Ayar Labs is making its co-packed optics (CPO) technologies available to hyperscalers and other vendors to build optically connected AI infrastructure through Nvidia’s NVLink Fusion platform.
Ayar Labs has also joined the ecosystem as a partner.
Ayar Labs’ CPO technology is designed to scale limits of connectivity in AI data centers. The technology would expand design headroom as bandwidth requirements grow and electrical constraints tighten.
The company said that AI factories are scaling larger GPU counts and data movement and power are only turning into critical bottlenecks. Now these companies are looking to improve interconnectivity by putting COP in architectures that are able to scale beyond the limits of copper.
“AI infrastructure is being co-designed from the ground up, and customers need more options to scale performance efficiently as bandwidth continues to rise,” said Mark Wade, CEO at Ayar Labs. “By joining the NVIDIA NVLink Fusion ecosystem, we’re introducing co-packaged optics as a foundational building block for customers deploying heterogeneous compute in NVIDIA AI factories.”
NVLink Fusion is a path toward integrating custom CPUs and XPUs into Nvidia’s rack-scale architecture and ecosystem. The goal is to allow users to accelerate system concepts to deployment at scale.
The technologies will support next generation AI factories and data centers, Ayar Labs said.
