Lambda runs Nvidia CPO switches on 10,000+ GPU GB300 NVL72 cluster, cutting latency up to ~10%
Beth_Kindig · x · 2026-11-22
Nvidia ($NVDA) says cloud startup Lambda is using its CPO (co-packaged optics) switches to connect a GB300 NVL72 cluster with more than 10,000 GPUs. The company reports network latency decreased by roughly 4% at rack scale and 10% at pod scale — an early production deployment of CPO switching in large-scale AI training infrastructure.
More from Infra
- Running Inkling 2 Mini locally on Mac: 20GB model, ~16 tok/s on M4 Pro — jacob_posel · 2026-11-22
- A Hands-On GPU Profiling Starter Path: Time PyTorch Attention with Stanford CS336 Assignment 2 — stanfordnlp · 2026-11-22
- 42 Sellers, 5x Price Spread: Open Models Enrich Big Inference Providers, Not Startups — AccBalanced · 2026-11-22
- Inkling 2 Mini tops efficiency charts: 26B looped transformer at $25.80 per correct answer — TheZachMueller · 2026-11-22
- 425MB Qwen3.5-0.8B decision model on CPU: custom 4-bit quant loses just 0.18pp, 2x faster — Danmoreng · 2026-11-22
- Gemini 4 Neon autonomously writes GPU kernels, 5.9x faster prefill on AMD hardware — sebkrier · 2026-11-22