Lambda Boosts AI Factory Resiliency and Efficiency With NVIDIA Photonics
Reported by SDxCentral and NVIDIA2 sources ↓
Lambda is using NVIDIA Photonics co-packaged optics switches to connect GB300 NVL72 racks with more than 10,000 GPUs. NVIDIA reports lower latency in tests, with throughput matching conventional pluggable optics.
Key points
- NVIDIA reports latency fell about 4% at rack scale and up to about 10% at pod scale in its tests
- NVIDIA says throughput matched conventional pluggable optics across rack, scalable unit and pod configurations
- Lambda reports median optics temperatures of 36°C, compared with 51°C for conventional optics
- Lambda and NVIDIA say the switches provide up to 15 times higher reliability than conventional optics
Sources
Details
SDxCentral reports the deployment is at Lambda’s 142-megawatt campus in Quebec, Canada, and uses NVIDIA Quantum-X Q3450-LD switches. Each switch has 144 ports running at 800Gb/s and 18 removable external light sources.
NVIDIA says its tests found latency reductions of about 4% at rack scale and up to about 10% at pod scale, while throughput matched conventional pluggable optics across the tested configurations. Lambda reports a median optics temperature of 36°C, versus 51°C with conventional optics; SDxCentral says Lambda linked the lower temperature to up to 15 times higher reliability.
NVIDIA says the switches use optical engines packaged alongside the switch chip, with field-replaceable external light sources. It says the design uses four times fewer lasers and provides up to five times better power efficiency than conventional pluggable optics.
Written from coverage as of
Images
- NVIDIA
- SDxCentral
Stocks in this story
Change since open: 20 Nov to 21 Nov 2026
Last refresh