NVIDIA contributed Vera Rubin rack designs to OCP, updated MGX 1.2, and published open specifications with AMD, Google, Meta, and Microsoft for scaling AI factories.
NVIDIA contributed key NVIDIA Vera Rubin rack and compute tray designs to OCP through the open NVIDIA MGX architecture.
The updated MGX 1.2 specification extends the open architecture for AI factories with rack-scale blind-mate liquid cooling.
The architecture supports rack deployments up to 250kW while simplifying serviceability, deployment and infrastructure reuse.
NVIDIA joined AMD, Google and Meta to publish a new open specification that establishes that foundation.
At the Open Compute Project (OCP) Global Summit in the San Jose McEnery Convention Center, NVIDIA announced new partnerships and contributions aimed at creating a common foundation for next-generation AI infrastructure, according to the NVIDIA source story. NVIDIA contributed key NVIDIA Vera Rubin rack and compute tray designs to OCP through the open NVIDIA MGX architecture. The updated MGX 1.2 specification extends the open architecture for AI factories with rack-scale blind-mate liquid cooling, interoperable rack-scale interconnect infrastructure and a liquid-cooled 5,000A busbar designed for higher-power AI systems. The architecture supports rack deployments up to 250kW while simplifying serviceability, deployment and infrastructure reuse.
Editorial interpretation: These open contributions matter because they give system builders a common foundation for developing next-generation AI factory infrastructure, as stated in the source. The MGX 1.2 specification adds native 19-inch equipment support and higher-density rack designs, which could help partners deploy specialized infrastructure while benefiting from a shared ecosystem. The Vera CPU rack design specification extends the NVIDIA MGX open architecture beyond accelerated computing, providing a purpose-built platform for the growing agentic AI era, supporting tool calling, orchestration and sandboxed execution. This means AI agents could interact with applications, data and services in real time, based on the source description.
A reader can verify that NVIDIA joined AMD, Google and Meta to publish the Open Data Center Specification v0.7, which defines a unified blueprint for next-generation AI factory facility design, from power architecture to cooling infrastructure to physical layout. The OCP Grid Resilience Specification, published by Google, Meta, Microsoft and NVIDIA, establishes how large-scale AI infrastructure should behave during utility grid disturbances: staying connected, responding automatically and restoring to near-full load within less than two seconds. NVIDIA also expanded DSX OS with open source components for configuring networks, validating clusters, managing racks and provisioning bare metal, developed from NVIDIA’s experience running NVIDIA DGX Cloud.
The supplied feed entry is a concise summary; consult the canonical article for full context. The source story does not provide details on the timeline for when these specifications will be adopted by manufacturers or how quickly the open standards will translate into deployable infrastructure. The source also does not specify the full list of partners contributing to the third-generation MGX AI data center reference design through OCP’s Open DC Initiative, beyond noting NVIDIA joined industry leaders. The practical performance impact of the 250kW rack deployments and the 5,000A busbar in real-world AI factory operations is not quantified in the source material.
The MGX 1.2 specification, as described in the NVIDIA source story, extends the open architecture for AI factories with rack-scale blind-mate liquid cooling, interoperable rack-scale interconnect infrastructure and a liquid-cooled 5,000A busbar designed for higher-power AI systems. It also adds native 19-inch equipment support and higher-density rack designs, supporting rack deployments up to 250kW.
According to the NVIDIA source, NVIDIA joined AMD, Google and Meta to publish the Open Data Center Specification v0.7. The OCP Grid Resilience Specification was published by Google, Meta, Microsoft and NVIDIA. These open standards aim to provide a common target for manufacturers building AI factory infrastructure.