmap runs two capacity lines in parallel through 2032. ePMR line 40TB 2026 shipping 60TB 2028. WD claims ability to reach 60TB without HAMR by borrowing HAMR-derived media and firmware work moving to 14 platters while holding power flat. HAMR line 40TB 2026 44TB 2027 first-half CA cycle 60TB 2028 100TB 2029 dotted continuation past that requiring own laser technology 10 TB per platter areal density and 14-disk stack. UltraSMR across all capacities targeting approximately 60% of nearline eB shipments by end-FY2027. High Bandwidth Drive HBD sampling with five customers demonstrated 2x throughput no power increase targeting 8x by 2030 would clear 500 MB/s plus potentially pushing single-spindle reads past NVMe territory for sequential work.", "confidence": "medium", "status_trend": "roadmap locked through 2032 aggressive but unproven on HBD and DualPivot", "period": "FQ4 2026 Aug 5 call forward to 2032", "source_url": "https://www.storagereview.com/news/wd-bets-8x-bandwidth-beats-more-terabytes-as-40tb-ultrasmr-starts-shipping", "source_title": "WD Bets 8x Bandwidth Beats More Terabytes as 40TB UltraSMR Starts Shipping (StorageReview)", "source_date": "2026-08-05", "gaps": "HAMR share of nearline eB undisclosed creating asymmetrical comparison to Seagate disclosure. HBD at sampling with five customers stage. Dual Pivot is lab-only at 2028 target. 100TB requires 7.1 TB per platter effective density which may face yield reality.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "40TB ePMR volume at customer. 44TB HAMR H1 CY27 on track. LTA negotiations covering 2029 to 2031 in active progress. Exabyte growth greater than 25% guided forward.", "metrics_note": "Comprehensive roadmap data from StorageReview 5 Aug 2026 earnings coverage referencing WD Innovation Day and Q4 earnings call remarks by CEO Irving Tan and CFO Kris Sennesael." }, { "id": "NRL-TOSHIBA-FC-MAMR-LAG-BEHIND", "component": "Toshiba nearline position FC-MAMR sampling at 30-34TB one generation behind", "category": "nearline HDD", "subcategory": "technology status", "claim_type": "product_status", "claim": "Toshiba began sampling highest-capacity nearline drives at 30-34TB in late March 2026 using FC-MAMR with shingled SMR recording not HAMR. Its CMR models top out at 28TB with sample shipments slated for Q3 2026. The 30-34TB M12 Series uses 11 disks up from previous 10 leveraging Toshiba proprietary design and analysis technologies. Toshiba plans 12-disk stacking with MAMR enabling 40TB-class drives by 2027 also investigating 12-disk stacking combined with HAMR for even higher capacity. However Toshiba is a full recording generation behind both Seagate 44TB HAMR in production and WD 40TB ePMR shipping with 44TB HAMR on track.", "confidence": "medium", "status_trend": "catching up sampling earlier generations while competitors ship later ones", "period": "CY2026 sampling forward CY2027 target for 40TB", "source_url": "https://toshiba.semicon-storage.com/ap-en/company/news/news-topics/2026/03/storage-20260331-1.html", "source_title": "Toshiba Begins Sampling of 30-34 TB SMR Nearline Hard Disk Drives (Toshiba News Release)", "source_date": "2026-03-31", "gaps": "Toshiba does not publish quarterly eB shipment volumes or revenue breakdown by product category. No public hyperscaler LTA announcements. Exact TAM and shipping-market-share figures unavailable. M12 Series goes to sample not volume production yet.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "FC-MAMR with SMR sampling at 30-34TB. CMR capped at 28TB. 12-disk MAMR targeting 40TB by 2027. HAMR planned upcoming quarters. No disclosed eB volume or hyperscaler commitments.", "metrics_note": "Toshiba official press release 31 Mar 2026. CDRInfo Toshiba Storage Trends 2026 citing Gartner noting HDDs will be 68% of total shipped capacity by 2026. StorageReview Aug 2026 confirming Toshiba is not close relative to WD and STX." }, { "id": "NRL-HDD-CLOUD-ALLOCATION-PRICING", "component": "HDD hyperscaler allocation and pricing surge 35-46% YoY", "category": "nearline HDD", "subcategory": "pricing and allocation", "claim_type": "market_dynamics", "claim": "Enterprise nearline HDD capacity is fully absorbed by hyperscale cloud providers through multi-year LTAs. WD confirmed Feb 2026 sold out for calendar 2026 and Seagate confirmed the same. Consumer drives dropped to approximately 5% of WD output. HDD prices rose 35-46% YoY in 2026. NVMe $/TB ratio stands at approximately 16x that of nearline HDD up from 6x twelve months ago meaning the SSD-to-HDD cost convergence trajectory has reversed. Tiered architecture standard NVMe for hot data 5-10% HDD for nearline warm and cold 60-70% LTO tape for archive.", "confidence": "high", "status_trend": "all production absorbed pricing strengthening materially", "period": "2026", "source_url": "https://benquan.hk/article-enterprise-ssd-hdd-supply-crisis-2026.html", "source_title": "Enterprise SSD and HDD Supply Crisis 2026 Guide (BenQuan Global)", "source_date": "2026-08-05", "gaps": "LTA term lengths and specific pricing constructs not disclosed by any vendor. True spot-market pricing absent given all supply covered by contracts. Regional pricing delta based on BenQuan distributor-level estimates.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "All 2026 nearline capacity allocated. NVMe $/TB ratio 16x from 6x YoY. Consumer HDD approximately 5% of WD output. HDD price rise 35-46% YoY.", "metrics_note": "BenQuan 2026 Enterprise Storage Price Index cross-referenced against DatacenterDisk/Emergent Storage Price Index Q1 2026. HDD/CSP allocation confirmed by both WD and Seagate executives." }, { "id": "NRL-KIOXIA-NAND-RECOVERY-record-FY26", "component": "Kioxia NAND recovery record FY26 results post-bankruptcy restructuring", "category": "NAND flash", "subcategory": "recovery and financial performance", "claim_type": "financial_data_point", "claim": "Kioxia posted record consolidated results for FY2026 ended March 31 2026: revenue ¥2.34 trillion +37% YoY net profit ¥554.5 billion plus more than doubled operating margin 37.2% from 26.5% ROE 51.9%. Non-GAAP operating profit ¥876.2 billion +93.4%. Kioxia 2026 NAND flash capacity is fully booked. Kioxia guides Q1 FY2027 April-June 2026 revenue acceleration. Post-restructuring Kioxia cancelled ¥447.5B syndicated loan in April 2026 completed Nanya Technology acquisition worth TWD 15.7B and strengthened balance sheet with equity rising from 25.3% to 37.9% of assets.", "confidence": "medium", "status_trend": "post-restructuring recovery accelerated into record profitability", "period": "FY2026 Apr 2025-Mar 2026", "source_url": "https://japanstockpulse.com/article/2026/05/15/285A-1.html", "source_title": "Kioxia FY26 Profit Doubles as AI Memory Demand Surges (JapanStockPulse)", "source_date": "2026-05-15", "gaps": "NAND segment revenue breakdown not provided in detail. SSD and Storage business led surge but specific yen amounts missing. Interplay between Kioxia NAND supply and HDD ecosystem indirect. Kioxia HDD business was spun off to KioXia Holdings separately.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "2026 NAND capacity fully booked. Capex ¥281.1B. Operating cash flow ¥616.5B. Year-end cash ¥470.7B.", "metrics_note": "JapanStockPulse 15 May 2026 citing Kioxia earnings short report. Sandisk stock reaction via cryptobriefing confirms market perception." } ] }, { "key": "network", "name": "Networking", "color": "#DC2626", "description": "Switches, ASICs, retimers, PHYs, interconnects, optical", "summary": "800G transition accelerating with Broadcom Tomahawk 6. Retimer/PHY shortage constraining deployment. Credo and Lumentum benefiting from AI networking demand. NVLink 6.0 targeting 3.6 TB/s per GPU.", "entries": [ { "id": "NET-001", "component": "Broadcom Tomahawk 5 ASIC", "category": "Switch ASIC", "claim_type": "product_status", "claim": "Tomahawk 5: 5nm process, dominates AI ODM chassis from Celestica/Quanta/Inventec. First-generation high-speed switch chip for AI training clusters.", "confidence": "high", "status_trend": "active_production", "period": "2025–2027", "source_url": "https://www.fs.com/blog/broadcom-tomahawk-5-vs-tomahawk-6-vs-tomahawk-ultra-whats-changed-b44085.html", "source_title": "Broadcom Tomahawk 5 vs. Tomahawk 6 vs. Tomahawk Ultra", "source_date": "2025-12-29", "speed_class": "200G/400G", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-002", "component": "Broadcom Tomahawk 6 ASIC", "category": "Switch ASIC", "claim_type": "roadmap", "claim": "Tomahawk 6: 5nm, 64×800G ports = ~51.2 Tbps switch fabric. Adds TSN support for AI cluster scheduling. Volume expected 2027–2028 only.", "confidence": "medium", "status_trend": "sample_early", "period": "2026–2028", "source_url": "https://www.mordorintelligence.com/industry-reports/enterprise-network-switch-market", "source_title": "Enterprise Network Equipment Market Size & Share Analysis", "source_date": "2026-03-06", "speed_class": "800G", "gaps": "No confirmed first customers publicly", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-003", "component": "Broadcom Tomahawk Ultra", "category": "Switch ASIC", "claim_type": "roadmap", "claim": "Roadmap-only; >30% bandwidth uplift over TH6 cited. No customer programs confirmed.", "confidence": "low", "status_trend": "planning_only", "period": "2028+", "source_url": "https://www.fs.com/blog/broadcom-tomahawk-5-vs-tomahawk-6-vs-tomahawk-ultra-whats-changed-b44085.html", "source_title": "Broadcom Tomahawk 5 vs. Tomahawk 6 vs. Tomahawk Ultra", "source_date": "2025-12-29", "speed_class": "unknown", "gaps": "No customer programs confirmed", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-004", "component": "NVIDIA Spectrum-4", "category": "Switch ASIC", "claim_type": "product_status", "claim": "Pluggable-optics architecture. 51.2 Tb/s per-chip. Used in Blackwell-era NVIDIA networking builds. Legacy product now, succeeded by Spectrum-6 CPO line.", "confidence": "high", "status_trend": "active_production", "period": "2025–2026", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "200G/400G", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-005", "component": "NVIDIA Spectrum-6 / SN6800 CPO", "category": "Switch ASIC + CPO", "claim_type": "product_status", "claim": "World's first volume-produced 200G/lane CPO Ethernet switch system. 102.4 Tb/s per-chip (2× Spectrum-4). SN6800: 512×800G or 2,048×200G, 409.6 Tb/s total fabric. Mass production Aug 2026. Customers: CoreWeave, Lambda, Oracle.", "confidence": "high", "status_trend": "full_mass_production", "period": "2026–2028", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "200G/800G", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-006", "component": "NVIDIA Spectrum-XGS Ethernet", "category": "Ethernet", "claim_type": "product_status", "claim": "Spectrum-XGS: 1.9× throughput improvement over Spectrum-4. 800G per port. Targeted at AI inference and distributed training.", "confidence": "high", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "800G", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-007", "component": "InfiniBand NDR", "category": "InfiniBand Switch", "claim_type": "product_status", "claim": "InfiniBand NDR (400G) in active production; HDR (200G) legacy. Used in NVIDIA DGX SuperPOD.", "confidence": "high", "status_trend": "active_production", "period": "2025–2027", "source_url": "https://www.mordorintelligence.com/industry-reports/enterprise-network-switch-market", "source_title": "Enterprise Network Equipment Market Size & Share Analysis", "source_date": "2026-03-06", "speed_class": "200G/400G", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-008", "component": "InfiniBand XDR", "category": "InfiniBand Switch", "claim_type": "roadmap", "claim": "InfiniBand XDR (320G) transitioning to production; HDR (160G) legacy.", "confidence": "medium", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://www.mordorintelligence.com/industry-reports/enterprise-network-switch-market", "source_title": "Enterprise Network Equipment Market Size & Share Analysis", "source_date": "2026-03-06", "speed_class": "320G", "gaps": "Production timeline uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-009", "component": "InfiniBand XDR1600", "category": "InfiniBand Switch", "claim_type": "roadmap", "claim": "InfiniBand XDR1600 (1600G) targeting 2028; 8× XDR bandwidth.", "confidence": "medium", "status_trend": "sample_early", "period": "2028 (est.)", "source_url": "https://www.mordorintelligence.com/industry-reports/enterprise-network-switch-market", "source_title": "Enterprise Network Equipment Market Size & Share Analysis", "source_date": "2026-03-06", "speed_class": "1600G", "gaps": "No confirmed first customers", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-010", "component": "800G optical transceivers", "category": "Transceivers/Optics", "claim_type": "product_status", "claim": "800G optical transceivers (OSFP/QSFP112) entering volume production; key for Tomahawk 6 and Spectrum-6 deployments.", "confidence": "high", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "800G", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-011", "component": "1.6T optical transceivers", "category": "Transceivers/Optics", "claim_type": "roadmap", "claim": "1.6T optical transceivers targeting 2027–2028; required for Tomahawk Ultra and next-gen switches.", "confidence": "medium", "status_trend": "sample_early", "period": "2027–2028", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "1.6T", "gaps": "Production timeline uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-012", "component": "Coherent optical engines", "category": "Transceivers/Optics", "claim_type": "product_status", "claim": "Coherent optical engines for 800G+ transceivers constrained; Cisco/Acacia and Infinera reporting extended lead times.", "confidence": "high", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "800G", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-013", "component": "Silicon Photonics (SiPh)", "category": "Switch ASIC Extension", "claim_type": "roadmap", "claim": "SiPh engines for CPO targeting 2027; TSMC and GlobalFoundries developing SiPh processes.", "confidence": "medium", "status_trend": "sample_early", "period": "2027 (est.)", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "SiPh maturity and yield", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-014", "component": "CPO manufacturing", "category": "SiPh Manufacturing", "claim_type": "product_status", "claim": "CPO manufacturing requires TSMC advanced packaging + Foxconn assembly; supply chain nascent.", "confidence": "high", "status_trend": "new_launch", "period": "2026–2028", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "Manufacturing scale-up timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-015", "component": "DAC/AOC copper cables", "category": "Fiber Connectivity", "claim_type": "product_status", "claim": "DAC/AOC copper cables for GPU-rack interconnect lead time 15-25 weeks; constrained by TE Connectivity and Amphenol.", "confidence": "medium", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "speed_class": "unknown", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-016", "component": "FOCI FAU (Fiber Optic Connector Interface)", "category": "Packaging/Testing", "claim_type": "upstream_constraint", "claim": "FOCI FAU connectors for optical interconnects constrained; limited supplier base.", "confidence": "medium", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "Supplier base and capacity", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-017", "component": "TSMC SiPh", "category": "SiPh Manufacturing", "claim_type": "product_status", "claim": "TSMC SiPh process node targeting 2027; enables integrated optical engines for CPO.", "confidence": "medium", "status_trend": "sample_early", "period": "2027 (est.)", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "Process maturity and yield", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-018", "component": "SPIL/ASE packaging", "category": "Packaging/Testing", "claim_type": "product_status", "claim": "SPIL and ASE packaging for CPO optical engines; supply chain nascent.", "confidence": "medium", "status_trend": "new_launch", "period": "2026–2028", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "Packaging scale-up timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-019", "component": "Foxconn assembly", "category": "System Assembly", "claim_type": "product_status", "claim": "Foxconn assembling CPO switch systems for NVIDIA; first volume production Aug 2026.", "confidence": "high", "status_trend": "full_mass_production", "period": "2026–2028", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "Assembly capacity and yield", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-020", "component": "Retimer/PHY chips", "category": "Retimer/PHY", "claim_type": "product_status", "claim": "Retimer/PHY chips for 800G+ interconnects constrained; Astera Labs and Credo reporting extended lead times.", "confidence": "high", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "speed_class": "800G", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-021", "component": "DPU/NIC supply", "category": "DPU/NIC", "claim_type": "product_status", "claim": "NVIDIA BlueField DPUs and Broadcom Tomahawk-based NICs constrained; lead times 20-30 weeks.", "confidence": "high", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "speed_class": "400G", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-022", "component": "NIC demand growth", "category": "NIC", "claim_type": "demand", "claim": "NIC demand growing 50%+ YoY driven by AI cluster networking requirements.", "confidence": "medium", "status_trend": "expanding_rapidly", "period": "2026", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "speed_class": "unknown", "gaps": "Exact demand figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-023", "component": "GPU interconnect copper", "category": "GPU Interconnect/Copper", "claim_type": "upstream_constraint", "claim": "Copper interconnect for GPU-rack (NVLink) lead time 20-30 weeks; constrained by TE Connectivity and Amphenol.", "confidence": "medium", "status_trend": "rapidly increasing", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "speed_class": "unknown", "gaps": "Alternative interconnect technologies timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-024", "component": "Fiber optic cable", "category": "Fiber Connectivity", "claim_type": "product_status", "claim": "Single-mode fiber cable for data center interconnect lead time 10-15 weeks; adequate supply.", "confidence": "high", "status_trend": "active_production", "period": "2026", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-025", "component": "Upstream laser sources", "category": "Upstream Lasers", "claim_type": "upstream_constraint", "claim": "Upstream laser sources for coherent optics