Subscribe

Vera Rubin architecture contributions to OCP (MGX 1.2)

At the OCP Global Summit NVIDIA contributed Vera Rubin rack and compute tray designs to the Open Compute Project through MGX 1.2, with Cisco, Google, Meta, Microsoft and Visionbay adopting parts of it.

NVIDIA contributed Vera Rubin rack and compute tray designs through the open MGX architecture. The MGX 1.2 specification adds rack-scale blind-mate liquid cooling, interoperable interconnect infrastructure and a liquid-cooled 5,000A busbar, and supports racks up to 250kW. NVIDIA also published 800 VDC power reference designs and described a Vera CPU rack for agentic AI workloads.

Date
Monday 12 October 2026
Lab
NVIDIA
Kind
infra
Access
open standard

Figures

MeasureValueMeasured by
Maximum rack power250kW
rack deployments supported by MGX 1.2
company
Busbar current5,000A
liquid-cooled busbar
company
Switch bandwidth102.4Tb/s
Spectrum-6 switches Meta is adopting
company

Partner items: Meta adopting Spectrum-6, Google working on Falcon Transport support for ConnectX-10, Cisco on USFP optical transceivers, Visionbay planning 800 VDC in its K-1 AI factory platform. No availability dates given.

Sources

  1. blogs.nvidia.com/blog/ocp-global-summit-vera-rubin-scale-ai-factories/

This record was partly confirmed: some claims could not be checked on 12 October 2026. How we check

Read the daily brief for 12 October 2026