An open infrastructure push at OCP
At the Open Compute Project (OCP) Global Summit at the San Jose McEnery Convention Center, NVIDIA announced new partnerships and contributions. The company contributed its Vera Rubin rack and compute tray designs to OCP through the open NVIDIA MGX architecture, giving system builders a common foundation for next-generation AI factory infrastructure.
What MGX 1.2 brings
The updated MGX 1.2 specification extends the open architecture for AI factories:
- Rack-scale blind-mate liquid cooling
- Interoperable rack-scale interconnect infrastructure
- A liquid-cooled 5,000A busbar for higher-power AI systems
- Native 19-inch equipment support and higher-density rack designs
- Configurable I/O bays and fanless liquid-cooled compute trays
The architecture supports rack deployments up to 250 kW and aims to cut assembly, test and service times.
Vera CPU rack and partner announcements
NVIDIA is extending the MGX open architecture beyond accelerated computing with a Vera CPU rack design specification. The platform is built to support tool calling, orchestration and sandboxed execution for agentic AI workloads. The event also featured open AI factory and grid resilience standards with hyperscalers, new open source software components for NVIDIA DSX AI Factory, work with Meta on Spectrum-X Ethernet SuperNIC and Spectrum-6, Visionbay's 800V DC AI factory design, the USFP standard with Cisco, and Falcon Transport support with Google.