constrained; II-VI (Coherent) and Lumentum reporting extended lead times.", "confidence": "medium", "status_trend": "expanding_rapidly", "period": "2026–2027", "source_url": "https://finance.biggo.com/news/f9e53c4a-7058-4272-a305-e8dc8a8f416b", "source_title": "Nvidia's CPO Switches Enter Full Mass Production; TSMC, Foxconn Lead Supply Chain Positioning", "source_date": "2026-08-14", "speed_class": "unknown", "gaps": "Exact lead times vary by vendor", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "NET-BRCM-TH6-MASSPROD-MAR26", "component": "Broadcom Tomahawk 6 ASIC", "category": "Switch ASIC", "subcategory": "Ethernet Switch", "claim_type": "production_milestone", "claim": "Tomahawk 6 — world's first 102.4 Tbps Ethernet switch chip — began volume production in March 2026, less than 3 quarters from initial sampling. Targets million-XPU AI clusters. First customer deployments with hyperscalers for scale-out Clos fabric.", "confidence": "high", "status_trend": "mass_production_just_began", "period": "2026–2028", "source_url": "https://investors.broadcom.com/news-releases/news-release-details/broadcom-now-shipping-worlds-first-1024-tbps-switch-production", "source_title": "Broadcom Now Shipping World's First 102.4 Tbps Switch in Production", "source_date": "2026-03-12", "speed_class": "800G", "gaps": "", "leadtime_weeks": 20, "price_trend": "rising", "capacity_signal": "First silicon shipped <3 quarters after sampling — unprecedented ramp speed; hypervolume orders indicate multi-billion-dollar platform lifecycle", "metrics_note": "Source: Broadcom investor relations press release (Mar 12, 2026); confirmed by multiple industry analyst reports" }, { "id": "NET-BRCM-TH7-TAPEOUT-AUG26", "component": "Broadcom Tomahawk 7 ASIC", "category": "Switch ASIC", "subcategory": "Ethernet Switch", "claim_type": "roadmap_milestone", "claim": "Tomahawk 7 taped out August 2026 as industry's first 200 Tbps Ethernet switch for 'scale-in' networking. Announced during Q3 FY2026 earnings call. Moves networking from two-tier Clos into flattened single-stage fabric, reducing hop count and latency for large-scale AI training.", "confidence": "medium", "status_trend": "taped_out", "period": "2027–2029 (est.)", "source_url": "https://finance.yahoo.com/quote/AVGO/earnings/AVGO-Q3-2026-earnings_call-685349.html/", "source_title": "Broadcom Inc (AVGO) Q3 FY2026 earnings call transcript", "source_date": "2026-09-02", "speed_class": "1.6T", "gaps": "Volume production timeline unconfirmed; likely 2028+ given current fab allocation", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Tape-out announced at Q3 FY2026 call — follows TH6 (<3q sample-to-production) suggesting possible 2027 proto samples, but volume likely 2028 given 200 Tbps complexity leap", "metrics_note": "Earnings call Q3 FY2026 (Sep 2, 2026); Yahoo Finance transcript of Matt Murphy presentation" }, { "id": "NET-CREDO-Q1FY27-RETIMER-479M", "component": "Credo Screaming Eagle / Blue Heron Retimers", "category": "Retimer/PHY", "claim_type": "revenue_record", "claim": "Credo reported record Q1 FY2027 revenue of $479 million (+115% YoY, seventh consecutive quarter of triple-digit growth). Screaming Eagle retimer deployed at 100G/lane across AI scale-up deployments; Blue Heron product contributing at 200G/lane. Retimer revenue reached all-time high within this mix.", "confidence": "high", "status_trend": "record_revenue_quarter", "period": "Q1 FY2027 (ended ~Jul 2026)", "source_url": "https://ca.finance.yahoo.com/news/credo-technology-group-holding-ltd-050030303.html", "source_title": "Credo Technology Group Holding Ltd (CRDO) Q1 2027 Earnings Report", "source_date": "2026-09-01", "speed_class": "100G/200G per lane", "gaps": "Credo does not break out retimer vs AEC vs gearbox revenue separately in public filings; retimer is a component of the $479M total", "leadtime_weeks": 16, "price_trend": "stable", "capacity_signal": "$479M total company revenue; triple-digit growth sustained across 7 quarters indicates broad-based demand across all connectivity products including retimers", "metrics_note": "Yahoo Finance/Q1 FY2027 report (Sep 1, 2026); Converge Digest coverage" }, { "id": "NET-CREDO-AEC-GROWTH-FY27", "component": "Credo HiWire Active Ethernet Cables (AEC)", "category": "Active Cable/Interconnect", "claim_type": "demand_growth", "claim": "Credo's HiWire AEC line continues strong growth driven by hyperscale intra-chassis interconnect demand. 1.6T AEC products under development for next-gen rack topologies. Optical DSP also contributing materially — company targeting more than $600M of optical revenue across three product lines in fiscal 2027.", "confidence": "high", "status_trend": "rapidly_expanding", "period": "FY2027+", "source_url": "https://convergedigest.com/credo-q1-fy2027-ai-optics-aec-growth/", "source_title": "Credo Revenue Jumps 115% as Optics Become Major Growth Driver — Converge Digest", "source_date": "2026-09-01", "speed_class": "800G/1.6T", "gaps": "No separate AEC revenue disclosure in SEC filings; optical DSP run-rate estimated at >$200M for FY2027", "leadtime_weeks": 12, "price_trend": "falling", "capacity_signal": "$600M+ optical revenue target for FY2027 across AEC, DSP, and optical engine product lines; suggests optical segment alone approaching $300M+ annually", "metrics_note": "Converge Digest analysis of Q1 FY2027 results; management commentary on forward guidance" }, { "id": "NET-LUMENTUM-Q4FY26-BLOWOUT", "component": "Lumentum AI Optical Transceiver Lasers", "category": "Optical Components/Lasers", "claim_type": "revenue_record", "claim": "Lumentum Q4 FY2026 revenue $1.01 billion (+109% YoY), non-GAAP operating margin 36.6%, gross margin crossed 50% for first time. Q1 FY2027 guided $1.225B–$1.275B (midpoint ~$1.25B, up ~130% YoY vs consensus of $1.16B). Primary supplier of VCSEL/EML lasers for 800G transceivers powering Tomahawk 6 and Spectrum-6 ecosystems.", "confidence": "high", "status_trend": "strong_acceleration", "period": "Q4 FY2026 (reported Aug/Sep 2026)", "source_url": "https://beststocks.com/research/lumentum-lite-stock-jump-q4-optical-ai-datacenter-blowout", "source_title": "Lumentum (LITE) Stock Rockets 8% on Record-Breaking Q4 Results and Strong Forward Guidance", "source_date": "2026-08-28", "speed_class": "800G", "gaps": "Revenue split between AI optics vs. traditional telecom not publicly disclosed separately", "leadtime_weeks": 14, "price_trend": "rising", "capacity_signal": "Gross margin >50% for first time signals leverage from AI optics scaling; raised guidance by ~8% above street estimate indicates strong underlying demand momentum", "metrics_note": "BestStocks/Blockonomi/Perplexity composite of Lumentum Q4 FY2026 earnings (Aug 28, 2026)" }, { "id": "NET-ASTERA-LABS-PHY-SHORTAGE", "component": "Astera Labs Retimer/PHY Chips", "category": "Retimer/PHY", "claim_type": "supply_constraint", "claim": "Astera Labs PCIe/PCIe Gen6 retimers and PHY chips experiencing 16-24 week lead times for early 2026 delivery. Company gained rapid traction as 800G GPU-rack interconnect adoption accelerates; SerDes IP licensing plus chip sales model differentiates from Credo's pure-chip approach.", "confidence": "high", "status_trend": "supply_shortage_early_delivery", "period": "2026–2027", "source_url": "https://intuitionlabs.ai/articles/nvidia-gb200-supply-chain", "source_title": "NVIDIA GB200 Supply Chain: The Global Ecosystem Explained", "source_date": "2026-04-18", "speed_class": "800G", "gaps": "Astera Labs private company; exact market share vs Credo not publicly quantified", "leadtime_weeks": 20, "price_trend": "rising", "capacity_signal": "Dual business model (IP licensing + retimer chips) means capacity limited by both wafer allocation and IP implementation cycles; constrained supply benefits pricing power", "metrics_note": "Intuition Labs supply chain analysis (Apr 2026); corroborated by multiple DPU/NIC ecosystem reports" }, { "id": "NET-BRCM-AI-SEM-REV-16_7B_Q3", "component": "Broadcom AI Semiconductor Revenue — Networking + XPUs", "category": "Revenue Metrics", "claim_type": "financial_data_point", "claim": "Broadcom Q3 FY2026 AI semiconductor revenue surged 221% YoY to $16.7 billion. Total company revenue $29.6 billion (+86% YoY). Semiconductor Solutions segment $20.8B (+127% YoY). AI-related bookings continue exceptional pace with optics and Ethernet switching contributing meaningfully to the growth trajectory.", "confidence": "high", "status_trend": "explosive_growth", "period": "Q3 FY2026 (ended ~Jul 2026)", "source_url": "https://www.lighthouse-canton.com/insights/broadcom-q3-fy2026-earnings-recap-equity-insights", "source_title": "Broadcom Q3 FY2026 Earnings: AI Revenue Guide to $230bn by 2027 — LightHouse Canton", "source_date": "2026-09-03", "speed_class": "multiple", "gaps": "Networking-specific revenue breakdown not separately disclosed; optics/switching embedded within broader AI semiconductor revenue", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "$16.7B AI semiconductor revenue in single quarter = ~$67B annualized rate; bandwidth of demand far exceeds historical networking TAM", "metrics_note": "LightHouse Canton recap of Q3 FY2026 results (Sep 3, 2026); cross-referenced with Yahoo Finance transcript data" }, { "id": "NET-BRCM-OCTHPULSE-143PCENT_YOY", "component": "Broadcom AI Revenue Trajectory — FY2026 Run-Rate", "category": "Revenue Metrics", "claim_type": "market_trend", "claim": "Broadcom Q2 FY2026 AI semiconductor revenue hit $10.8 billion (+143% YoY) in a single quarter. Company guides Q3 FY2026 revenue to ~$29.4B with Wall Street penciling EPS near $3.24. AI semiconductor revenue projected to reach $115 billion by 2027, indicating networking/interconnect/optics TAM expansion well beyond traditional enterprise markets.", "confidence": "high", "status_trend": "accelerating_expansion", "period": "FY2026–FY2027", "source_url": "https://www.whatjobs.com/news/broadcom-beats-estimates-guides-ai-revenue-to-115-billion-in-2027-on-anthropic-and-openai-demand/", "source_title": "Broadcom Beats Estimates, Guides AI Revenue to $115 Billion in 2027", "source_date": "2026-09-03", "speed_class": "multiple", "gaps": "$115B figure includes custom silicon (TPU-class), XPU designs, and networking components combined", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "$115B by 2027 implies ~$57B compound growth from FY2025 baseline; networking and interconnect represent an increasing share as GPU-count-per-cluster grows non-linearly", "metrics_note": "WhatJobs/MarketBeat aggregation of investor presentations and analyst estimates (Sep 2026)" } ] }, { "key": "power", "name": "Power", "color": "#EA580C", "description": "UPS, PDUs, power supplies, electrical infrastructure", "summary": "Static UPS lead times at 30-48 weeks. Vertiv and Eaton benefiting from data center power demand surge. $95B Dell backlog signals massive power infrastructure buildout ahead.", "entries": [ { "id": "UPS-LEAD-TIME-STATIC-500-1500", "component": "Static UPS, 500-1500 kVA", "category": "UPS", "subcategory": "Static UPS Systems", "claim_type": "lead_time", "claim": "Static UPS, 500-1500 kVA: lead time 30-48 weeks. Suppliers: ABB, Schneider Electric, Vertiv, Eaton.", "confidence": "high", "status_trend": "worsening at top vendors", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "UPS-LEAD-TIME-MODULAR-2000+", "component": "Modular UPS, 2000+ kVA", "category": "UPS", "subcategory": "Modular UPS Systems", "claim_type": "lead_time", "claim": "Modular UPS, 2000+ kVA: lead time 36-52 weeks. Suppliers: Vertiv, Schneider Electric, EATON.", "confidence": "medium-high", "status_trend": "stable-to-worsening", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "UPS-LEAD-TIME-ROTARY", "component": "Rotary UPS or Diesel Rotary UPS (DRUPS)", "category": "UPS", "subcategory": "Rotary UPS Systems", "claim_type": "lead_time", "claim": "Rotary UPS or Diesel Rotary UPS (DRUPS): lead time 40-60 weeks. Suppliers: ABB, EATON, Schneider.", "confidence": "medium", "status_trend": "no significant improvement since 2023", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "UPS-BATTERY-LEAD-TIME", "component": "UPS Battery Systems", "category": "UPS", "subcategory": "UPS Battery", "claim_type": "lead_time", "claim": "UPS battery systems (lead-acid/lithium): lead time 12-20 weeks. Suppliers: Cadex, Elipse, Samsung SDI.", "confidence": "medium", "status_trend": "slowly improving post-pandemic peak", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Battery chemistry affects lead time", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "GEN-LEAD-TIME-5MW", "component": "Generator, 5-10 MW", "category": "Generator", "subcategory": "Generator", "claim_type": "lead_time", "claim": "Generator, 5-10 MW: lead time 26-40 weeks. Suppliers: Caterpillar, Cummins, MTU, Wärtsilä.", "confidence": "high", "status_trend": "rapidly increasing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "GEN-LEAD-TIME-10MW+", "component": "Generator, 10+ MW", "category": "Generator", "subcategory": "Generator", "claim_type": "lead_time", "claim": "Generator, 10+ MW: lead time 36-52 weeks. Suppliers: Caterpillar, Cummins, MTU, Wärtsilä.", "confidence": "high", "status_trend": "rapidly increasing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "GEN-NODE-PLATFORM-5MW", "component": "Generator Node Platform, 5 MW", "category": "Generator", "subcategory": "Generator Node Platform", "claim_type": "lead_time", "claim": "Generator Node Platform, 5 MW: lead time 20-30 weeks. Suppliers: Caterpillar, Cummins.", "confidence": "medium-high", "status_trend": "rapidly growing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "GEN-NODE-PLATFORM-10MW", "component": "Generator Node Platform, 10 MW", "category": "Generator", "subcategory": "Generator Node Platform", "claim_type": "lead_time", "claim": "Generator Node Platform, 10 MW: lead time 26-40 weeks. Suppliers: Caterpillar, Cummins.", "confidence": "medium-high", "status_trend": "rapidly growing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "PDU-LEAD-TIME", "component": "PDU (Power Distribution Unit)", "category": "PDU", "subcategory": "PDU", "claim_type": "lead_time", "claim": "PDU: lead time 12-20 weeks. Suppliers: EATON, Vertiv, ABB, Schneider.", "confidence": "high", "status_trend": "slowly improving post-pandemic peak", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "BUSWAY-LEAD-TIME", "component": "Busway/Trunkway", "category": "Busway/Trunkway", "subcategory": "Busway/Trunkway", "claim_type": "lead_time", "claim": "Busway/Trunkway: lead time 16-24 weeks. Suppliers: EATON, ABB, Schneider.", "confidence": "high", "status_trend": "slowly improving post-pandemic peak", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "SWITCHGEAR-LEAD-TIME", "component": "Switchgear", "category": "Switchgear", "subcategory": "Switchgear", "claim_type": "lead_time", "claim": "Switchgear: lead time 20-36 weeks. Suppliers: ABB, Schneider, Eaton, Siemens.", "confidence": "high", "status_trend": "rapidly increasing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "RACK-DENSITY-100KW", "component": "Rack Density 100kW+", "category": "Rack Density", "subcategory": "Rack Density", "claim_type": "trend", "claim": "AI rack power density increasing to 100kW+ per rack; traditional 10-15kW designs obsolete.", "confidence": "high", "status_trend": "rapidly increasing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "RACK-DENSITY-50KW", "component": "Rack Density 50kW", "category": "Rack Density", "subcategory": "Rack Density", "claim_type": "trend", "claim": "50kW per rack becoming standard for AI workloads; requires liquid cooling.", "confidence": "high", "status_trend": "rapidly increasing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "RACK-DENSITY-200KW", "component": "Rack Density 200kW+", "category": "Rack Density", "subcategory": "Rack Density", "claim_type": "trend", "claim": "200kW+ per rack emerging for next-gen AI clusters; requires direct-to-chip cooling.", "confidence": "medium", "status_trend": "rapidly increasing", "period": "2027 (est.)", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Commercial viability timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-TRANSITION-AC-DC", "component": "AC to DC Power Transition", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "trend", "claim": "Data centers transitioning from AC to DC power distribution for efficiency; 48V DC becoming standard.", "confidence": "medium", "status_trend": "newly implemented", "period": "2026–2027", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Transition timeline varies by facility", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-RELIABILITY-99.999", "component": "Power Reliability 99.999%", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "requirement", "claim": "AI data centers requiring 99.999% power availability; driving investment in redundant UPS and generator systems.", "confidence": "high", "status_trend": "newly implemented", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-COST-AI", "component": "Power Cost Impact", "category": "Power/Grid", "subcategory": "Market Context", "claim_type": "economic", "claim": "Power costs for AI data centers increasing 30-50% due to demand and grid constraints.", "confidence": "medium", "status_trend": "rapidly increasing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Cost varies by region and utility", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-RENEWABLE-requirement", "component": "Renewable Power Requirements", "category": "Power/Grid", "subcategory": "Market Context", "claim_type": "requirement", "claim": "Hyperscalers requiring 100% renewable power for new data centers; driving investment in on-site generation.", "confidence": "high", "status_trend": "newly implemented", "period": "2026–2027", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Renewable availability varies by region", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-BATTERY-STORAGE", "component": "Battery Energy Storage", "category": "UPS", "subcategory": "UPS Battery", "claim_type": "trend", "claim": "Battery energy storage systems (BESS) for grid stabilization growing 50%+ YoY.", "confidence": "medium", "status_trend": "rapidly increasing", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "BESS capacity and deployment timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-FUEL-CELL", "component": "Fuel Cell Backup", "category": "Generator", "subcategory": "Generator", "claim_type": "trend", "claim": "Fuel cell backup power systems for data centers growing; 100kW-1MW range.", "confidence": "medium", "status_trend": "increasing interest for AI transient response", "period": "2026–2027", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Fuel cell maturity and cost", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-GRID-INTERCONNECT", "component": "Grid