Synthesis history

The main addition this pass is institutional corroboration of the power constraint theme: the IEA [^14030], WEF [^22062], Data Center Knowledge [^29565], and KTS Law [^29563] have each independently documented grid conn…

Three significant new themes entered this pass. First, SemiAnalysis launched an H100 1-Click Rental Index documenting a 40% price increase in GPU rentals since October 2025, characterizing the move as structurally drive…

New items (28612–28615) add nuance to the NVIDIA-Intel foundry story: whereas last pass multiple publications framed the Feynman GPU as confirmed to use Intel Foundry, this pass includes sources characterizing the relat…

The NVIDIA-Intel foundry story has acquired a specific product name: multiple publications now report NVIDIA's 'Feynman' GPU architecture will use Intel Foundry for some components [^27840][^27841][^27842][^27843], addi…

The substantive new development is TrendForce's report that NVIDIA is evaluating Intel's 18A process node for a multi-die GPU design [^27751], extending the Intel foundry narrative beyond the confirmed Google TPU deal t…

The Google-Intel TPU foundry deal, which appeared in the prior pass as a single-source report from The Information, has been confirmed across multiple independent publications including Quartz, Mobile World Live, and Fi…

Three substantive developments this pass: SK Hynix and NVIDIA formalized a multi-year HBM co-development partnership covering Vera Rubin and future platforms [^26485], Bernstein projected HBM4 pricing nearly doubling to…

All nine new items this pass are social media posts (Twitter/X, Instagram, LinkedIn) amplifying the Micron HBM thesis among retail investors — including a call for $MU to reach $2,000 [^26461] and restatements that HBM …

Seven new items this pass [^25973][^25974][^25975][^25976][^25977][^25978][^25979] all relate to the MLCC shortage angle. None carried extractable claims, but their titles and sources — spanning Digitimes, EE News Europ…

Two substantive additions this pass. SemiAnalysis extended its supply chain analysis to MLCCs, flagging each AI server rack's requirement for tens of thousands of sub-$1 passive components as an overlooked constraint no…

New items this pass are thin. The one substantive addition is confirmation that Intel's Crescent Island inference chip does not use HBM [^25082], which deepens the existing observation that Intel is competing entirely o…

Two new angles this pass. SemiAnalysis corrected a potential misread of AMD's competitive position: rack photos of the MI455 and VR200 at CoreWeave and Microsoft are engineering samples with incomplete software stacks a…

The most significant new development is SK Hynix's disclosure that even doubling wafer capacity within five years will not relieve memory supply tightness until at least 2030 [^23422] — a substantial extension of the sh…

The most significant new material this pass is quantitative confirmation of the shortage thesis from TSMC's own financials: +41% YoY growth with all-time-high margins [^23115], foundry industry at a record $48.8 billion…

The most significant new development is the extension of the TSMC shortage horizon beyond 2027 [^23032], pushing the supply constraint from a near-term bottleneck into a potentially multi-year structural regime — a mate…

The most significant new development is empirical confirmation of the HBM shortage thesis: sold-out order books at both SK Hynix and Micron through 2026 [^22728][^20307] move SemiAnalysis's central claim from forecast t…

The primary constraint on frontier AI hardware has migrated from GPU availability to High Bandwidth Memory (HBM) wafer supply, with CoWoS packaging no longer the binding bottleneck [^22548][^22656]. SemiAnalysis resear…