NVIDIA opens Vera Rubin rack and software stack for AI factory deployments
The company is widening the shared hardware and software base for large AI infrastructure, with new MGX and DSX pieces aimed at faster deployment.
NVIDIA used the OCP Global Summit to open up more of its Vera Rubin AI factory stack, adding new rack, cooling and software building blocks meant to make large-scale AI infrastructure easier to standardize and deploy. The company contributed Vera Rubin rack and compute tray designs through its MGX architecture, and said the updated MGX 1.2 spec supports blind-mate liquid cooling, interoperable rack-scale interconnects, a 5,000A busbar and rack deployments up to 250kW. It also released open source DSX AI Factory software components for network configuration, cluster readiness checks, rack management and bare-metal provisioning so providers can automate more of the path from hardware to production clusters. NVIDIA said it and partners, including hyperscalers, Meta, Cisco and Google, also announced open standards and networking additions across AI factories, grid resilience, USFP and Falcon Transport support.
Why it matters
For cloud providers and other large operators, the shift means more of the AI factory stack is being defined around common components and open software. That can make it easier to standardize rack design, power, cooling and cluster setup across sites, while partner-backed networking updates push the ecosystem toward more interoperable infrastructure.
Keep or strike?
Does this story matter, or is it hype? Mark it before you see what everyone else did.
Sources
- NVIDIA Blog