Interconnection", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "constraint", "claim": "Grid interconnection queue times 2-5 years for new data center connections.", "confidence": "high", "status_trend": "prohibitive", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Interconnection queue varies by utility", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-TRANSFORMER-requirement", "component": "Transformer Requirements", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "requirement", "claim": "AI data centers requiring 50-100MVA transformers per facility; lead times 12-28 months.", "confidence": "high", "status_trend": "worsening at top vendors", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Transformer lead times vary by manufacturer", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-MICROGRID", "component": "Microgrid Integration", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "trend", "claim": "Data centers integrating microgrids for power resilience; solar + battery + generator hybrid.", "confidence": "medium", "status_trend": "newly implemented", "period": "2026–2027", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Microgrid integration complexity", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-POWER-QUALITY", "component": "Power Quality Requirements", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "requirement", "claim": "AI workloads requiring ultra-clean power; harmonic distortion <3%.", "confidence": "high", "status_trend": "newly implemented", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-ENERGY-EFFICIENCY", "component": "Energy Efficiency (PUE)", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "requirement", "claim": "AI data centers targeting PUE <1.1; driving innovation in power distribution design.", "confidence": "high", "status_trend": "newly implemented", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-DYNAMIC-LOAD", "component": "Dynamic Load Response", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "requirement", "claim": "AI workloads create dynamic power loads; UPS and generator systems must respond within milliseconds.", "confidence": "high", "status_trend": "newly implemented", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-COLD-CHAIN", "component": "Power Cold Chain", "category": "Power/Grid", "subcategory": "Power Distribution", "claim_type": "requirement", "claim": "Power cold chain (uninterrupted power during maintenance) critical for AI training jobs that run weeks.", "confidence": "high", "status_trend": "newly implemented", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-FUTURE-TECH", "component": "Future Power Technologies", "category": "Power/Grid", "subcategory": "Market Context", "claim_type": "roadmap", "claim": "Solid-state transformers and wide-bandgap semiconductors (SiC, GaN) targeting 2028 for power distribution.", "confidence": "medium", "status_trend": "proposed/proposing", "period": "2028 (est.)", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "Commercial viability timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "POWER-SUPPLIER-CONCENTRATION", "component": "Power Supplier Concentration", "category": "Power/Grid", "subcategory": "Market Context", "claim_type": "market_concentration", "claim": "Power equipment market highly concentrated: ABB, Schneider, Eaton, Vertiv control ~70% of market.", "confidence": "high", "status_trend": "oligopoly persists; no significant M&A activity in 2026", "period": "2026", "source_url": "https://terrapincg.com/news/switchgear-transformer-generator-lead-times-2026", "source_title": "Switchgear, Transformer & Generator Lead Times 2026", "source_date": "2026-06-10", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "VRT-Q2-2026-GUIDANCE-RAISE", "component": "Vertiv Full-Year Guidance Raised FY2026", "category": "Financial Signal", "subcategory": "Revenue Guidance", "claim_type": "fact", "claim": "Vertiv raised full-year 2026 revenue guidance to $14.0B (midpoint), up $250M from prior, representing 37% YoY growth. Q3 FY2026 guided to $3.75B (+40% YoY). Adjusted EPS raised to $6.70 (+60% over 2025).", "confidence": "high", "status_trend": "raised", "period": "FY2026 / Q3 2026", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/07/vertiv-vrt-q2-2026-earnings-call-transcript/", "source_title": "Vertiv (VRT) Q2 2026 Earnings Call Transcript — Motley Fool", "source_date": "2026-08-07", "gaps": null, "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "$14B guided revenue implies significant demand surge; inventory rose to $2.523B (from $1.457B at YE 2025) to support H2 production ramp", "metrics_note": "Guidance raise announced Q2 2026 earnings call 29 Jul 2026; source: Motley Fool transcript 7 Aug 2026" }, { "id": "VRT-Q2-2026-DEFERRED-REVENUE", "component": "Vertiv Deferred Revenue Signals Backlog Strength", "category": "Financial Signal", "subcategory": "Deferred Revenue", "claim_type": "fact", "claim": "Vertiv deferred revenue reached $3.634 billion in Q2 2026, driven by project advanced payments and ongoing milestone collections on large infrastructure deployments.", "confidence": "high", "status_trend": "strong and growing", "period": "Q2 2026", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/07/vertiv-vrt-q2-2026-earnings-call-transcript/", "source_title": "Vertiv (VRT) Q2 2026 Earnings Call Transcript — Motley Fool", "source_date": "2026-08-07", "gaps": null, "leadtime_weeks": null, "price_trend": null, "capacity_signal": "$3.634B deferred revenue represents advanced customer payments signaling contracted but not-yet-delivered backlog", "metrics_note": "Q2 2026 deferred revenue figure from Vertiv earnings release via Motley Fool transcript" }, { "id": "VRT-NET-LIVERAGE-NHFL", "component": "Vertiv Negative Net Leverage Provides Capital Flexibility", "category": "Financial Signal", "subcategory": "Balance Sheet", "claim_type": "fact", "claim": "Vertiv ended Q2 2026 with negative net leverage of 0.1x, providing significant strategic flexibility for capital deployment including capacity expansion.", "confidence": "high", "status_trend": "positive", "period": "Q2 2026", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/07/vertiv-vrt-q2-2026-earnings-call-transcript/", "source_title": "Vertiv (VRT) Q2 2026 Earnings Call Transcript — Motley Fool", "source_date": "2026-08-07", "gaps": null, "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Cash-rich balance sheet supports $1.1B capacity expansion program targeting 2027 delivery", "metrics_note": "Net leverage 0.1x from Vertiv Q2 2026 financial results" }, { "id": "ETN-DATA-CENTER-65PCT-OFFICIAL", "component": "Eaton Data Center Organic Revenue +65% Q2 2026", "category": "Financial Signal", "subcategory": "Segment Growth", "claim_type": "fact", "claim": "Eaton data center organic revenue grew 65% in Q2 2026, significantly outperforming the underlying market growth rate of 23%.", "confidence": "high", "status_trend": "accelerating above market", "period": "Q2 2026", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/07/eaton-etn-q2-2026-earnings-call-transcript/", "source_title": "Eaton (ETN) Q2 2026 Earnings Call Transcript — Motley Fool", "source_date": "2026-08-08", "gaps": null, "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Electrical Americas recorded ~$5B new orders YTD 2026 alone; sequential +$700M additional backlog after Q2", "metrics_note": "65% organic growth cited by CEO Paulo Ruiz on Q2 2026 earnings call 8 Aug 2026" }, { "id": "ETN-US-DC-BACKLOG-307GW", "component": "US Data Center Backlog Reaches 307 Gigawatts", "category": "Financial Signal", "subcategory": "Backlog", "claim_type": "fact", "claim": "Total U.S. data center backlog at Eaton reached 307 gigawatts, equivalent to 15 years at 2025 build rates, up from 12 years previously. Majority converts to deliveries in 2028 and beyond.", "confidence": "high", "status_trend": "massively elongated", "period": "Q2 2026", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/07/eaton-etn-q2-2026-earnings-call-transcript/", "source_title": "Eaton (ETN) Q2 2026 Earnings Call Transcript — Motley Fool", "source_date": "2026-08-08", "gaps": "Conversion timeline depends on site permitting, grid connection, and construction pace", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "307 GW = 15 years of backlog at 2025 rates vs 12 years previous update", "metrics_note": "CEO statement on Q2 2026 earnings call; backed by total Electrical backlog up 43% YoY" }, { "id": "ETN-BOYD-THERMAL-INTEGRATION", "component": "Eaton Acquires Boyd Thermal — 'Grid to Chip' Portfolio Complete", "category": "Financial Signal", "subcategory": "M&A Strategy", "claim_type": "fact", "claim": "Eaton acquired Boyd Thermal to complete its 'grid to chip' portfolio covering utility grid through semiconductor-level power delivery and liquid cooling. Boyd provides $1.8B projected annual revenue.", "confidence": "high", "status_trend": "integrating", "period": "2026-2027", "source_url": "https://www.fool.com/earnings/call-transcripts/2026/08/07/eaton-etn-q2-2026-earnings-call-transcript/", "source_title": "Eaton (ETN) Q2 2026 Earnings Call Transcript — Motley Fool", "source_date": "2026-08-08", "gaps": null, "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Acquisition adds cold plate and CDU capabilities, completing Eaton's liquid cooling offering", "metrics_note": "Strategic integration announced and tracked through Q2 2026 earnings cycle" }, { "id": "GEV-ELECTRIFICATION-Q1-2025-DC-ORDERS", "component": "GE Vernova Electrification: $2.4B Data Center Orders in Q1 2025", "category": "Financial Signal", "subcategory": "Data Center Orders", "claim_type": "fact", "claim": "GE Vernova's electrification segment booked $2.4 billion in data center orders in Q1 2025 alone, exceeding all of 2025 guidance, with book-to-bill near 2.5.", "confidence": "medium", "status_trend": "extremely strong", "period": "Q1 2025", "source_url": "https://finance.yahoo.com/energy/articles/restructured-energy-monopoly-no-brainer-151947187.html", "source_title": "This Restructured Energy Monopoly Is a No-Brainer Buy — FinanceBuzz/Yahoo Finance", "source_date": "2026-06-16", "gaps": "Secondary-source reporting ($2.4B / BtB 2.5); primary GE Vernova Q1 2025 transcript not yet accessed — re-verify against GE IR next wave", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Book-to-bill of 2.5 indicates orders far exceed shipments; sustained demand acceleration", "metrics_note": "Quoted from Yahoo Finance analysis citing GE Vernova disclosures" }, { "id": "CAT-POWER-ENERGY-Q2-2026", "component": "Caterpillar Power & Energy Q2 2026: Sales >$20B, Orders Through 2030", "category": "Financial Signal", "subcategory": "Generator/Engine Sales", "claim_type": "fact", "claim": "Caterpillar Power & Energy segment reported Q2 2026 sales surpassing $20 billion. Data center generators driving demand. Customers placing orders through 2030. Company's total order backlog is $72B with ~59% attributed to power and energy.", "confidence": "high", "status_trend": "record pace", "period": "Q2 2026", "source_url": "https://www.utilitydive.com/news/caterpillar-sales-surpass-20b-growing-data-center-demand-q2-2026/827569/", "source_title": "Caterpillar sales surpass $20B as generators for data centers take off — Utility Dive", "source_date": "2026-08-11", "gaps": "Total $20B includes all P&E categories; exact DC allocation not broken out in headline data", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "$72B total order backlog with 59% in power/energy; customers booking through 2030", "metrics_note": "Utility Dive reporting on Caterpillar Q2 2026 financial results 11 Aug 2026" }, { "id": "GEV-Q2-2026-ORDERS-BACKLOG", "component": "GE Vernova Q2 2026 Orders & Backlog", "category": "Power Generation", "subcategory": "Gas Power Equipment", "claim_type": "supply_status", "claim": "GE Vernova Q2 2026 orders reached $24.2B (+88% organically) led by Power and Electrification; total company backlog grew $13.0B sequentially to $176B, driven by equipment and services.", "confidence": "high", "status_trend": "rising", "period": "Q2 2026", "source_url": "https://www.gevernova.com/news/press-releases/ge-vernova-reports-second-quarter-2026-financial-results-raises-2026-financial", "source_title": "GE Vernova reports second quarter 2026 financial results and raises 2026 financial guidance", "source_date": "2026-07-22", "gaps": "Company-reported figure; independent backlog verification not performed.", "price_trend": "rising", "capacity_signal": "Backlog of $176B at quarter-end, up $13B QoQ", "metrics_note": "Direct from GE Vernova Q2 2026 press release (Jul 22, 2026), accessed 2026-09-14." }, { "id": "GEV-GAS-TURBINE-116GW-BACKLOG", "component": "Gas Turbines (GE Vernova slot reservations)", "category": "Power Generation", "subcategory": "Gas Power Equipment", "claim_type": "supply_status", "claim": "GE Vernova Gas Power equipment backlog and slot reservation agreements grew from 100 to 116 GW in Q2 2026; the company now expects at least 125 GW of gas equipment under contract by year-end 2026.", "confidence": "high", "status_trend": "tightening", "period": "2026", "source_url": "https://www.gevernova.com/news/press-releases/ge-vernova-reports-second-quarter-2026-financial-results-raises-2026-financial", "source_title": "GE Vernova reports second quarter 2026 financial results and raises 2026 financial guidance", "source_date": "2026-07-22", "gaps": "Slot reservations are reservations, not all converted firm orders; conversion timing not disclosed.", "capacity_signal": "116 GW under contract/reserved; target >=125 GW by end-2026", "metrics_note": "Direct from GE Vernova Q2 2026 press release, accessed 2026-09-14." }, { "id": "GEV-DC-ORDERS-5B-H1-2026", "component": "GE Vernova Electrification — Data Center Orders", "category": "Power Generation", "subcategory": "Electrification (Grid Solutions)", "claim_type": "fact", "claim": "GE Vernova booked $2.7B of data center orders in Electrification in Q2 2026, bringing first-half 2026 data center orders to over $5B — more than double the full-year 2025 total. CEO: 'data center orders reaching over $5 billion year-to-date, more than double our 2025 total.'", "confidence": "high", "status_trend": "rising sharply", "period": "H1 2026", "source_url": "https://www.gevernova.com/sites/default/files/gev_webcast_transcript_07222026.pdf", "source_title": "GE Vernova 2Q 2026 Earnings Call Webcast Transcript", "source_date": "2026-07-22", "gaps": "Electrification-segment data center orders only; does not include data-center-related gas equipment orders.", "metrics_note": "From GE Vernova Q2 2026 webcast transcript (lines ~95-99, 84 of transcript), accessed 2026-09-14." }, { "id": "GEV-GAS-TURBINE-CAPACITY-RAMP", "component": "Gas Turbine Production Capacity (GE Vernova)", "category": "Power Generation", "subcategory": "Gas Power Equipment", "claim_type": "capacity", "claim": "GE Vernova is on track to deliver 20 GW of annual gas turbine output starting in Q3 2026, with 24 GW in 2028 and actions underway to reach 30 GW by 2030, in response to demand driven largely by data centers and grid modernization.", "confidence": "high", "status_trend": "expanding", "period": "2026-2030", "source_url": "https://www.gevernova.com/news/press-releases/ge-vernova-reports-second-quarter-2026-financial-results-raises-2026-financial", "source_title": "GE Vernova reports second quarter 2026 financial results and raises 2026 financial guidance", "source_date": "2026-07-22", "gaps": "Capacity targets are company plans; actual ramp execution not yet verified.", "capacity_signal": "20 GW annual output from Q3 2026; 24 GW 2028; 30 GW 2030 target", "metrics_note": "CEO quote in GE Vernova Q2 2026 press release; corroborated by webcast transcript, accessed 2026-09-14." }, { "id": "GEV-US-TRANSFORMER-ORDERS-800M", "component": "US Power Transformers (GE Vernova)", "category": "Grid Equipment", "subcategory": "Transformers", "claim_type": "supply_status", "claim": "GE Vernova booked $800M in US transformer orders in H1 2026, with significant growth in US substations, switchgear and transformers to support data center development.", "confidence": "high", "status_trend": "rising", "period": "H1 2026", "source_url": "https://www.gevernova.com/sites/default/files/gev_webcast_transcript_07222026.pdf", "source_title": "GE Vernova 2Q 2026 Earnings Call Webcast Transcript", "source_date": "2026-07-22", "gaps": "US-only; global transformer order total not disclosed in this source.", "capacity_signal": "$800M H1 2026 US transformer orders", "metrics_note": "From GE Vernova Q2 2026 webcast transcript (line ~103 and ~321-323), accessed 2026-09-14." }, { "id": "SE-Q3-FY26-RECORD-ORDERS-17-9BN", "component": "Siemens Energy Q3 FY2026 Orders & Backlog", "category": "Power Generation", "subcategory": "Gas Power + Grid Technologies", "claim_type": "supply_status", "claim": "Siemens Energy Q3 FY2026 (Apr-Jun 2026) orders hit a record €17.9B (+8.5% comparable YoY); order backlog rose to a record €162bn; book-to-bill 1.57. Growth driven primarily by US demand, including large data center-related orders at Gas Services.", "confidence": "high", "status_trend": "rising", "period": "Q3 FY2026 (Apr-Jun 2026)", "source_url": "https://assets.siemens-energy.com/dam/30018eed-275d-4e73-811e-b49d004cdc3d/2026-08-05-Earnings_Release_Q3_FY26_EN-pdf_Original%20file.pdf", "source_title": "Siemens Energy Earnings Release Q3 FY 2026", "source_date": "2026-08-05", "gaps": "Company-reported; segment split: Gas Services orders €10.0B, Grid Technologies €5.4B.", "price_trend": "rising", "capacity_signal": "Record order backlog €162bn; book-to-bill 1.57", "metrics_note": "Direct from Siemens Energy Q3 FY2026 earnings release PDF (pp. 1-2), accessed 2026-09-14." }, { "id": "SE-Q3-FY26-GRID-TECH-TRANSFORMER-DC", "component": "Siemens Energy Grid Technologies — Transformers & Data Center", "category": "Grid Equipment", "subcategory": "Transformers / Data Center Grid Connection", "claim_type": "supply_status", "claim": "Siemens Energy Grid Technologies orders rose 27.6% comparable to €5.4B in Q3 FY2026, with the major contribution from substantial transformer business growth including data center projects; order backlog reached €51bn and book-to-bill was 1.48. Pre-close call: ~€2bn of US data-center-related orders booked in H1 FY2026.", "confidence": "high", "status_trend": "rising", "period": "Q3 FY2026 (Apr-Jun 2026)", "source_url": "https://assets.siemens-energy.com/dam/30018eed-275d-4e73-811e-b49d004cdc3d/2026-08-05-Earnings_Release_Q3_FY26_EN-pdf_Original%20file.pdf", "source_title": "Siemens Energy Earnings Release Q3 FY 2026", "source_date": "2026-08-05", "gaps": "Data center order attribution from pre-close call (~€2bn H1) is management-stated, not itemized in the release.", "capacity_signal": "Grid Technologies backlog €51bn; book-to-bill 1.48; ~€2bn H1 data center orders (US)", "metrics_note": "Earnings release p.3 + Q3 FY26 pre-close call transcript (2026-06-29, line ~139) 'booked around €2 billion orders related to data centers in the first half', accessed 2026-09-14." }, { "id": "SE-LPT-SWITCHGEAR-CAPACITY-50PCT", "component": "Large Power Transformers & Switchgear Capacity (Siemens Energy)", "category": "Grid Equipment", "subcategory": "Transformers / Switchgear", "claim_type": "lead_time", "claim": "Siemens Energy states growth remains limited by execution capacity rather than demand, and will increase capacity for Large Power Transformers and Switchgears by 50% between 2026 and 2030; a further big step-up in capacity additions is planned for 2029-2030.", "confidence": "high", "status_trend": "expanding (still capacity-constrained)", "period": "2026-2030", "source_url": "https://assets.siemens-energy.com/dam/630d3b4f-2d26-4803-804e-b479007691c5/2026-06-29_Q3FY26_Pre-Close-Call_Transcript-pdf_Original%20file.pdf", "source_title": "Siemens Energy Q3 FY2026 Pre-Close Call Transcript", "source_date": "2026-06-29", "gaps": "Percentage is company target, not observed; lead-time impact on specific orders not quantified in source.", "capacity_signal": "LPT + switchgear capacity +50% 2026-2030; execution capacity is the binding constraint", "metrics_note": "Management statements in pre-close call transcript (lines ~160-163, 163), accessed 2026-09-14." }, { "id": "SE-GAS-SERVICES-SLOT-RESERVATIONS", "component": "Gas Turbine Slot Reservations (Siemens Energy)", "category": "Power Generation", "subcategory": "Gas Power Equipment", "claim_type": "supply_status", "claim": "Siemens Energy Gas Services Q3 FY2026 orders reached a record €10.0B (+61.9% comparable YoY) driven by US demand including large data-center-related orders plus Middle East and Asia new power plants; Gas Services backlog is €73bn with book-to-bill 2.65. Pre-close call: gas equipment order backlog ~60 GW as of Q2 FY2026; slot reservations described as structured, non-speculative.", "confidence": "high", "status_trend": "tightening", "period": "Q3 FY2026 (Apr-Jun 2026)", "source_url": "https://assets.siemens-energy.com/dam/30018eed-275d-4e73-811e-b49d004cdc3d/2026-08-05-Earnings_Release_Q3_FY26_EN-pdf_Original%20file.pdf", "source_title": "Siemens Energy Earnings Release Q3 FY 2026", "source_date": "2026-08-05", "gaps": "~60 GW figure is as of Q2 FY2026 per pre-close call; Q3-end GW figure not stated in release.", "capacity_signal": "Gas Services backlog €73bn; ~60 GW gas order backlog (Q2 FY26); book-to-bill 2.65", "metrics_note": "Earnings release p.2 + pre-close call transcript (lines ~92, 99) 'large order backlog of around 60 GW as of Q2 Fiscal Year 2026', accessed 2026-09-14." }, { "id": "DC-MORATORIUM-374-2026", "component": "US Data Center Moratoriums (community opposition)", "category": "Grid / Permits", "subcategory": "Demand-side constraints", "claim_type": "constraint", "claim": "Wells Fargo report (as covered by Newstarget, Sep 11 2026): US data center moratoriums quadrupled to 374 as local opposition mounts ahead of midterms; analysts flagged exposure of IPPs and utilities (CEG, NRG, AEP, DTE) to state-level political risk.", "confidence": "low", "status_trend": "worsening", "period": "2026", "source_url": "https://newstarget.com/2026-09-11-data-center-moratoriums-quadruple-local-opposition-mounts.html", "source_title": "Data Center Moratoriums Quadruple to 374 as Local Opposition Mounts Ahead of Midterms, Wells Fargo Report Shows", "source_date": "2026-09-11", "gaps": "Tier-3 aggregator reporting on a Wells Fargo report; original report not accessed; jurisdictional breakdown not captured. Flagged as thin signal.", "metrics_note": "Newstarget article (Edison Reed, Sep 11 2026), accessed 2026-09-14; underlying Wells Fargo report not verified." } ] }, { "key": "grid", "name": "Grid", "color": "#9333EA", "description": "Transmission, transformers, grid infrastructure, utility capacity", "summary": "Data center power demand tripling by 2030. Transformer lead times at 144 weeks. UHV investment and reconductor capacity expansion critical for AI growth.", "entries": [ { "id": "EQ-001", "component": "Transformer lead time", "category": "transformer_lead_time", "claim_type": "fact", "claim": "Large power transformers historically shipped within ~2 years of order; in 2026, average lead times stretched to 12–28 months, high-capacity units 4–5 years.", "confidence": "medium-high", "status_trend": "worsening", "period": "2026 (estimates as of early 2026)", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-002", "component": "Transformer demand growth", "category": "transformer_demand_growth", "claim_type": "fact", "claim": "Transformer demand growing 30%+ YoY driven by AI data center construction and grid modernization.", "confidence": "high", "status_trend": "stretched", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-003", "component": "Transformer supply capacity", "category": "transformer_demand_growth", "claim_type": "fact", "claim": "Transformer manufacturing capacity constrained; major manufacturers (Siemens, ABB, GE) operating at full capacity.", "confidence": "high", "status_trend": "growing", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortance-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-004", "component": "Grid interconnection queue", "category": "queue_backlog", "claim_type": "fact", "claim": "Grid interconnection queue times 2-5 years for new data center connections.", "confidence": "high", "status_trend": "prohibitive", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-005", "component": "Grid interconnection cost", "category": "interconnection_cost", "claim_type": "fact", "claim": "Grid interconnection costs for new data centers $50M-$200M+ depending on location and capacity.", "confidence": "medium", "status_trend": "prohibitive", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Costs vary significantly by location", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-006", "component": "Transmission construction decline", "category": "transmission_construction_decline", "claim_type": "fact", "claim": "US transmission construction declined 50%+ since 2015; grid modernization lagging AI demand growth.", "confidence": "high", "status_trend": "stagnant at worst-case levels", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-007", "component": "Virginia grid constraints", "category": "regional_virginia", "claim_type": "fact", "claim": "Northern Virginia data center corridor facing grid capacity constraints; interconnection queue 3-4 years.", "confidence": "high", "status_trend": "prohibitive", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-008", "component": "Texas ERCOT queue", "category": "regional_ercot_queue", "claim_type": "fact", "claim": "Texas ERCOT interconnection queue backlog growing; AI data center demand straining grid capacity.", "confidence": "high", "status_trend": "stagnant at worst-case levels", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-009", "component": "FERC interconnection reform", "category": "ferc_action", "claim_type": "fact", "claim": "FERC implementing interconnection reform (Order No. 2023) to streamline queue process; implementation ongoing.", "confidence": "high", "status_trend": "newly implemented", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Implementation effectiveness uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-010", "component": "DOE grid modernization", "category": "doe_directive", "claim_type": "fact", "claim": "DOE directing grid modernization investments; $3B+ allocated for transmission and distribution upgrades.", "confidence": "medium-high", "status_trend": "newly implemented", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Funding timeline and allocation", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-011", "component": "Texas LLIS reform", "category": "texas_llis_reform", "claim_type": "fact", "claim": "Texas LLIS (Large Load Interconnection Study) reform proposed to accelerate AI data center interconnection.", "confidence": "medium", "status_trend": "proposed/proposing", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Reform timeline and outcomes uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-012", "component": "PJM generation requirement", "category": "pjm_generation_requirement_discussion", "claim_type": "fact", "claim": "PJM discussing generation requirements for data centers; potential mandate for on-site generation.", "confidence": "medium", "status_trend": "proposed/proposing", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Regulatory outcome uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-013", "component": "Global queue backlog", "category": "global_queue_backlog", "claim_type": "fact", "claim": "Global data center interconnection queue backlog growing; average wait times 2-5 years across major markets.", "confidence": "high", "status_trend": "massive funding gap", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-014", "component": "EU grid upgrade need", "category": "eu_grid_upgrade_need", "claim_type": "fact", "claim": "EU grid upgrade need estimated at €1.2 trillion through 2030; AI data center demand accelerating timeline.", "confidence": "medium-high", "status_trend": "massive funding gap", "period": "2026-2030", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Funding availability uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-015", "component": "Chinese export surge", "category": "chinese_export_surge", "claim_type": "fact", "claim": "Chinese transformer exports surging; China capturing 40%+ of global transformer market.", "confidence": "medium", "status_trend": "prohibitive", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Export restrictions and tariffs", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-016", "component": "Consumer cost PJM", "category": "consumer_cost_pjm", "claim_type": "fact", "claim": "PJM ratepayers bearing increasing costs for grid upgrades to serve data centers.", "confidence": "medium", "status_trend": "prohibitive", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Cost allocation methodology", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-017", "component": "Transmission delay cost", "category": "transmission_delay_cost", "claim_type": "fact", "claim": "Transmission project delays adding $10B+ to overall grid upgrade costs.", "confidence": "medium", "status_trend": "stagnant at worst-case levels", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Cost estimates vary by source", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-018", "component": "Regional Texas grid", "category": "regional_texas", "claim_type": "fact", "claim": "Texas grid capacity constraints worsening; ERCOT implementing emergency protocols for data center load.", "confidence": "high", "status_trend": "stagnant at worst-case levels", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-019", "component": "Regional other utilities", "category": "regional_other_utilities", "claim_type": "fact", "claim": "Other major utilities (Dominion, Duke, ConEd) reporting similar grid capacity constraints.", "confidence": "high", "status_trend": "stagnant at worst-case levels", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-020", "component": "FERC reform process", "category": "ferc_reform_process", "claim_type": "fact", "claim": "FERC Order No. 2023 implementation progressing; first wave of queue repositioning expected 2027.", "confidence": "medium", "status_trend": "proposed/proposing", "period": "2026-2027", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Implementation timeline uncertain", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-021", "component": "Market size grid upgrades", "category": "market_size", "claim_type": "fact", "claim": "Global grid upgrade market projected to reach $200B+ by 2030; AI data center demand driving growth.", "confidence": "medium", "status_trend": "massive funding gap", "period": "2026-2030", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Market size estimates vary by source", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-022", "component": "Demand share data centers", "category": "demand_share", "claim_type": "fact", "claim": "Data centers projected to consume 8%+ of global electricity by 2030; AI workloads driving 50%+ of growth.", "confidence": "medium", "status_trend": "massive funding gap", "period": "2030", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Projections vary by source", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-023", "component": "Wait time transformer", "category": "wait_time", "claim_type": "fact", "claim": "Average wait time for transformer delivery 12-28 months; high-capacity units 4-5 years.", "confidence": "high", "status_trend": "stretched", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "EQ-024", "component": "Withdrawal rate queue", "category": "withdrawal_rate", "claim_type": "fact", "claim": "Data center interconnection withdrawal rate 30-40%; projects abandoned due to queue times and costs.", "confidence": "medium", "status_trend": "prohibitive", "period": "2026", "source_url": "https://kxy-group.com/data-center-switchgear-transformer-shortage-2026/", "source_title": "Why Data Centers Face a Switchgear & Transformer Shortage in 2026 — KXY Group", "source_date": "2026-08-04", "gaps": "Withdrawal rate varies by region", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "GRID-QBACK-2061GW-2026", "component": "US interconnection queue size", "category": "interconnection_queue", "claim_type": "fact", "claim": "US interconnection queue contains 2,061 GW of generation and storage projects seeking grid connection as of 2026, down from ~2,300 GW at end of 2024 due primarily to withdrawals (75% long-term withdrawal rate), not construction. Average queue wait for projects energized in 2024 was 55 months.", "confidence": "high", "status_trend": "shrinking on paper but deeper problem — most capacity withdrawn rather than built", "period": "2026 (end of 2024 baseline: 2,300 GW)", "source_url": "https://osakawire.com/en/the-grid-is-the-bottleneck-copper-queues-permits/", "source_title": "The Grid Is the Bottleneck — 2,061 GW Stuck in Queues (OsakaWire Intelligence, Aug 2026) citing Berkeley Lab EIP/EMP", "source_date": "2026-08-28", "gaps": "Breakdown by RTO/region, technology type, and cluster status not granularly sourced; Berkeley Lab primary data blocked by Cloudflare", "leadtime_weeks": 240, "price_trend": null, "capacity_signal": "2,061 GW queued globally ~2,500 GW per IEA; 75% withdrawal rate means effective pipeline << 2,061 GW", "metrics_note": "Ledger [1] OsakaWire citing Berkeley Lab (emp.lbl.gov) which blocked via Cloudflare; [2] OsakaWire cites IEA Electricity 2026 for global 2,500 GW figure" }, { "id": "GRID-GS-TRANSFORMER-LT-144WK-2026", "component": "Generator step-up transformer lead time", "category": "transformer_lead_time", "claim_type": "lead_time", "claim": "Generator step-up (GSU) transformer lead times averaged 144 weeks (~2.8 years) in 2026 per Wood Mackenzie Q2 2025 survey cited by CoBank review; power transformers averaged 128 weeks. 2025 US supply deficit: ~30% for power transformers, ~10% for distribution transformers.", "confidence": "high", "status_trend": "stretched — worst equipment category in grid chain", "period": "2026", "source_url": "https://natlawreview.com/article/grids-bottleneck-isnt-just-money-its-materials", "source_title": "The Grid's Bottleneck Isn't Just Money. It's Materials. (Foley & Lardner LLP / Scott D. Ellis, National Law Review, Sep 2026)", "source_date": "2026-09-03", "gaps": "Lead times vary by manufacturer, size, voltage class; figures are averages from industry press reports of Wood Mackenzie survey", "leadtime_weeks": 144, "price_trend": "rising", "capacity_signal": "Wood Mackenzie 2025 US supply deficit: 30% power transformers, 10% distribution; manufacturers committed ~$2B new capacity since 2023 but plants take 2-4 years to stand up", "metrics_note": "CoBank July 2026 sector review cites Wood Mackenzie Q2 2025 transformer survey via industry publications (Utility Dive, etc.). GSU-specific: 144 weeks. Ledger [3]" }, { "id": "GRID-TX-LLIS-QUADRUPOLE-2025", "component": "Texas ERCOT large-load interconnection queue growth", "category": "interconnection_queue", "subcategory": "ERCOT", "claim_type": "fact", "claim": "ERCOT large-load interconnection queue nearly quadrupled in one year, growing from 63 GW to 226 GW (Sept 2025 to Sept 2026), driven overwhelmingly by data center demand.", "confidence": "high", "status_trend": "rapidly worsening — massive demand surge from AI data centers", "period": "2025–2026", "source_url": "https://www.latitudemedia.com/news/ercots-large-load-queue-has-nearly-quadrupled-in-a-single-year/", "source_title": "ERCOT's Large Load Queue Has Nearly Quadrupled in a Single Year (Latitude Media, Dec 2025)", "source_date": "2025-12-03", "gaps": "Detailed breakdown by customer type (hyperscale vs colocation), specific counties/districts, and resolution rate not fully sourced", "leadtime_weeks": 200, "price_trend": "rising", "capacity_signal": "226 GW ≈ 2× Texas current peak demand of ~80 GW; represents unprecedented concentration of AI-driven load in a single RTO", "metrics_note": "Latitude Media primary report. Ledger [5]. Cross-referenced with OsakaWire Section 01 noting 'data centres quadrupled' in Texas LL queue." }, { "id": "GRID-DATA-CENTER-DEMAND-TRIPLE-2030", "component": "Data center grid-power demand trajectory", "category": "demand_growth", "claim_type": "fact", "claim": "Data center grid-power demand to rise 22% in 2025, then nearly triple by 2030. Global data center capacity projected to nearly double to 200 GW by 2030. SP Global/451 Research: 143 GW total grid-power demand by 2030. Goldman Sachs estimates ~$720B required grid upgrades through 2030 to accommodate growth.", "confidence": "medium", "status_trend": "rapidly worsening", "period": "2025–2030", "source_url": "https://www.spglobal.com/energy/en/news-research/latest-news/electric-power/101425-data-center-grid-power-demand-to-rise-22-in-2025-nearly-triple-by-2030", "source_title": "Data center grid-power demand to rise 22% in 2025, nearly triple by 2030 (S&P Global, Oct 2024)", "source_date": "2024-10-14", "gaps": "Concentration risk: demand growing 'in a handful of regions rather than spread'; exact geographic distribution not granularly sourced here", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "$720B Goldman Sachs grid-upgrade estimate; 200 GW global data center capacity by 2030 doubles current ~100 GW", "metrics_note": "S&P Global primary. Cross-ref: Introl blog aggregates same SP Global data + Goldman Sachs. Ledger [7], [8]" }, { "id": "GRID-CN-UHV-INVESTMENT-4T-YUAN-2026", "component": "China UHV transmission investment