CoreWeave says Vera Rubin NVL72 delivers 10x better tokens per megawatt
mark_k · x · 2026-07-23
CoreWeave says NVIDIA Vera Rubin NVL72 delivered about 10x more tokens per megawatt than GB200 Blackwell NVL72 on DeepSeek R1.
- The post cites CoreWeave’s first measured silicon results for Vera Rubin NVL72.
- It claims similar per-user responsiveness while cutting cost per token dramatically.
- The chart shows roughly 800,000 TPS/MW for Vera Rubin versus 80,000 TPS/MW for GB200 NVL72, framed as a 10x improvement.
- The implication is that this kind of efficiency jump could make large-scale, always-on reasoning and agent workloads more economical.
Related event: NVIDIA Launches Vera Rubin with 10x Energy Efficiency(6 posts)→
More from Infra
- Engram's random reads don't suit SSDs; CPU-memory over NVLink could serve all 72 GPUs — bookwormengr · 2026-09-11
- 80% of the DIY LLM inference hype posters have already quit — it's brutally hard systems work — abhijithneil · 2026-09-11
- Hugging Face's Ultra Scale Playbook: a free book on training LLMs on GPU clusters — mdancho84 · 2026-09-11
- Is inference latency becoming the biggest bottleneck for production AI agents? — Euphoric_Sea632 · 2026-09-11
- LLM Serving Metrics Thread: Why TPOT and Uptime Make or Break User Experience — abhijithneil · 2026-09-11
- PlanetScale launches sharded Postgres: 768 servers acting as one, 1PB scale — dhruv2038 · 2026-09-11