Nvidia · Compute · Blackwell · Datacenter Dynamics
Nvidia opens up Vera Rubin rack and tray designs to OCP
Compiled by KHAO Editorial — aggregated from 1 source. See llms.txt for citation guidance.
◌ Single Source
Nvidia is contributing its Vera Rubin rack and tray designs to the Open Compute Project (OCP), a move it claims will foster open standards.
Key facts
- The Vera Rubin update ties to version 1.2 of its modular reference design standard, MGX
- Nvidia is contributing its Vera Rubin rack and tray designs to the Open Compute Project (OCP), a move it claims will foster open standards
- A separate spec extends the MGX beyond GPUs to cover the new Vera CPU rack, which Nvidia revealed has started shipping
- Included in the tech giant’s infra gift are cable-and fan-free liquid-cooled compute trays, interoperable rack-scale interconnect infrastructure, and a liquid-cooled 5,000 amp busbar built
Summary
Included in the tech giant’s infra gift are cable-and fan-free liquid-cooled compute trays, interoperable rack-scale interconnect infrastructure, and a liquid-cooled 5,000 amp busbar built for higher-power systems. The Vera Rubin update ties to version 1.2 of its modular reference design standard, MGX. This latest effort doubles down on liquid cooling, featuring connector solutions for inside trays that allow for cold plates to be swapped out more easily without any unwanted spillage. “The bottleneck for AI factory buildouts is how fast we can manufacture and deploy,” Joe DeLaere, Nvidia’s product lead for AI infrastructure solutions, said during a press briefing.