commitment", "category": "transmission_investment", "subcategory": "China", "claim_type": "financial", "claim": "State Grid Corporation committed 4 trillion yuan (~$580B) for 2026–2030 period (40% increase over prior five-year plan), targeting 15 new ultra-high-voltage (UHV) lines. China UHV network grew from ~28,000 km to >40,000 km during 2021–2025 plan, operating 45 UHV projects (19 AC + ≥23 DC). Approval-to-energisation cycle: ~2 years.", "confidence": "high", "status_trend": "accelerating — world-leading build rate", "period": "2026–2030", "source_url": "https://osakawire.com/en/the-grid-is-the-bottleneck-copper-queues-permits/", "source_title": "Two Years in China, a Decade in the West (OsakaWire Intelligence, Aug 2026, section 05)", "source_date": "2026-08-28", "gaps": "Specific 2026 project-level breakdowns not sourced; Chinese planning documents difficult to access independently", "leadtime_weeks": 104, "price_trend": "stable", "capacity_signal": "40,000+ km UHV by late 2025; 12,000 km added in 5 years; Changji-Guquan link: 3,293 km at 1,100 kV carrying 12,000 MW", "metrics_note": "OsakaWire citing ChinaTalk, State Grid announcements. Approval-to-energy cycle of ~2 years contrasts sharply with Western 10–20 year permitting cycles. Ledger [1]" }, { "id": "GRID-EU-GRIDS-PKG-584EUR-2025", "component": "EU grids investment requirement", "category": "transmission_investment", "subcategory": "EU", "claim_type": "fact", "claim": "European Commission estimates €584B grid investment required before end of 2026/2028/2030 decade. European Grids Package (Dec 2025) proposes tacit approval (permitting capped at 2 years standard / 3 years complex), raising Connecting Europe Facility for Energy from €5.84B (2021-2027) to €29.91B (2028-2034).", "confidence": "high", "status_trend": "newly proposed — implementation TBD", "period": "2025–2030", "source_url": "https://osakawire.com/en/the-grid-is-the-bottleneck-copper-queues-permits/", "source_title": "What the Evidence Actually Supports (OsakaWire Intelligence, Aug 2026, sections 02 & 07)", "source_date": "2026-08-28", "gaps": "Final legislative text may differ; actual EU budget appropriations subject to co-decision process", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "€584B is 50x the current CEF-Energy envelope (€5.84B); tacit approval mechanism removes administrative clock but does not bind physical manufacturing timeline", "metrics_note": "OsakaWire citing European Commission December 2025 package. Ledger [1]" }, { "id": "GRID-TRANSFORMER-PRICE-INFLATION-2026", "component": "Transformer price index and input material costs", "category": "transformer_costs", "claim_type": "fact", "claim": "Power transformer prices up ~77% since 2019. Distribution transformers up 78-95%. Generator step-up units up ~45%. Grain-oriented electrical steel (GOES) costs doubled since January 2020. Cleveland-Cliffs is sole domestic GOES producer in US — creates single-source dependency for US transformer manufacturers.", "confidence": "high", "status_trend": "rising — confirmed real scarcity signal", "period": "2019–2026", "source_url": "https://natlawreview.com/article/grids-bottleneck-isnt-just-money-its-materials", "source_title": "The Grid's Bottleneck Isn't Just Money. It's Materials. (National Law Review / Foley & Lardner, Sep 2026)", "source_date": "2026-09-03", "gaps": "Price indices by region/country not comprehensively sourced; GOES price doubling attributed to general market trend, not single report", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Demand growth for GSU transformers: +274% 2019-2025. Overall power-transformer demand: +119%. Cleveland-Cliffs = only US GOES producer", "metrics_note": "Wood Mackenzie Q2 2025 survey data cited by industry publications, verified through CoBank/Cleveland-Cliffs sourcing. Ledger [3]" }, { "id": "GRID-SWITCHGEAR-PRIVILEGE-LOCKOUT-2026", "component": "Medium-voltage switchgear procurement timelines", "category": "switchgear_lead_time", "claim_type": "fact", "claim": "Terrapin Construction Group reports medium-voltage switchgear procurement times of 52–80 weeks and pad-mount distribution transformers 40–65 weeks based on active 2026 manufacturer-slot reservations. Eaton committed $340M Feb 2025 to expand SC plant (Jonesville) to third US three-phase transformer facility. Siemens Energy building first US large PT plant ($150M Charlotte NC) expecting 57 units/year full capacity ~early 2027.", "confidence": "medium", "status_trend": "constrained — but manufacturers investing heavily", "period": "2026", "source_url": "https://natlawreview.com/article/grids-bottleneck-isnt-just-money-its-materials", "source_title": "The Grid's Bottleneck Isn't Just Money. It's Materials. (National Law Review / Foley & Lardner, Sep 2026)", "source_date": "2026-09-03", "gaps": "Terrapin data reflects contractor-reserved slots, not national survey; vendor-specific pricing unavailable", "leadtime_weeks": 66, "price_trend": "rising", "capacity_signal": "~$2B total manufacturer capital committed to new/expanded transformer capacity since 2023. Hitachi Energy: $1B+ including $457M South Boston VA (nation's largest LT plant by 2028)", "metrics_note": "Terrapin ranges from active contractor reservation data. Manufacturer commitments: Eaton, Siemens Energy, Hitachi Energy separately reported. Ledger [3]" }, { "id": "GRID-UTILITY-CAPEX-FORECAST-2026-2030", "component": "US utility capital expenditure forecast", "category": "utility_capex", "claim_type": "financial", "claim": "S&P Global Regulatory Research projects $1.295T utility capex 2026-2030 (up from $1.169T in Dec 2025 forecast). EEI member companies invested $208B in 2025 (vs $139.8B in 2020, +49%). EEI projects $1.1T cumulative capex 2025-2029. Transmission investment reached $32.6B in 2024, projected $39.9B in 2025.", "confidence": "medium", "status_trend": "strong demand signal but constrained by manufacturing timeline", "period": "2025–2030", "source_url": "https://natlawreview.com/article/grids-bottleneck-isnt-just-money-its-materials", "source_title": "The Grid's Bottleneck Isn't Just Money. It's Materials. (National Law Review / Foley & Lardner, Sep 2026)", "source_date": "2026-09-03", "gaps": "Capex projections ≠ actual delivered infrastructure; regulatory approvals may delay/scale spending. Regional breakdowns not comprehensively sourced.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "EIA data: distribution capex rose $31.4B (+160%) between 2003-2023, with $6.5B increase alone 2022-2023 ($50.9B total). Utilities spent $104B on distribution in 2025.", "metrics_note": "S&P Global upward revision in 4 months underscores changing cost trajectory. EEI/EIA independent government/data sources. Ledger [3]" }, { "id": "GRID-RECONDUCTORING-CAPACITY-DOUBLE-2035", "component": "Reconductoring capacity potential", "category": "transmission_upgrade", "claim_type": "fact", "claim": "Replacing conventional conductors with advanced composite-core designs in existing rights-of-way could roughly double American transmission capacity by 2035, cut wholesale electricity costs by 3-4%, save ~$85B in system costs by 2035 and $180B by 2050. Delivered in months-to-a-year vs 10-15 years for greenfield new lines.", "confidence": "medium", "status_trend": "opportunity — underutilised solution", "period": "2026–2035", "source_url": "https://osakawire.com/en/the-grid-is-the-bottleneck-copper-queues-permits/", "source_title": "Copper, Steel and Skilled Hands (OsakaWire Intelligence, Aug 2026, section 04)", "source_date": "2026-08-28", "gaps": "Analysis source not directly accessible; economic assumptions around conductor cost curves and corridor availability need verification", "leadtime_weeks": 52, "price_trend": "stable", "capacity_signal": "Potential: ~2× US transmission capacity without new right-of-way. Fastest available capacity gain (months-vs-years timeline)", "metrics_note": "OsakaWire section 04 cites analysis on advanced conductors. Key differentiator: no litigation, no new corridor, substitutes materials science for legal process. Ledger [1]" } ] }, { "key": "oem", "name": "OEM / Server", "color": "#0891B2", "description": "Server OEMs, system integrators, custom silicon partners", "summary": "Dell, HPE, Supermicro, Lenovo all reporting record AI server orders. Supermicro $60B Q4 new orders. Allocation hierarchy favors AI over traditional workloads.", "entries": [ { "id": "HPE-01", "component": "HPE Cray EX254n", "category": "portfolio", "product": "HPE Cray EX254n", "claim_type": "product_status", "claim": "8U supercomputing blade with 8× NVIDIA Grace Hopper GH200 liquid-cooled GPUs. First/primary Cray adopter of GH200 for exascale computing.", "confidence": "high", "status_trend": "standard SKU — widely available, lower differentiation than Neptune line", "period": "2025–2026", "source_url": "https://www.servethehome.com/8x-nvidia-grace-hopper-superchips-arm-in-a-blade-hpe-cray-ex254n-at-gtc-2024/", "source_title": "8x NVIDIA Grace Hopper Superchips in a Blade — HPE Cray EX254n at GTC 2024", "source_date": "2024-03-18", "customer_channel": "supercomputing", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "HPE-02", "component": "HPE Neptune AI", "category": "portfolio", "product": "HPE Neptune AI", "claim_type": "product_status", "claim": "HPE Neptune: GPU-optimized AI server line; differentiates from standard Cray offerings.", "confidence": "high", "status_trend": "standard SKU — widely available, lower differentiation than Neptune line", "period": "2026", "source_url": "https://www.storagereview.com/news/hpe-expands-next-gen-cray-supercomputing-portfolio-for-the-ai-hpc-era", "source_title": "HPE Expands Next-Gen Cray Supercomputing Portfolio for the AI-HPC Era", "source_date": "2025-11-14", "customer_channel": "enterprise", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "HPE-03", "component": "HPE AI server backlog", "category": "financial", "product": "HPE AI servers", "claim_type": "financial", "claim": "HPE AI server backlog converting to revenue Q3+; servers declining but backlog providing visibility.", "confidence": "high", "status_trend": "transition — servers declining but backlog converting to revenue Q3+", "period": "2026", "source_url": "https://www.techi.com/hpe-earnings-preview-juniper-networking-ai-june-2026/", "source_title": "HPE Earnings Preview: Juniper Makes Networking the AI Story", "source_date": "2026-05-28", "customer_channel": "enterprise", "gaps": "Exact backlog figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "HPE-04", "component": "HPE GPU allocation", "category": "supply_chain", "product": "HPE NVIDIA GPUs", "claim_type": "supply_chain", "claim": "HPE NVIDIA GPU allocation constrained; hyperscalers capture 40-60% of NVIDIA GPU supply.", "confidence": "high", "status_trend": "managed risk but structural barriers persist", "period": "2026", "source_url": "https://www.chokepoints.ai/stack/compute-hardware/ai-server-oems", "source_title": "AI Server OEMs: Dell Leads with 20% Share in 2024 as Hyperscalers Lock 40–60% of Production Capacity", "source_date": "2026-07-15", "customer_channel": "enterprise", "gaps": "Exact allocation figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "HPE-05", "component": "HPE India portfolio", "category": "customer_channel", "product": "HPE India", "claim_type": "market_access", "claim": "HPE completing full-stack 'Made in India' portfolio; India tender exclusion remains unresolved.", "confidence": "low", "status_trend": "active — completing full-stack 'Made in India' portfolio", "period": "2026", "source_url": null, "source_title": null, "source_date": null, "customer_channel": "government", "gaps": "Tender exclusion timeline Downgraded from medium — no independent verification of HPE-specific India tender exclusion Downgraded from medium — no independent verification of HPE-specific India tender exclusion [F5 downgrade] Downgraded from medium — no independent verification of HPE-specific India tender exclusion", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "HPE-06", "component": "HPE integration complexity", "category": "supply_chain", "product": "HPE systems", "claim_type": "operational", "claim": "HPE system integration complexity does not scale linearly with demand; lead times 12-20 weeks.", "confidence": "medium", "status_trend": "stable — integration complexity does not scale linearly with demand", "period": "2026", "source_url": "https://www.storagereview.com/news/hpe-expands-next-gen-cray-supercomputing-portfolio-for-the-ai-hpc-era", "source_title": "HPE Expands Next-Gen Cray Supercomputing Portfolio for the AI-HPC Era", "source_date": "2025-11-14", "customer_channel": "enterprise", "gaps": "Exact lead times vary by configuration", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "HPE-07", "component": "HPE supply chain risk", "category": "supply_chain", "product": "HPE supply chain", "claim_type": "risk", "claim": "HPE supply chain risk: NVIDIA dependency, CoWoS packaging, HBM3e supply.", "confidence": "high", "status_trend": "managed risk but structural barriers persist", "period": "2026–2027", "source_url": "https://www.chokepoints.ai/stack/compute-hardware/ai-server-oems", "source_title": "AI Server OEMs: Dell Leads with 20% Share in 2024 as Hyperscalers Lock 40–60% of Production Capacity", "source_date": "2026-07-15", "customer_channel": "enterprise", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "LEN-01", "component": "Lenovo ThinkSystem SR670 V3", "category": "portfolio", "product": "Lenovo ThinkSystem SR670 V3", "claim_type": "product_status", "claim": "8× NVIDIA B200 GPU server; liquid-cooled option available.", "confidence": "high", "status_trend": "standard SKU — widely available, lower differentiation than Neptune line", "period": "2026", "source_url": "https://news.lenovo.com/pressroom/press-releases/lenovo-unveils-hybrid-ai-solutions-delivering-power-of-generative-ai-to-enterprises-with-nvidia/", "source_title": "Lenovo Unveils Hybrid AI Solutions — NVIDIA Partnership Announcement (Tech World)", "source_date": "2025-05", "customer_channel": "enterprise", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "LEN-02", "component": "Lenovo AI server portfolio", "category": "portfolio", "product": "Lenovo AI servers", "claim_type": "product_status", "claim": "Lenovo AI server portfolio: ThinkSystem SR670 V3, NE900 supercomputer, liquid-cooled variants.", "confidence": "high", "status_trend": "standard SKU — widely available, lower differentiation than Neptune line", "period": "2026", "source_url": "https://biztechdaily.com/article/lenovo-q4-fy26-record-ai-server-revenue", "source_title": "Lenovo Hit a Record $21.6B in Q4. AI Server Revenue Jumped 84%.", "source_date": "2026-06-01", "customer_channel": "enterprise", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "LEN-03", "component": "Lenovo GPU allocation", "category": "supply_chain", "product": "Lenovo NVIDIA GPUs", "claim_type": "supply_chain", "claim": "Lenovo NVIDIA GPU allocation constrained; hyperscalers capture 40-60% of NVIDIA GPU supply.", "confidence": "high", "status_trend": "managed risk but structural barriers persist", "period": "2026", "source_url": "https://www.chokepoints.ai/stack/compute-hardware/ai-server-oems", "source_title": "AI Server OEMs: Dell Leads with 20% Share in 2024 as Hyperscalers Lock 40–60% of Production Capacity", "source_date": "2026-07-15", "customer_channel": "enterprise", "gaps": "Exact allocation figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "LEN-04", "component": "Lenovo India portfolio", "category": "customer_channel", "product": "Lenovo India", "claim_type": "market_access", "claim": "Lenovo 'Made in India' portfolio; India tender exclusion remains unresolved.", "confidence": "high", "status_trend": "active — completing full-stack 'Made in India' portfolio", "period": "2026", "source_url": "https://economictimes.indiatimes.com/tech/artificial-intelligence/lenovo-to-make-ai-servers-in-india-as-nations-tech-push-deepens/articleshow/113420058.cms", "source_title": "Lenovo to Make AI Servers in India as Nation's Tech Push Deepens", "source_date": "2024-09-17", "customer_channel": "government", "gaps": "Tender exclusion timeline", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "LEN-05", "component": "Lenovo integration complexity", "category": "supply_chain", "product": "Lenovo systems", "claim_type": "operational", "claim": "Lenovo system integration complexity does not scale linearly with demand; lead times 12-20 weeks.", "confidence": "medium-low", "status_trend": "stable — integration complexity does not scale linearly with demand", "period": "2026", "source_url": "https://www.livemint.com/companies/news/chinese-checkers-how-pc-maker-lenovo-stayed-safe-11733835393829.html", "source_title": "Chinese Checkers: How PC Maker Lenovo Stayed Safe", "source_date": "2024-12-09", "customer_channel": "enterprise", "gaps": "Exact lead times vary by configuration Downgraded from medium — lead time estimate inferred from comparable OEM analysis, not Lenovo-specific source Downgraded from medium — lead time estimate inferred from comparable OEM analysis, not Lenovo-specific source", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "LEN-06", "component": "Lenovo supply chain risk", "category": "supply_chain", "product": "Lenovo supply chain", "claim_type": "risk", "claim": "Lenovo supply chain risk: NVIDIA dependency, CoWoS packaging, HBM3e supply.", "confidence": "high", "status_trend": "managed risk but structural barriers persist", "period": "2026–2027", "source_url": "https://www.chokepoints.ai/stack/compute-hardware/ai-server-oems", "source_title": "AI Server OEMs: Dell Leads with 20% Share in 2024 as Hyperscalers Lock 40–60% of Production Capacity", "source_date": "2026-07-15", "customer_channel": "enterprise", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "LEN-07", "component": "Lenovo financial performance", "category": "financial", "product": "Lenovo AI servers", "claim_type": "financial", "claim": "Lenovo AI server revenue growing 50%+ YoY in 2026; driven by B200 adoption.", "confidence": "high", "status_trend": "transition — servers declining but backlog converting to revenue Q3+", "period": "2026", "source_url": "https://biztechdaily.com/article/lenovo-q4-fy26-record-ai-server-revenue", "source_title": "Lenovo Hit a Record $21.6B in Q4. AI Server Revenue Jumped 84%.", "source_date": "2026-06-01", "customer_channel": "enterprise", "gaps": "Exact revenue figures proprietary", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-01", "component": "PowerEdge XE9680", "category": "portfolio", "product": "Dell PowerEdge XE9680", "claim_type": "product_status", "claim": "Dell's first 8-way GPU server - 4U rackmount, up to 8x NVIDIA GPUs (Blackwell GB200/B200 and Hopper H100). Liquid cooling required for high-end configs; supports NVLink, NVSwitch, Infiniband, Spectrum-X interconnects.", "confidence": "medium", "status_trend": "standard SKU - BTO through channel/partner network and direct enterprise sales", "period": "2026", "source_url": "https://www.dell.com/en-us/work/product/ps12/poweredge-xe9680", "source_title": "Dell PowerEdge XE9680 Product Page", "source_date": "2026", "customer_channel": "enterprise", "gaps": "Product page heavily JS-rendered; exact specs derived from brief summary which notes limited text extraction possible. No PDF spec sheet directly fetched.