Open Compute Project Expands Open Data Center Ecosystem for AI

Open Compute Project Foundation announced in a press release a set of new technical contributions, projects, and partnerships to advance its Open Data Center Ecosystem for AI. The initiative expands work announced in 2025 and covers both IT and physical infrastructure for AI data centers.

The newly approved contributions include design principles for machine learning infrastructure, low voltage direct current distribution architectures, safety guidelines for energy storage systems, telemetry integration methods, and Ethernet-based networking requirements for AI clusters. The update also introduces the Open Rack Wide specification and new reference architectures for AI clusters, along with the Foundation Chiplet System Architecture.

New projects focus on AI data center facilities, power distribution, networking, and the AI Computing Continuum. These projects aim to improve energy efficiency, increase power density, and enhance interoperability in large-scale AI deployments. The AI Computing Continuum project seeks to define modular infrastructure standards that can be applied across hyperscale and enterprise environments.

The foundation also formed alliances with Electric Power Research Institute (EPRI) and IOWN Global Forum. The collaboration with EPRI focuses on integrating data centers as flexible grid resources, while the partnership with IOWN Global Forum will develop a roadmap for distributed, high-bandwidth, low-latency compute and network infrastructure.