TSMC denies 3D DRAM plans at Nvidia's request, stays pure-play logic foundry
rohanpaul_ai · x · 2026-11-08
TSMC has publicly denied a report that it began drafting 3D DRAM plans at the request of its largest customer, Nvidia, saying it will remain a pure-play logic foundry.
- Context: AI demand has drained DRAM supply, making memory the bottleneck that decides whether accelerators ship — giving Nvidia reason to look for new supply
- 3D DRAM stacks memory cells vertically to raise capacity per wafer, but even Micron plans at least 4 more planar nodes before adopting it, meaning a TSMC entry couldn't ease shortages for at least 5 years — and TSMC lacks a competitive planar DRAM process to build from
- TSMC already profits from memory without making it: it fabricates HBM base dies and bonds memory to logic via SoIC hybrid bonding
- Separately, a Korean report claims Nvidia wants Samsung, SK hynix and Micron to build its custom NVHBM base dies on TSMC's 3nm process, giving TSMC memory revenue without DRAM's price swings
More from Infra
- 'GPU homelessness' on the rise: compute owners hunt for colo space with 10G and reliable power — markjeffrey · 2026-11-08
- In-flight Wi-Fi speedtest shows 122Mbps down, "best flight internet" yet — JoeJustice · 2026-11-08
- Data centers are scaling so fast that electricity and grid capacity are now the biggest bottleneck for superintelligence — XFreeze · 2026-11-08
- Running 50-100 Agents, Dev Weighs Local Decision Models to Cut Latency — f1yingbanana · 2026-11-08
- TSMC denies developing 3D DRAM at Nvidia's request, staying a pure logic foundry — rohanpaul_ai · 2026-11-08
- Why is everyone sleeping on Modal live inference endpoints for agents? — vboykis · 2026-11-08