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-02", "component": "PowerEdge XE9980", "category": "portfolio", "product": "Dell PowerEdge XE9980", "claim_type": "product_status", "claim": "Next-generation multi-socket / dense configuration AI server for Rubin-era architectures. Dell builds AI servers to NVIDIA reference designs including GB200 NVL72 chassis specifications.", "confidence": "medium", "status_trend": "pipeline - early orders accepted for Rubin era (mid-to-late 2027)", "period": "2027 (est.)", "source_url": null, "source_title": null, "source_date": null, "customer_channel": "enterprise", "gaps": "Thin signal - Rubin-era delivery timelines speculative extrapolation per brief limitations section", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-03", "component": "AI server revenue & backlog", "category": "financial", "product": "Dell AI servers", "claim_type": "financial", "claim": "Full-year AI server revenue forecast raised to $74 billion in fiscal 2026 - revision from ~13% growth estimate six months prior. Implies ~35-40% of Dell's total annual revenue. Growth of 200+% over fiscal 2025.", "confidence": "high", "status_trend": "rapid growth - accelerating from elevated base", "period": "FY2026", "source_url": "https://investors.delltechnologies.com", "source_title": "Dell Q2 FY2027 Earnings Release", "source_date": "2026-08-26", "customer_channel": "enterprise", "gaps": "Exact breakdown of AI server revenue within ISG line item not separately disclosed", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-04", "component": "GPU supply allocation constraints", "category": "supply_chain", "product": "NVIDIA GPU server systems", "claim_type": "supply_chain", "claim": "NVIDIA Blackwell GB200 faced thermal issues with NVL72 configurations and initial interconnect technology problems on Blackwell silicon, prompting TSMC production method revisions. These limited initial supply and created allocation challenges. Priority allocation goes to largest/pre-existing enterprise relationships.", "confidence": "high", "status_trend": "managed risk - NVIDIA ramping production; thermal fixes deployed", "period": "2026–2027", "source_url": "https://siliconangle.com", "source_title": "SiliconAngle Analysis - Dell AI Server Revenue Growth", "source_date": "2026-09-01", "customer_channel": "enterprise", "gaps": "Exact lead time figures (weeks/months) inferred from industry patterns, not Dell-specific", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-05", "component": "Manufacturing & supply chain dependency", "category": "supply_chain", "product": "Dell supply chain", "claim_type": "risk", "claim": "Dell integrates NVIDIA GPUs at contract manufacturing facilities (Foxconn and others). Third-party component sourcing creates single-source dependencies. Geopolitical trade policy affects semiconductor supply. CoWoS packaging at TSMC and HBM3e supply constrain overall capacity.", "confidence": "high", "status_trend": "structural barriers persist - upstream dependencies at NVIDIA/TSMC/Samsung/Micron/SK Hynix", "period": "2026–2027", "source_url": "https://www.chokepoints.ai/stack/compute-hardware/ai-server-oems", "source_title": "AI Server OEMs: Dell Leads with 20% Share in 2024 as Hyperscalers Lock 40–60% of Production Capacity", "source_date": "2026-07-15", "customer_channel": "enterprise", "gaps": "", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-06", "component": "Financial performance Q2 FY2027", "category": "financial", "product": "Dell Technologies", "claim_type": "financial", "claim": "Q2 FY2027 total revenue $31.6 billion, up 4% year-over-year. ISG strong growth with traditional servers & networking +122%, storage +26%, client solutions +20%. Broadened growth across servers, storage, and client segments.", "confidence": "high", "status_trend": "accelerating - multiple business units showing strong growth", "period": "Q2 FY2027", "source_url": "https://investors.delltechnologies.com", "source_title": "Dell Q2 FY2027 Earnings Press Release", "source_date": "2026-08-26", "customer_channel": "enterprise", "gaps": "ISG revenue split (estimated $13-15B from context); geographic revenue breakdown pattern consistent with 10-K filings but exact percentages not extracted", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-07", "component": "Competitive differentiation", "category": "market", "product": "Dell PowerEdge ecosystem", "claim_type": "competitive_position", "claim": "Dell recognized as one of the top two (alongside Samsung) NVIDIA GPU system integrators. Gartner identifies Dell as key infrastructure vendor for AI workloads. Differentiators: global manufacturing/distribution footprint, deep hyperscaler partnerships dating back decades, Dell ProSupport service ecosystem, co-engineering capability with NVIDIA/Intel/AMD.", "confidence": "medium", "status_trend": "strong - top-two positioning solidified by Blackwell demand", "period": "2026", "source_url": "https://www.wccftech.com", "source_title": "Wccftech - NVIDIA Blackwell NVL72 Thermal Issue Reporting", "source_date": "2026", "customer_channel": "enterprise", "gaps": "Market share percentage from chokepoints.ai (20% in 2024); IDC/Gartner specific ranking not accessible without subscription", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "OEM-COMP-01", "component": "Dell vs HPE vs Lenovo GPU allocation", "category": "supply_chain", "product": "Dell vs HPE vs Lenovo", "claim_type": "comparison", "claim": "All three OEMs constrained by NVIDIA's capacity lock-in - hyperscalers capture 40-60% of NVIDIA GPU allocation. Dell leads with 20% share among OEMs in 2024. Dell has priority access via deeper NVIDIA co-engineering relationship (engineering iterations on GB200 NVL72), giving it a structural advantage over HPE and Lenovo.", "confidence": "high", "status_trend": "managed risk but structural barriers persist - Dell edges out competitors via direct NVIDIA partnership", "period": "2026", "source_url": "https://www.chokepoints.ai/stack/compute-hardware/ai-server-oems", "source_title": "AI Server OEMs: Dell Leads with 20% Share in 2024 as Hyperscalers Lock 40–60% of Production Capacity", "source_date": "2026-07-15", "customer_channel": "enterprise", "gaps": "Exact allocation figures proprietary; Dell vs HPE vs Lenovo differential breakdown not publicly available", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "OEM-COMP-02", "component": "Dell vs HPE vs Lenovo India/market access", "category": "customer_channel", "product": "Dell vs HPE vs Lenovo", "claim_type": "comparison", "claim": "Lenovo holds advantage with 'Made in India' portfolio and local AI server manufacturing commitment (announced Sept 2024). HPE completing full-stack Made in India portfolio. Dell operates through global distribution hubs including APAC but lacks same-level local manufacturing commitment. All face India tender exclusion from government purchasing.", "confidence": "medium", "status_trend": "Lenovo leads on local presence; HPE catching up; Dell faces steeper market entry barrier", "period": "2026", "source_url": "https://economictimes.indiatimes.com/tech/artificial-intelligence/lenovo-to-make-ai-servers-in-india-as-nations-tech-push-deepens/articleshow/113420058.cms", "source_title": "Lenovo to Make AI Servers in India as Nation's Tech Push Deepens", "source_date": "2024-09-17", "customer_channel": "government", "gaps": "HPE tender exclusion timeline unverifiable; Dell-specific India strategy not independently documented [F5 downgrade applies]", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "OEM-COMP-03", "component": "Dell vs HPE vs Lenovo differentiation", "category": "portfolio", "product": "Dell vs HPE vs Lenovo", "claim_type": "comparison", "claim": "Dell: top-two NVIDIA GPU integrator (alongside Samsung), global scale, ProSupport ecosystem, hyperscaler partnerships dating decades. HPE: Neptune GPU-optimized line differentiates from standard Cray. Lenovo: ThinkSystem RM70 V3 with B200, liquid-cooled options. Dell's channel partner network + direct enterprise model gives broader reach than either competitor.", "confidence": "medium", "status_trend": "stable - Dell's scale and NVIDIA partnership remain hard-to-match moat", "period": "2026", "source_url": "https://www.storagereview.com/news/hpe-expands-next-gen-cray-supercomputing-portfolio-for-the-ai-hpc-era/", "source_title": "HPE Expands Next-Gen Cray Supercomputing Portfolio for the AI-HPC Era", "source_date": "2025-11-14", "customer_channel": "enterprise", "gaps": "Differentiation may evolve as new product cycles roll out; Samsung's role as wild card not quantified", "leadtime_weeks": null, "price_trend": null, "capacity_signal": null, "metrics_note": null }, { "id": "DELL-Q2-FY2027-ISG-BACKLOG", "component": "Dell Technologies — AI Server Demand Surge", "category": "Server OEMs", "subcategory": "Revenue / Backlog Guidance", "claim_type": "fact", "claim": "Dell reported Q2 FY2027 revenue of $47.0B (+58% YoY). ISG (servers+networking) delivered $12.9B (+69% YoY); total ISG reached $16.8B. AI server backlog surged to $95.0B at end of Q2 (up from $51.3B in Q1). Full-year FY2027 revenue guidance raised to $192.0B (midpoint, up from prior ~$165B); non-GAAP EPS raised to $25.50 (from $17.90). Fiscal 2026 saw >$64B in AI-optimized server orders booked.", "confidence": "high", "status_trend": "record momentum", "period": "Q2 FY2027 / FY2026 annual", "source_url": "https://www.blocksandfiles.com/ai-ml/2026/09/02/dells-revenue-bonanza-benefits-from-killer-ai-workloads/5293866", "source_title": "Dell's AI servers sell like hotcakes again — SiliconANGLE / Blocks & Files", "source_date": "2026-09-02", "gaps": "ISG detail broken as servers+networking ($12.9B) plus storage/services making up remainder; precise split not always provided in secondary sources.", "leadtime_weeks": 5, "price_trend": "rising", "capacity_signal": "$95B AI server backlog signals multi-year delivery pipeline; Dell entering Q3 FY2027 with record backlog, suggesting capacity utilization at or near maximum", "metrics_note": "Lead time sourced from ReluTech industry survey (March 2026): Dell OEM pre-crisis 2–3 weeks → 4–6 weeks current (median 5w); price change +15–20% 2025–2026. Dollar figures from Dell earnings announcements and industry reporting Sept 2026." }, { "id": "DELL-RYSE-FY2027-GUIDANCE", "component": "Dell Raises FY2027 Guidance — $192B Revenue / $25.50 EPS", "category": "Server OEMs", "subcategory": "Guidance Raise", "claim_type": "fact", "claim": "In Q2 FY2027 earnings (Aug/Sep 2026), Dell raised full-year revenue guidance by ~$25B to $192.0B midpoint and non-GAAP EPS guidance to $25.50 from $17.90, reflecting unprecedented AI infrastructure demand. Adjusted EPS came in at $7.04 in Q2 itself.", "confidence": "high", "status_trend": "raised materially", "period": "FY2027 annual guidance / Q2 FY2027 actuals", "source_url": "https://siliconangle.com/2026/09/01/dells-ai-servers-sell-like-hot-cakes-again-driving-a-stunning-earnings-beat/", "source_title": "Dell's AI servers sell like hotcakes again — SiliconANGLE", "source_date": "2026-09-01", "gaps": null, "leadtime_weeks": 5, "price_trend": "rising", "capacity_signal": "$95B backlog + $192B guided revenue implies Dell is scaling production to fill ~$74B AI server revenue run-rate; COO warned investors of \"unprecedented\" memory shortage affecting cost basis", "metrics_note": "Guidance raise announced alongside Q2 FY2027 earnings report, early September 2026. Source: Seeking Alpha news summary, SiliconANGLE live coverage." }, { "id": "SUPERMICRO-FY2026-ANNUAL-RESULTS", "component": "Supermicro FY2026 Annual — $39.1B Revenue, 78% Growth", "category": "Server OEMs", "subcategory": "Annual Results", "claim_type": "fact", "claim": "Super Micro Computer reported fiscal 2026 full-year revenue of $39.06B (+77.8% YoY vs FY2025). Q4 FY2026 alone generated $11.12B in net sales. Gross margin dropped modestly from 11.1% to 10.8% overall; Q4 GAAP gross margin rose sharply to 17.5% (vs 9.9% QoQ, 9.5% YoY). Diluted EPS Q4: $1.62 beat est. of $0.96. Company projects Q1 FY2027 revenue of $14.5–15.5B (analysts expected $11.68B).", "confidence": "high", "status_trend": "surge continued into FY2027", "period": "FY2026 annual / Q4 FY2026 / Q1 FY2027 guidance", "source_url": "https://finance.yahoo.com/markets/stocks/articles/supermicro-q4-fy2026-earnings-margins-205145427.html", "source_title": "Supermicro Q4 FY2026 earnings: margins surge — Yahoo Finance / MarketWatch", "source_date": "2026-08-18", "gaps": "Fiscal 2026 breakdown by quarter available via 10-K filing; Q1/Q2/Q3 quarterly splits not captured here. Margin compression at annual level despite Q4 improvement needs monitoring.", "leadtime_weeks": 6, "price_trend": "rising", "capacity_signal": "Entered Q1 FY2027 with record order backlog after generating >$60B in new orders in Q4 alone; received ~$39B in new AI server orders from 20+ customers (June 2026 announcement); supply chain expansion via $7B financing plan noted in mid-2026", "metrics_note": "Lead time inferred from industry-wide OEM survey data: SMCI follows similar pattern to peers (5–8 weeks). Margin recovery from 9.5% to 17.5% QoQ in Q4 is notable structural improvement." }, { "id": "SUPERMICRO-Q4-FY2026-MARGIN-SWING", "component": "Supermicro Q4 FY2026 — Gross Margin Doubled to 17.5%", "category": "Server OEMs", "subcategory": "Margin Improvement", "claim_type": "fact", "claim": "Supermicro's Q4 FY2026 gross margin reached 17.5% GAAP — nearly double the prior quarter's 9.9% and up from 9.5% a year ago. The margin swing is among the sharpest in the company's recent history. Non-GAAP gross margin also improved significantly. Revenue came in $610M below consensus near $11.55B, yet stock still gained 7–15% on the margin surprise and record order book signal.", "confidence": "high", "status_trend": "sharp turnaround", "period": "Q4 FY2026", "source_url": "https://tradune.com/earnings/super-micro-q4-fy2026/", "source_title": "Super Micro Q4 FY2026: gross margin doubled — Tradune", "source_date": "2026-08-18", "gaps": "Exact non-GAAP vs GAAP margin reconciliation needs 10-K filing details. Reasons for Q3 margin collapse and recovery need deeper investigation.", "leadtime_weeks": 6, "price_trend": "stable", "capacity_signal": "Record order backlog at transition to FY2027; Q1 revenue guide $14.5–15.5B far exceeded consensus of $11.68B", "metrics_note": "Source: Tradune earnings analysis 18 Aug 2026, cross-ref'd with Yahoo Finance/Reuters reporting." }, { "id": "HPE-Q3-FY2026-CLOUD-AI", "component": "HPE Q3 FY2026 — Cloud & AI Revenue $9.0B (+25.4%)", "category": "Server OEMs", "subcategory": "Segment Performance", "claim_type": "fact", "claim": "HPE reported Q3 FY2026 Cloud & AI revenue of $9.0B (+25.4% YoY) with 17.0% operating profit margin (up from 7.0% YoY). Networking revenue surged 74.9% YoY, driven largely by the Juniper Networks acquisition. Total revenue grew ~11% YoY. Management called AI a 'multi-year growth driver' and raised full-year FY2027 revenue growth framework to 13–17%, with Networking specifically guided to grow 14–17%.", "confidence": "high", "status_trend": "accelerating", "period": "Q3 FY2026", "source_url": "https://www.hpe.com/us/en/newsroom/press-release/2026/09/hpe-reports-fiscal-2026-third-quarter-results.html", "source_title": "HPE reports fiscal 2026 third quarter results — HPE Newsroom", "source_date": "2026-09-02", "gaps": "Total FY2026 full-year revenue not separately captured yet; Q3 is latest complete quarter. Cloud & AI includes both traditional server and AI systems — AI-specific split within segment not fully disaggregated in public reporting.", "leadtime_weeks": 6, "price_trend": "rising", "capacity_signal": "Networking for AI orders reached $700M in Q3 (+ triple digits), cumulative FY2026 NA orders at $2.2B; HPE's AI systems backlog reached $5.9B earlier in fiscal year", "metrics_note": "Lead time from industry survey: HPE OEM 2–4 weeks pre-crisis → 5–7 weeks current (median 6w); price change +15%. Q3 source: HPE official IR press release 2 Sep 2026." }, { "id": "HPE-NETWORKING-JUNIPER-GROWTH", "component": "HPE Networking Revenue Surges 74.9% on Juniper Acquisition", "category": "Server OEMs", "subcategory": "M&A Impact", "claim_type": "fact", "claim": "HPE's Networking segment revenue surged 74.9% YoY in Q3 FY2026, primarily driven by consolidation of Juniper Networks results. Prior quarter saw networking revenue up 148.2% YoY. Networks for AI orders reached $700M in Q3 (triple-digit growth), bringing cumulative FY2026 NA orders to $2.2B. HPE's stock climbed ~120% YTD following the momentum.", "confidence": "high", "status_trend": "rapidly growing", "period": "Q3 FY2026", "source_url": "https://247wallst.com/investing/2026/09/02/live-can-hpes-q3-earnings-tonight-keep-its-120-ytd-rally-going/", "source_title": "Can HPE's Q3 Earnings Tonight Keep Its 120% YTD Rally Going? — 24/7 Wall St.", "source_date": "2026-09-02", "gaps": "Juniper contribution blended with legacy HPE networking in segment reporting; exact standalone Juniper numbers not publicly separated yet.", "leadtime_weeks": 6, "price_trend": "rising", "capacity_signal": "Juniper integration accelerating HPE's positioning in AI network infrastructure; networking remains a major growth engine beyond just servers", "metrics_note": "Sources combined from HPE official Q3 IR release, 24/7 Wall St. analysis, TradingKey transcript." }, { "id": "LENOVO-FY2026-ANNUAL-AI-Doubles", "component": "Lenovo FY2026 — $83B Annual Revenue, AI Sales Double (+105%)", "category": "Server OEMs", "subcategory": "Annual Results", "claim_type": "fact", "claim": "Lenovo Group reported FY2026 annual revenue of $83.1B. AI-related sales grew 105% for the full year. Management predicted revenue would reach the $100B milestone in coming years. Q1 FY2026/27 was strongest quarter in group history: $26.9B revenue (+43% YoY, highest growth rate in past 5 years), adjusted net profit surged 176% to $1.075B (clearing $1B mark first time). Shares surged ~20–22% on the print.", "confidence": "high", "status_trend": "strongest quarter in group history", "period": "FY2026 annual / Q1 FY2026/27", "source_url": "https://qz.com/lenovo-record-earnings-ai-revenue-doubles-052226", "source_title": "Lenovo Q4 FY2026 earnings: record revenue as AI nearly doubles — Quartz", "source_date": "2026-08-13", "gaps": "AI server unit volumes not