AI silicon demand at 110-120% of leading-edge wafer capacity, Bajarin argues all foundries win
BenBajarin · x · 2026-10-07
Analyst Ben Bajarin reiterates that all viable foundries will be used, citing his July report "The GPU Tsunami."
- Leading-edge wafer demand is running at roughly 110-120% of available capacity, with AI accelerators, custom ASICs, mobile, and CPUs competing for the same nodes
- Samsung and Intel matter without displacing TSMC—they fill in where TSMC is constrained, expensive, or strategically concentrated
- Foundry is shifting from wafer capacity to platform capacity: customers now buy a full path to deployable AI silicon spanning packaging, memory integration, qualification, timing, and geography
- Chiplets mark the shift: under 1% of total semiconductor units but possibly 10-20% of HPC/AI processors
A follow-up report argues wafer fab equipment spending will surge further.
More from Venture
- a16z maps consumer AI: winners in search, creative tools, productivity; gaps everywhere else — omooretweets · 2026-10-07
- Earthling VC doubles down on bio with 7th deal of 2026, likening it to robotics in 2023 — arian_ghashghai · 2026-10-07
- Coinbase Institute paper: AI agents are the internet's second customer, and stablecoins fit micro-payments — kleffew94 · 2026-10-07
- ARIA's £50M Scaling Trust program reveals first 12 teams securing AI agent coordination — DavideCrapis · 2026-10-07
- AI chip testing is the new bottleneck: winstek approves NT$13.4B capex on NT$2.7B revenue — tengyanAI · 2026-10-07
- Another B2B SaaS Crushed by Google's January Update: 100+ Sites Full of Listicles Hit — lilyraynyc · 2026-10-07