Open Compute Project Launches AI Portal for Next-Gen Clusters

The Open Compute Project Foundation has launched a new AI portal on its marketplace, aimed at accelerating the deployment of next-generation AI clusters. Announced in a press release, this portal will serve as a comprehensive resource for AI cluster designers and builders, offering access to the latest AI infrastructure products, white papers, and best practice documents.

The initiative addresses the challenges faced by hyperscale operators, such as compute density and power distribution, by developing open standards to overcome these bottlenecks. The OCP community, consisting of over 400 corporate members and 6,000 engineers, is actively working on standardizing silicon, power, cooling, and interconnects to support the growth of AI infrastructure.

In conjunction with the portal launch, Meta Platforms, Inc. has contributed its Catalina AI Compute Shelf specification, designed to support high-density AI systems with NVIDIA GB200. This complements previous contributions by Nvidia, enhancing the collaborative efforts within the OCP community to develop scalable AI infrastructure solutions.