disclosed separately from overall AI-related sales figure. China domestic vs international split not captured.", "leadtime_weeks": 7, "price_trend": "rising", "capacity_signal": "Management targeting $100B revenue milestone; aggressive AI expansion through Hybrid AI strategy combining edge-to-cloud; Q1 showed highest growth rate in 5 years", "metrics_note": "Annual revenue from Lenovo ESG report reference (FY2026/27 English); Q1 figures from Reuters 13 Aug 2026. Lenovo OEM industry survey: 2–3 weeks pre-crisis → 6–8 weeks current (median 7w); price change +15%." }, { "id": "OEM-INDUSTRY-LEADTIMES-2026", "component": "OEM Server Lead Times 2026 — All Major Vendors Significantly Elongated", "category": "Server OEMs", "subcategory": "Supply Chain Lead Times", "claim_type": "lead_time", "claim": "OEM server lead times stretched dramatically in 2026 due to DRAM/memory allocation constraints driven by AI infrastructure demand. Pre-crisis norms of 2–4 weeks expanded to 4–8+ weeks depending on vendor and configuration. Memory-heavy configurations face the longest delays. DRAM contract prices climbed ~50% in 2025, with server DRAM up 60%+ in some configurations. Major OEMs (Dell, HP, Lenovo, HPE) implemented ~15% system-wide price increases in Q1 2026.", "confidence": "high", "status_trend": "chronically elongated", "period": "2026", "source_url": "https://relutech.com/blogs/hardware/the-2026-server-hardware-crisis", "source_title": "The 2026 Server Hardware Crisis: OEM Lead Times & Rising Costs Explained — ReluTech", "source_date": "2026-03-30", "gaps": "Lead times vary by configuration complexity and volume commitments. Hyperscaler tier gets priority allocation (~80% of requested volumes); major OEMs get ~70%; smaller resellers get 30–50%. Real-time weekly fluctuations not tracked.", "leadtime_weeks": 6, "price_trend": "rising", "capacity_signal": "Gartner projects combined DRAM and SSD prices will surge 130% by end of 2026. TSMC signaling 10%+ price increases for sub-5nm nodes. U.S. tariffs on Chinese electronics and China rare-earth export restrictions add further pressure. Gartner warning Feb 2026.", "metrics_note": "Source: ReluTech March 2026 article synthesizing IT Vortex, ITWeb Global Memory Shortage, SoftwareSeni Hardware Procurement Strategy, Gartner Feb 2026, Sourceability 2026 Semiconductor Outlook. Vendor-specific data: Dell 4-6w (+15-20%), Lenovo 6-8w (+15%), HPE 5-7w (+15%)." }, { "id": "DRAM-ALLOCATION-HIERARCHY", "component": "DRAM Allocation Priority — AI Infrastructure Diverts Capacity from Enterprise", "category": "Server OEMs", "subcategory": "Supply Chain Disruption", "claim_type": "supply_status", "claim": "Memory manufacturers Samsung, SK Hynix, and Micron shifted significant fabrication capacity toward HBM and high-density DDR5 for AI training/inference clusters. This created a tiered allocation hierarchy: hyperscalers (AWS, Azure, Google Cloud) secure ~80% of requested allocations; major OEMs (Dell, HP, Lenovo) receive ~70%; module makers/resellers get 30–50%; rest competes on spot market at premium prices. Result: enterprise server memory costs surged 60%+ in several configurations.", "confidence": "medium", "status_trend": "structural shortage through 2026", "period": "2025–2026", "source_url": "https://relutech.com/blogs/hardware/the-2026-server-hardware-crisis", "source_title": "The 2026 Server Hardware Crisis — ReluTech (citing BaCloud, SoftwareSeni, IT Vortex)", "source_date": "2026-03-30", "gaps": "Allocation percentages are industry estimates rather than official disclosures. Spot market pricing is highly variable by SKU. Exact impact varies by OEM sourcing agreements.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Memory availability now dictates delivery timelines more than chassis or CPU supply. DRAM contract prices climbed ~50% in 2025, server DRAM exceeding 60% gains in some configs.", "metrics_note": "Synthesized from three independent industry sources citing BaCloud Global Server Hardware Price Trends 2025, SoftwareSeni Hardware Procurement Strategy, IT Vortex 2026." }, { "id": "SUPERMICRO-$60B-Q4-NEW-ORDERS", "component": "Supermicro Generated >$60B New Orders in Q4 FY2026", "category": "Server OEMs", "subcategory": "Order Book", "claim_type": "fact", "claim": "Supermicro CEO Charles Liang reported the company 'generated more than $60 billion in new orders' in Q4 FY2026 alone, starting FY2027 with a record backlog. The company also highlighted it had previously received ~$39B in new AI server orders from 20+ customers (June 2026 announcement). Stock jumped 10–15% on preliminary earnings forecast releasing Q4 order data.", "confidence": "high", "status_trend": "record order book", "period": "Q4 FY2026 / Filing into FY2027", "source_url": "https://siliconangle.com/2026/07/21/ai-server-maker-supermicros-stock-gains-60b-order-backlog-stronger-margins/", "source_title": "AI server maker Supermicro's stock gains on $60B order backlog — SiliconANGLE", "source_date": "2026-07-21", "gaps": "New orders ≠ recognized revenue; conversion timing uncertain. Order backlog composition by customer not disclosed.", "leadtime_weeks": 6, "price_trend": "rising", "capacity_signal": ">$60B Q4-only orders + previous $39B in new AI server orders indicates massive multi-quarter pipeline; liquid-cooling solutions remain key differentiator", "metrics_note": "SiliconANGLE 21 Jul 2026 covering preliminary Q4 earnings forecast statement. Cross-ref with formal Q4 results (Aug 11, 2026) confirming record backlog." } ] }, { "key": "cooling", "name": "Cooling", "color": "#0284C7", "description": "Liquid cooling, immersion, thermal management, heat exchangers", "summary": "Vertiv $14.06B backlog signals massive cooling infrastructure demand. Liquid cooling transitioning from niche to standard for AI workloads. Immersion cooling adoption accelerating.", "entries": [ { "id": "VRT-FY2026-Q2-EARNINGS-BACKLOG-GUIDANCE", "component": "Vertiv Holdings — FY2026 Q2 Financial Results & Full-Year Guidance", "category": "Cooling", "subcategory": "Data Center Infrastructure", "claim_type": "revenue_record", "claim": "FY2026 Q2 net sales reached $3.27 billion (+50% YoY), adjusted diluted EPS grew 60%. Backlog hit record $14.06 billion (up from $8.12 billion year earlier). Management raised full-year guidance to $13.5–$14.0 billion revenue and $6.30–$6.40 adjusted EPS. Americas region (approx. 70% of revenue) grew 44% YoY driven by Microsoft, Meta, Google, AWS hyperscaler builds.", "confidence": "high", "status_trend": "record_backlog_and_guidance", "period": "FY2026 Q2 (ended ~Jul 2026)", "source_url": "https://investors.vertiv.com/news/news-details/2026/Vertiv-Reports-Strong-Second-Quarter-2026-with-Diluted-EPS-Growth-of-53-Adjusted-Diluted-EPS-Growth-of-60-Raises-Full-Year-2026-Guidance-Across-All-Key-Metrics/default.aspx", "source_title": "Vertiv Reports Strong Second Quarter 2026 — Investor Relations", "source_date": "2026-07-29", "gaps": "Revenue breakdown between liquid cooling vs. power distribution products not disclosed separately; likely >40% of total revenue is now cooling-related given Blackwell TDP-driven conversion.", "leadtime_weeks": 24, "price_trend": "rising", "capacity_signal": "Record backlog $14.06B provides >3 quarters visibility; AI rack density escalation to 145kW+ accelerates CDU deployment cycles", "metrics_note": "Q2 FY2026 earnings release; Forbes Apr 2026 analysis confirmed $15B order book; 24-week lead time reflects supply chain constriction for liquid cooling SKUs amid capacity expansion lag." }, { "id": "MOD-FA2026-Q4-SALES-GROWTH-4B-DEAL", "component": "Modine Manufacturing — Airedale Data Center Cooling Revenue Surge & $4B Contract", "category": "Cooling", "subcategory": "Cooling Solutions", "claim_type": "revenue_record", "claim": "Fiscal 2026 Q4 data center cooling sales soared 158% YoY overall; companywide sales up 53% (47% organic, ~5 points acquisitions). Management targeted >$1 billion annual data center revenue run rate for fiscal 2026. Secured landmark $4 billion long-term capacity agreement through 2029 with a strategic data center customer; includes $165 million upfront payment for capacity expansion. Fourth consecutive quarter beating both earnings and revenue estimates.", "confidence": "high", "status_trend": "four_quarters_consecutive_beat", "period": "FY2026 Q4 (ended Jun 2026)", "source_url": "https://www.prnewswire.com/news-releases/modine-announces-landmark-4-billion-long-term-capacity-agreement-through-2029-with-strategic-data-center-customer-for-airedale-by-modine-cooling-solutions-302779610.html", "source_title": "Modine Announces Landmark $4 Billion Long-Term Capacity Agreement Through 2029 — PR Newswire", "source_date": "2026-05-27", "gaps": "Remaining revenue run-up before $4B deal recognized; FY2027 guidance details not fully public.", "leadtime_weeks": 18, "price_trend": "rising", "capacity_signal": "$4B multi-year contract with $165M capex commitment indicates sustained demand through 2029; new capacity build-in-progress at Airedale facilities", "metrics_note": "Seeking Alpha article (Aug 2026) confirms Q4 sales +158%; Rallies.ai reports 53% companywide growth; multiple analyst articles corroborate >$1B annual target." }, { "id": "NVT-FY2026-Q2-RECORD-DATA-CENTER-$2B-Target", "component": "nVent Electric — Record Q2 2026 Earnings & $2B Data Center Target", "category": "Cooling", "subcategory": "Data Center Solutions", "claim_type": "backlog_milestone", "claim": "FY2026 Q2 record sales with 53% YoY revenue growth and 69% jump in adjusted EPS. Data center segment targeting $2 billion in revenue for full year 2026 — representing more than twofold increase versus previous year. Backlog stood at $2.5B-$2.6B providing visibility through end of 2026 and into 2027. Infrastructure vertical approached ~45% of total company revenue.", "confidence": "high", "status_trend": "data_center_doubles_annually", "period": "FY2026 Q2 (ended Jul 2026)", "source_url": "https://finance.yahoo.com/markets/stocks/articles/nvent-electric-plc-nvt-q2-210059323.html", "source_title": "nVent Electric PLC (NVT) Q2 2026 Earnings Call Highlights — Yahoo Finance", "source_date": "2026-08-04", "gaps": "nVent does not break out direct-to-chip vs. immersion revenue; portion of $2B target attributable to NVIDIA partnership unclear.", "leadtime_weeks": 16, "price_trend": "rising", "capacity_signal": "$2.5-2.6B backlog at nearly 2x last year's base; manufacturing capacity expansion underway to convert backlog into future revenue", "metrics_note": "Yahoo Finance transcript of Q2 FY2026 call; Motley Fool transcript available; Zacks reporting $2.6B backlog." }, { "id": "EATON-BOYD-THERMAL-ACQUISITION-$9.5B-LIQUID-COOLING", "component": "Eaton Corporation — $9.5B Acquisition of Boyd Thermal Division", "category": "Cooling", "subcategory": "Corporate Strategy", "claim_type": "acquisition", "claim": "Eaton acquired the Boyd Thermal division of Boyd Corporation from Goldman Sachs Asset Management for $9.5 billion. Boyd Thermal has expected revenue of $1.7 billion for fiscal 2026, with $1.5 billion derived from liquid cooling for data centers (core activity). Combined offering creates integrated power-and-cooling solution from chip to power grid. Valuation at 22.5x EBITDA signals strong management conviction on DLC growth trajectory.", "confidence": "high", "status_trend": "integration_phase", "period": "Announced Nov 2025, expected close 2025-2026", "source_url": "https://itdaily.com/news/business/eaton-boyd-thermal/", "source_title": "Eaton Pays 9.5 Billion Dollars for Data Center Cooling Specialist Boyd Thermal — ITdaily", "source_date": "2025-11-05", "gaps": "Final acquisition close date and any regulatory approval status uncertain; post-merger integration plan details limited.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "$1.5B annual liquid cooling revenue base from Boyd Thermal adds ~5K employees across NA/EU/Asia; Eaton's combined power + cooling scale creates #2 position after Vertiv", "metrics_note": "ITdaily Nov 2025 report corroborated by LinkedIn industry commentary noting same trend; comparable: Schneider Electric completed Motivair acquisition ($850M) in Oct 2024." }, { "id": "SCHNEIDER-MOTIVAIR-DTC-ACQUISITION-850M", "component": "Schneider Electric — $850M Motivair Direct-to-Chip Acquisition", "category": "Cooling", "subcategory": "Corporate Strategy", "claim_type": "acquisition", "claim": "Schneider Electric signed agreement to acquire controlling interest (75%) in Buffalo-based Motivair Corp. for $850 million (all cash including tax step-up value). Motivair valued at mid-single digit multiple of projected FY2025 revenue. Transaction strengthens Schneider's direct-to-chip liquid cooling portfolio and high-capacity thermal solutions offerings, directly competing with Vertiv's integrated approach.", "confidence": "high", "status_trend": "completed_oct2024", "period": "Completed Oct 2024", "source_url": "https://www.reuters.com/markets/deals/schneider-electric-buy-data-centre-cooling-firm-motivair-850-million-2024-10-17/", "source_title": "Schneider Electric to buy data centre cooling firm Motivair for $850 million — Reuters", "source_date": "2024-10-17", "gaps": "Motivair's standalone revenue figure not publicly disclosed; post-integration contribution to Schneider's fiscal 2026 results unclear.", "leadtime_weeks": 20, "price_trend": "rising", "capacity_signal": "Motivair specialized in DTC cold plates and CDU integration for GPU clusters; acquisition gives Schneider immediate competitive capability in fastest-growing cooling segment", "metrics_note": "Reuters primary source; Facilities Dive corroborated terms; Motivair corporate website confirms transaction completion." }, { "id": "DTCC-MARKET-SIZE-333B-1731B-CAGR265", "component": "Direct-to-Chip Cooling Market — Size Forecast 2026-2032", "category": "Cooling", "subcategory": "Market Sizing", "claim_type": "forecast", "claim": "Global data center direct-to-chip cooling market projected to grow from USD 3.33 billion in 2026 to USD 17.31 billion by 2032, at CAGR of 26.5%. By cooling technology type, direct-to-chip holds largest share. Liquid cooling manifold market separately projected to reach USD 6.33 billion by 2033 at CAGR 31.2%. Direct-to-chip architecture accounts for ~44% of total cooling solutions deployed in 2026.", "confidence": "medium", "status_trend": "rapid_expansion_early_stage", "period": "2026-2032 forecast", "source_url": "https://www.marketsandmarkets.com/Market-Reports/data-center-direct-to-chip-cooling-market-164825126.html", "source_title": "Data Center Direct-to-chip Cooling Market — MarketsandMarkets Report", "source_date": "2026-07-03", "gaps": "Multiple market research firms use different definitions for 'direct-to-chip' boundary (cold plate only vs. TCS incl. CDU); reported sizes may vary ±20%.", "leadtime_weeks": null, "price_trend": "falling", "capacity_signal": "CAGR 26.5% outpaces overall data center infrastructure spend growth — implies supply chain expansion must accelerate to prevent bottleneck", "metrics_note": "MarketsandMarkets primary; Valuespectrum cross-reference shows manifold-specific market at $0.94B in 2026 growing to $6.33B by 2033 (CAGR 31.2%). Future Market Insights confirms direct-to-chip at ~44% share of cooling architecture mix." }, { "id": "CDU-MARKET-SIZE-12B-15B-CAGR25", "component": "CDU Coolant Distribution Unit Market — Size Forecast 2025-2035", "category": "Cooling", "subcategory": "Market Sizing", "claim_type": "forecast", "claim": "Global data center liquid cooling coolant distribution unit (CDU) market estimated at US $1.2 billion in 2025, projected to grow at CAGR >25% from 2025 to 2035 to reach approximately US $15 billion by 2035. CDU represents critical enabling infrastructure layer between chilled water / secondary loops and server-side cold plates or immersion tanks.", "confidence": "medium", "status_trend": "accelerating_demand", "period": "2025-2035 forecast", "source_url": "https://kghresearch.com/data-center-liquid-cooling-cdus-market/", "source_title": "Data Center Liquid Cooling CDUs Market – KGH Research Market Overview", "source_date": "2026-07-15", "gaps": "Single-source estimate; not independently cross-referenced by a second major analyst firm in accessible sources.", "leadtime_weeks": 24, "price_trend": "rising", "capacity_signal": ">25% CAGR for CDU alone suggests supply chain scaling needed beyond cold plate manufacturing", "metrics_note": "KGH Research proprietary model; aligns broadly with preceedence research noting CDU market to attain ~$14.6B by 2035." }, { "id": "OCP-STANDARDIZATION-DLC-UQD-COLDPLATE", "component": "OCP Cold Plate Sub-Project — Liquid Cooling Standardization Roadmap", "category": "Cooling", "subcategory": "Standards & Interoperability", "claim_type": "industry_development", "claim": "Open Compute Project advancing standardization of data center liquid cooling across five functional areas: Cold Plate, CDU, Immersion, Door Heat Exchanger, and Heat Reuse. Key deliverables include UQD (Universal Quick Disconnect) specification revision to v2.0, coolant guidelines for water-based fluids (July 2024), propylene glycol-based fluids (updated Nov 2025), and two-phase dielectric fluid guidelines (July 2024). Participants include NVIDIA, Meta, Dell, Intel, BP/Castrol. Meta completed EPDM hose evaluation for Blind Mate Quick Connector (BMQC) on OpenRack v3 in Jan 2026.", "confidence": "high", "status_trend": "specifications_drafting_phase", "period": "2024-2026 (ongoing)", "source_url": "https://sector-signals.net/en/themes/datacenter-power/articles/ocp-liquid-cooling-standardization", "source_title": "OCP Data Center Liquid Cooling Standards Explained — Sector Signals", "source_date": "2026-06-26", "gaps": "Standard adoption timeline vs. proprietary deployments still in competition phase; some hyperscalers pursuing custom designs alongside OCP specs.", "leadtime_weeks": null, "price_trend": "falling", "capacity_signal": "Standardized UQD connectors + coolant compatibility will reduce vendor lock-in, potentially enabling additional suppliers to enter market and drive price-down pressure on commoditized components", "metrics_note": "Sector Signals deep-dive (Jun 2026) compiled from OCP primary documents; BMQC spec published by Meta Jan 14, 2026." }, { "id": "BLACKWELL-Ultra-RACK-DENSITY-THERMAL-CHALLENGES", "component": "AI Server Rack Density Escalation — Thermal Design Limits 2026-2028", "category": "Cooling", "subcategory": "System Architecture", "claim_type": "industry_trend", "claim": "GPU TDP exceeding 1,000 watts makes traditional air cooling non-viable for next-gen systems. Current AI racks averaging 40 kW per rack, planned 145 kW racks with NVIDIA Blackwell Ultra GPUs entering near-term roadmap. First 1 MW racks expected by 2028. This thermal density escalation drives mandatory transition to liquid cooling (direct-to-chip preferred for GPU/CPU pairs), fundamentally restructuring data center design and increasing cooling spend per rack from <$50K to >$150K+.", "confidence": "high", "status_trend": "mandatory_transition_phase", "period": "2026-2028", "source_url": "https://itdaily.com/news/business/eaton-boyd-thermal/", "source_title": "Eaton Pays 9.5 Billion Dollars for Data Center Cooling Specialist Boyd Thermal — ITdaily", "source_date": "2025-11-05", "gaps": "Timeline uncertainty on 145 kW rack commercial availability; 1 MW rack estimate extrapolative based on current growth trajectory.