NVIDIA's Vera Rubin NVL72 could cut GPU needs 75% for training giant MoE models
teortaxesTex · x · 2026-10-26
Per a wallstengine post, NVIDIA's Vera Rubin NVL72 could train large mixture-of-experts models with 75% fewer GPUs than Grace Blackwell: training a 10T-parameter model on 100T tokens in one month would take roughly 16,000 Rubin GPUs versus 64,000 GB200s. Commentator teortaxesTex notes the setup is somewhat wasteful and Rubin will cost more, but argues the efficiency jump will significantly accelerate AI development.
More from Infra
- Anthropic S-1 reveals 48 cents of compute cost per dollar of Q2 revenue — nmasc_ · 2026-10-26
- Persistent AI joins UR network to route private AI inference with encrypted, IP-hidden traffic — 0xSammy · 2026-10-26
- Anthropic's disclosed compute commitments: $114.5B Google, $161.2B Broadcom top the list — BenBajarin · 2026-10-26
- SpaceX strikes $40B deal with Apollo Global to lease AI chips and data center hardware — Polymarket · 2026-10-26
- SpaceX AI revenue up 830%, Claude designs 318K antivenom proteins in 64 hours — PeterDiamandis · 2026-10-26
- Meta plans up to $95 billion stock sale to fund AI buildout — The Decoder · 2026-10-26