Lambda, the Superintelligence Cloud, today announced the world’s first large-scale deployment of co-packaged optics (CPO) networking. The company now connects 10,000+ NVIDIA Blackwell Ultra GPUs through NVIDIA Photonics networking switches at its Quebec campus.
Lambda first announced its early adoption of NVIDIA CPO networking in November 2025. Since then, the company has benchmarked the technology against conventional optics. The tests show lower latency and improved signal stability at scale.
As co-packaged optics AI clusters grow, networking infrastructure becomes increasingly critical. It must connect thousands of GPUs efficiently. Conventional pluggable optics convert electrical signals to light at the switch’s front panel. CPO works differently. It places optical engines alongside the switch ASIC. This shortens the electrical path from inches to millimeters. As a result, signal loss drops. Moreover, the design eliminates the need for retimer chips. It also increases network resiliency.
“Networking has to keep pace with the demands of increasingly large AI clusters. Bringing co-packaged optics into a deployment of this scale is an important step toward building the efficient and resilient infrastructure needed to support the next generation of AI,” said Dave Ward, President of Cloud Services at Lambda.
NVIDIA Photonics Switches Deliver Lower Latency and Cooler Optics
The deployment uses NVIDIA Photonics switches. Each switch features 144 800Gb/s ports, liquid cooling, and 18 removable external light sources. Lambda ran benchmarks across rack, scalable unit, and pod configurations. In every test, CPO maintained identical throughput. Additionally, it delivered more stable signal integrity than conventional pluggable optics.
co-packaged optics also delivered lower latency at every scale. Notably, the gap grew from 3.8 percent at rack to 10.1 percent at pod.
The CPO optics ran cooler as well. They operated at a lower median temperature of 36°C. In comparison, conventional pluggable optics reached 51°C. Cooler, more stable optics drive up to 15x higher reliability. Consequently, large training jobs keep running.
“In an AI factory, the network is a business accelerator. Every interruption stalls the job, and every watt counts. NVIDIA Photonics cuts network power, improves resiliency and delivers up to 15x higher reliability, so Lambda’s customers get more productive GPU hours from their investment,” said Gilad Shainer, Senior Vice President Networking at NVIDIA.
Explore IT Tech News for the latest advancements in Information Technology & insightful updates from industry experts!
News Source: Businesswire.com