NVIDIA Opens Vera Rubin Architecture to Open Compute Project, Tackling AI Factory Scaling Bottlenecks

NVIDIA Addresses AI Factory Scaling Through Open Standards

NVIDIA announced at the OCP Global Summit in San Jose that it's contributing key Vera Rubin rack and compute tray designs to the Open Compute Project through its NVIDIA MGX architecture1

1

2

2

. The move gives system builders common blueprints for developing next-generation AI factory infrastructure as the industry faces manufacturing and deployment bottlenecks. "The real bottleneck for the industry right now is how fast the world can manufacture and deploy AI factories," said AI Infrastructure Solutions Lead Joe DeLaere2

2

. Ian Buck, vice president of hyperscale and high-performance computing at NVIDIA, discussed how open infrastructure combined with NVIDIA's extreme codesign is accelerating deployment while improving power efficiency and tokens-per-watt performance1

1

.

Vera Rubin Architecture Advances With MGX 1.2 Specification

The updated MGX 1.2 specification extends the NVIDIA MGX 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 systems1

1

. The Vera Rubin rack designs feature compute trays stripped of cables, hoses, and fans—everything is liquid-cooled and modular, reducing tray assembly times from hours to minutes and speeding up overall rack build times3

3

. Combined with open liquid-cooling technologies such as MQD (MGX Quick Disconnect) and MQDB (blind mating), the effort helps turn proven AI factory designs into industry standards that others can adopt and manufacture1

1

3

3

. The architecture supports rack deployments up to 250kW while simplifying serviceability and infrastructure reuse1

1

.

Vera CPU Rack Design Targets Agentic AI Era

The Vera CPU rack design specification extends the NVIDIA MGX open architecture beyond accelerated computing to support the growing agentic AI era1

1

. While Vera Rubin NVL72 serves as the foundation for large-scale training and inference, the Vera CPU rack is designed for tool calling, orchestration, and sandboxed execution—helping AI agents interact with applications, data, and services in real time1

1

. "Between model calls, agents need to do orchestration. They need to execute code. They need to process data, run evaluations," DeLaera explained2

2

. NVIDIA claims its Vera CPU delivers around 2x the performance of its previous Grace GPU generation on several workload benchmarks, with Meta measuring 2x performance for Hyperscale Cloud using DCPerf and the Texas Advanced Computing Center seeing a 2x uplift on HPC GoldenSuite3

3

.

Strategic Partnerships With Meta, Google, and Cisco Drive Networking Advances

Meta is adopting NVIDIA Spectrum-6 Ethernet switches for its next-generation AI factory network, which will connect directly to Vera Rubin deployments2

2

3

3

. Spectrum-6 is NVIDIA's 102.4 terabit-per-second Ethernet switching chip designed to handle massive traffic in fabric networks2

2

. With multi-plane networks, when a component fails, the SuperNIC detects it in hardware and reroutes traffic in less than 3 milliseconds—30 times faster than software-based alternatives3

3

. NVIDIA is adding support for Google's Falcon Transport to its Spectrum-X Ethernet networking ecosystem, beginning with future ConnectX-10 SuperNICs and extending to BlueField DPUs1

1

2

2

. NVIDIA and Cisco jointly announced the Ultra-Dense Small Form Factor Pluggable (USFP) standard, the industry's first module for 400Gb/s-based networking fabrics1

1

2

2

. The USFP design offers four times the bandwidth density of the existing 200G OSFP form factor with a 1.3x improvement in power efficiency2

2

.

Open Source Software Components Expand DSX AI Factory Platform

NVIDIA expanded its DSX OS with open source components for configuring networks, validating clusters, managing racks, and provisioning bare metal1

1

. The DSX AI factory platform applies extreme codesign to unify computing, networking, power, cooling, facilities, and software—helping AI factories operate as a single optimized system that maximizes AI output within available infrastructure constraints1

1

. At the center of this unified system is DSX OS, the open operating layer that turns the DSX architecture into coordinated, repeatable operations at scale1

1

. Visionbay unveiled plans for K-1, a next-generation AI factory platform designed around an 800 VDC architecture built with future Vera Rubin deployments in mind3

3

. The project combines NVIDIA's 800 VDC reference architecture with power and thermal technologies from Vertiv to create a scalable foundation for high-density AI infrastructure, addressing the challenge of rising rack power densities as AI workloads expand3

3

.