NVIDIA's first GB300 FP4 counter-submission still 1.64x worse perf per dollar than TPUv7
xennygrimmato_ · x · 2026-10-20
Per SemiAnalysis, after being shocked by last week's TPUv7 FP8 submission showing better performance per dollar, NVIDIA worked overtime on a first GB300 FP4 counter-submission.
It still falls short: across most of the end-to-end latency curve, the GB300 FP4 submission is 1.64x worse in performance per dollar than TPUv7 for external customers. With E2EL on the x-axis, TPUv7 holds a higher perf-per-$ curve across most of the Pareto frontier.
SemiAnalysis views this as NVIDIA's first attempt to close the gap and expects more counter-submissions to follow.
More from Infra
- Together Launches Post-Training Platform: Custom SFT & RL for Open-Weight Models — simran_s_arora · 2026-10-20
- RTX 5090 Laptop GPU hangs with Xid 8 under Linux — CUDA graphs bug still unfixed — SOC_FreeDiver · 2026-10-20
- Agent inference costs spiral as GPU supply tightens — devs compare cloud, self-hosted, decentralized options — aruthesa · 2026-10-20
- Taco: hardware-agnostic LLM compression with one hash across 7 backends — phoenix10701 · 2026-10-20
- Scaling past 1T params: routing, hybrid attention, and 1M-token contexts — philipkiely · 2026-10-20
- NVIDIA partners with Germany on industrial AI, backs four German startups — nvidia · 2026-10-20