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "40 kW/rack → 145 kW/rack → 1 MW/rack trajectory means each subsequent generation requires exponentially more cooling capacity per sq ft", "metrics_note": "ITdaily/Aussteins analysis; correlated with Vertiv's admission that traditional air cooling 'no longer viable' at >1000W GPU TDP levels." }, { "id": "MOLEX-BoundaryCool-INVESTMENT-ACTIVE-LIQUID-COOLING", "component": "Molex Investment in CAEPlus BoundaryCool Active Liquid Cooling", "category": "Cooling", "subcategory": "Supply Chain Diversification", "claim_type": "supply_development", "claim": "Molex invested in CAEPlus BoundaryCool, characterized as part of Molex's data center thermal management strategy. Goal: advance active liquid cooling technology for AI data centers. BoundaryCool operates outside traditional cold plate paradigm, offering alternative liquid cooling approaches that could diversify supply chain dominance held by Vertiv, nVent, Modine.", "confidence": "low", "status_trend": "technology_validation", "period": "2026", "source_url": "https://www.facebook.com/TheDataCenterEngineer/posts/molex-backs-caeplus-boundarycool-active-liquid-cooling-for-ai-data-centers-more-/122134400337097412/", "source_title": "Molex invests in CAEPlus BoundaryCool liquid cooling for AI data centers — Facebook / Data Center Engineer", "source_date": "2026-08-26", "gaps": "BoundaryCool technology stage unknown; Facebook post is social media source rather than formal press release or investor communication.", "leadtime_weeks": null, "price_trend": "null", "capacity_signal": "Molex involvement signals tier-1 connector supplier sees liquid cooling as structurally important to data center buildout going forward", "metrics_note": "Facebook post from TheDataCenterEngineer — lower confidence source; Molex data center thermal management push corroborated by Molex's general product announcements." }, { "id": "MOD-DC-Q1FY2027-REVENUE-90PCT-GROWTH", "component": "Modine Manufacturing — Data Centers Segment, Q1 Fiscal 2027 Results", "category": "cooling", "subcategory": "Data Center Infrastructure", "claim_type": "revenue_record", "claim": "Modine's Data Centers segment (Airedale) reported Q1 fiscal 2027 sales of $348.6 million, up ~90% YoY from $183.7 million; total company net sales +28% to $874.1 million. Data Centers is now reported as a separately disclosed segment (split from Climate Solutions) and FY2027 guidance is unchanged at net sales +20-35% and adjusted EBITDA $650-680 million.", "confidence": "high", "status_trend": "rapid_revenue_ramp", "period": "Q1 FY2027 (ended June 30, 2026)", "source_url": "https://www.prnewswire.com/news-releases/modine-reports-first-quarter-fiscal-2027-results-302838002.html", "source_title": "Modine Reports First Quarter Fiscal 2027 Results — PR Newswire", "source_date": "2026-07-29", "gaps": "No customer-level disclosure of which segment share maps to the $4B 2029 capacity agreement; segment gross margin figure not captured in extracted text.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Data Centers revenue ~doubling YoY within one year of the $4B 2029 capacity deal; separate-segment reporting signals it is now a major, fast-growing business line", "metrics_note": "Primary: Modine press release via PR Newswire, July 29 2026; '90%' figure from CEO Neil Brinker's quoted segment commentary." }, { "id": "MOD-DC-Q1FY2027-SUPPLY-CHAIN-BOTTLENECK", "component": "Modine — Data Centers Segment Near-Term Supply Chain Bottlenecks", "category": "cooling", "subcategory": "Supply Chain", "claim_type": "supply_status", "claim": "In its Q1 fiscal 2027 release, Modine management disclosed 'near-term supply chain challenges in our Data Centers segment' and stated it is 'taking decisive actions to resolve these bottlenecks and have already made significant progress' — a first public acknowledgement of cooling-component supply constraint at a major DLC vendor.", "confidence": "high", "status_trend": "bottleneck_acknowledged", "period": "Q1 FY2027 (disclosed July 29, 2026)", "source_url": "https://www.prnewswire.com/news-releases/modine-reports-first-quarter-fiscal-2027-results-302838002.html", "source_title": "Modine Reports First Quarter Fiscal 2027 Results — PR Newswire", "source_date": "2026-07-29", "gaps": "Bottleneck specifics (which component: cold plates, CDUs, manifolds, or coolant) are not named; no quantified impact on delivery timing disclosed.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "First explicit public supply-constraint admission from a top-3 data center cooling vendor, corroborating the broader 2026 DLC supply-tightness signal", "metrics_note": "Primary: Modine press release, July 29 2026, quoted CEO commentary; no independent corroboration yet." }, { "id": "NVT-Q2-2026-LIQUID-COOLING-EXPANSION-GUIDANCE-RAISE", "component": "nVent Electric — Q2 2026 Results: Liquid Cooling Manufacturing Expansion & FY Guidance Raise", "category": "cooling", "subcategory": "Data Center Solutions", "claim_type": "capacity_expansion", "claim": "nVent announced 'another manufacturing expansion for liquid cooling to meet continued data center demand' alongside record Q2 2026 sales of $1.5 billion (+53% YoY; +47% organic) and adjusted EPS of $1.45 (+69% YoY). Full-year 2026 sales guidance raised to +37-39% reported (+32-34% organic) from +26-28% previously, and adjusted EPS guidance raised to $5.00-$5.10. Management credited 'significant data center growth' plus new products with more than 30 points of sales growth.", "confidence": "high", "status_trend": "capacity_expansion_and_guidance_raise", "period": "Q2 2026 (announced July 31, 2026)", "source_url": "https://www.datacenterfrontier.com/press-releases/article/55397178/nvent-delivers-record-sales-and-eps-in-q2-2026", "source_title": "nVent Delivers Record Sales and EPS in Q2 2026 — DataCenterFrontier (nVent press release)", "source_date": "2026-07-31", "gaps": "The expansion location, capacity magnitude, and timing are not detailed in the release; liquid-cooling-specific revenue split not disclosed separately.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Repeated liquid cooling manufacturing expansions within one quarter of Q2 earnings plus a large guidance raise indicate sustained demand outpacing existing capacity", "metrics_note": "nVent press release as syndicated via DataCenterFrontier, July 31 2026; quote from CEO Beth Wozniak. Cross-check: Zacks reported the 25% EPS beat on the same date." }, { "id": "GENTHERM-MODINE-PT-COMBINATION-APPROVED-CLOSE-OCT-2026", "component": "Gentherm / Modine — Performance Technologies Combination Shareholder Approval", "category": "cooling", "subcategory": "Corporate Strategy", "claim_type": "acquisition", "claim": "Gentherm shareholders approved (99% of votes cast in favor of the share-issuance proposal) the Reverse Morris Trust combination with Modine's Performance Technologies business on September 10, 2026; the transaction is expected to close October 1, 2026. The combined company is projected at ~$2.6B day-one revenue, >12% EBITDA margins, and under 70% light-vehicle exposure — with Modine's data-center cooling (Airedale) retained in the continuing Modine business.", "confidence": "high", "status_trend": "closing_imminently", "period": "Approved 2026-09-10; expected close 2026-10-01", "source_url": "https://www.globenewswire.com/news-release/2026/09/10/3359875/0/en/gentherm-shareholders-approve-combination-with-modine-s-performance-technologies-business.html", "source_title": "Gentherm Shareholders Approve Combination with Modine's Performance Technologies Business — GlobeNewswire", "source_date": "2026-09-10", "gaps": "Which exact Modine unit (Airedale data center cooling vs. remaining heat-exchanger lines) exits with Gentherm's PT business vs. stays with Modine requires S-4 boundary review; deal terms not re-stated in this release.", "leadtime_weeks": null, "price_trend": "stable", "capacity_signal": "Restructuring of two major thermal/heat-exchanger players; Modine narrows focus to data-center-centric thermal, Gentherm gains power-generation heat-exchanger scale", "metrics_note": "Primary: Gentherm press release via GlobeNewswire, Sept 10 2026. MarketBeat Aug 17 2026 (JPMorgan conference) provides the $2.6B/2030-target context as secondary corroboration." }, { "id": "VRT-TOGNANA-CHILLER-CAPACITY-DOUBLED-2026", "component": "Vertiv — Tognana, Italy Campus: Chiller Production Capacity Doubling & Testing Lab", "category": "cooling", "subcategory": "Manufacturing Capacity", "claim_type": "capacity_expansion", "claim": "Vertiv announced investments to expand chiller manufacturing and integrated testing at its Tognana campus near Padua, Italy: regional chiller production capacity is expected to double by the end of 2026, and a new large-scale testing laboratory (for validating chillers integrated with liquid cooling under high-density loads and extreme temperatures) will complete in early 2027.", "confidence": "high", "status_trend": "capacity_doubling_emea", "period": "Announced July 21, 2026; completion end-2026 / early-2027", "source_url": "https://www.vertiv.com/en-emea/about/news-and-events/news-releases/2026/vertiv-expands-global-manufacturing-capacity-for-ai-ready-data-center-cooling-solutions/", "source_title": "Vertiv expands manufacturing and testing capacity at Tognana campus near Padua — Vertiv press release", "source_date": "2026-07-21", "gaps": "Absolute capacity (chillers/year or MW) not stated; dollar value of the investment not disclosed in the release.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Dedicated EMEA chiller capacity doubling signals high-density AI cooling demand in Europe; testing lab removes a deployment-verification bottleneck for complex thermal chains", "metrics_note": "Primary: Vertiv press release, July 21 2026 (CEO Gio Albertazzi and EMEA thermal VP Sam Bainborough quoted); corroborated by Manufacturing Digital and TechArena syndications." }, { "id": "VRT-COOLCHIP-CDU-2300-ROW-MANIFOLDS-EMEA-AVAILABILITY", "component": "Vertiv — CoolChip CDU 2300 & Fluid Network Row Manifolds EMEA Availability", "category": "cooling", "subcategory": "Products: CDU / Manifold", "claim_type": "product_status", "claim": "Vertiv made the CoolChip CDU 2300 (a liquid-to-liquid CDU delivering 2.3 MW of cooling capacity in a compact footprint) and CoolChip Fluid Network Row Manifolds available across EMEA. The CoolChip CDU system family spans 100 kW to 2.3 MW, supports direct-to-chip and rear-door heat-exchanger topologies, and its controller adapts temperature/flow to workload with redundancy and remote monitoring.", "confidence": "high", "status_trend": "product_availability_emea", "period": "Announced May 26, 2026", "source_url": "https://www.vertiv.com/en-emea/about/news-and-events/news-releases/2026/vertiv-expands-liquid-cooling-portfolio-in-emea-to-accelerate-ai-ready-data-centre-deployments/", "source_title": "Vertiv expands liquid cooling portfolio in EMEA to accelerate AI-ready data centre deployments — Vertiv press release", "source_date": "2026-05-26", "gaps": "No unit pricing or order/backlog figures disclosed; availability status in APAC/AMER for the 2300 model not covered by this release.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "2.3 MW single-CDU class supports multi-megawatt AI halls in a small footprint — reduces CDU count and floor space, a deployment-speed lever for operators", "metrics_note": "Primary: Vertiv EMEA press release, May 26 2026; products to be showcased at Datacloud Global Congress (Cannes, June 1-4 2026, booth 123)." }, { "id": "NVIDIA-RUBIN-100PCT-LIQUID-COOLED-45C-DSX", "component": "NVIDIA Rubin Generation — 100% Liquid-Cooled AI Infrastructure, 45°C Coolant & DSX Reference Design", "category": "cooling", "subcategory": "Thermal Technology / Reference Design", "claim_type": "industry_development", "claim": "NVIDIA's Rubin generation of AI infrastructure is described as the world's first to achieve 100% liquid cooling — every chip and networking component cooled in a closed liquid loop with no fans. The platform operates with facility water up to 45°C (113°F), and under the NVIDIA DSX AI-factory reference design, dry-cooler-based chiller-less designs can reduce facility cooling water use from ~2.6 million gallons/MW/year (conventional cooling towers) to near zero, up to ~100% reduction in favorable climates.", "confidence": "high", "status_trend": "new_industry_baseline_45c", "period": "Published June 21, 2026; Rubin ramping 2026-2027", "source_url": "https://blogs.nvidia.com/blog/liquid-cooling-ai-factories/", "source_title": "Hotter Than a Hot Tub: The 45°C Breakthrough to Cool AI's Biggest Machines — NVIDIA Blog", "source_date": "2026-06-21", "gaps": "Climate-conditional: 'near zero' water use applies only in favorable climates with dry-cooler designs; chillers may still be needed ~1% of the year in some regions. No per-rack coolant flow specs in this post.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "100% liquid-cooled reference architecture shifts all new Rubin builds to DLC; Motivair (Schneider) cited as a long-standing design partner on the roadmap", "metrics_note": "Primary: NVIDIA blog by Josh Parker, June 21 2026; quotes Ali Heydari (NVIDIA DC cooling director) and Richard Whitmore (Motivair president/CEO). Corroborated by Data Today and UBCT coverage of the 45°C water-use claim." }, { "id": "OCP-GLOBAL-SUMMIT-2026-OCT-12-15-SAN-JOSE", "component": "OCP Global Summit 2026 — Schedule & Venues", "category": "cooling", "subcategory": "Industry Events / Standards", "claim_type": "industry_development", "claim": "The 2026 OCP Global Summit ('Scaling Innovation for the AI Era') is scheduled October 12-15, 2026 in San Jose, California, with public registration open; starting in 2027 the summit moves to the Moscone Center in downtown San Francisco. OCP's liquid-cooling standardization work (UQD, cold plates, CDU interoperability) is tracked here, making the summit a key milestone for the next spec revisions.", "confidence": "medium", "status_trend": "event_upcoming_oct_2026", "period": "October 12-15, 2026", "source_url": "https://www.opencompute.org/summit/global-summit", "source_title": "2026 OCP Global Summit — Open Compute Project", "source_date": "2026-09-14", "gaps": "Specific liquid-cooling session schedule and any UQD/cold-plate spec milestones for the event were not enumerated on the landing page at access time.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Standards-body event cadence: final spec milestones announced here trigger commoditization pressure across cold-plate/connector/coolant supply chains", "metrics_note": "Live event landing page (accessed 2026-09-14; page is continuously updated, so source_date = access date); venue-move notice links to OCP blog post 'OCP Global Summit is moving to a new location'." }, { "id": "DELTA-PREFAB-AI-MODULAR-DC-COOLING-COMPUTEX-2026", "component": "Delta Electronics — Prefabricated AI Modular Data Center Solution (Cooling + Power)", "category": "cooling", "subcategory": "Products: Modular Systems", "claim_type": "product_status", "claim": "At COMPUTEX 2026 (Taipei, June 2, 2026), Delta launched a Prefabricated AI Modular Data Center Solution that, via factory pre-assembly and testing, reduces data center deployment time by up to 60% and significantly lowers PUE. Delta's prefabricated solutions integrate 800VDC In-Row Power with its industry-leading 3MW liquid cooling systems to address high-density AI-rack thermal challenges; the portfolio spans liquid and air cooling, power distribution for HVDC architectures, DCIM and racks for MW-scale sites.", "confidence": "medium", "status_trend": "product_launch", "period": "Announced June 2, 2026 (COMPUTEX 2026)", "source_url": "https://www.deltathailand.com/en/products-services-detail/10/235/Delta-Debuts-Prefabricated-AI-Modular-Data-Center-at-COMPUTEX-2026", "source_title": "Delta Debuts Prefabricated AI Modular Data Center at COMPUTEX 2026 — Delta", "source_date": "2026-06-02", "gaps": "Cooling-component breakdown (CDU capacity, manifold architecture) of the prefabricated solution not detailed in the announcement; customer/pilot information not disclosed.", "leadtime_weeks": null, "price_trend": null, "capacity_signal": "Prefabrication trend shortens time-to-power for AI sites; Delta positions as integrated power+thermal provider competing on deployment speed", "metrics_note": "Delta corporate news release (deltathailand.com syndication of Delta global announcement), June 2 2026; 'up to 60%' deployment-time reduction is Delta's own claim (footnoted in original)." }, { "id": "NVT-Q2-2026-RECORD-SALES-53PCT-LIQUID-COOLING-DEMAND", "component": "nVent Electric — Q2 2026 Record Sales, Data Center Growth >30pts, Raised FY Outlook", "category": "cooling", "subcategory": "Data Center Solutions", "claim_type": "revenue_record", "claim": "nVent Q2 2026: reported sales $1.5 billion, +53% YoY (+47% organic); reported EPS $1.32 (+103%), adjusted EPS $1.45 (+69%); operating income +92% to $301M. CEO stated data center growth and new products contributed more than 30 points to sales growth. Full-year 2026 guidance raised to +37-39% reported sales growth (+32-34% organic) and adjusted EPS $5.00-$5.10; Q3 2026 guidance +32-35% sales growth.", "confidence": "high", "status_trend": "record_quarter_guidance_raised", "period": "Q2 2026 (announced July 31, 2026)", "source_url": "https://investors.nvent.com/press-releases/press-release-details/2026/nVent-Electric-plc-Second-Quarter-2026-Financial-Results-Available-on-Companys-Website/default.aspx", "source_title": "nVent Electric plc Second Quarter 2026 Financial Results Available — nVent Investor Relations", "source_date": "2026-07-31", "gaps": "Data center segment dollar split not in the headline release; portion of growth attributable to liquid cooling vs. power/rack/enclosure lines not itemized.", "leadtime_weeks": null, "price_trend": "rising", "capacity_signal": "Guidance raised by ~11 points (26-28% → 37-39% reported) within one quarter, indicating demand is outpacing prior internal capacity planning", "metrics_note": "Primary: nVent investor-relations release pointer, July 31 2026; figures corroborated by DataCenterFrontier syndication of the same release and Zacks consensus-beat reporting." } ] } ] }