NVIDIA Opens Vera Rubin Architecture to Scale AI

At the OCP Global Summit, NVIDIA announced that it is opening its Vera Rubin architecture to help the industry scale AI factories from megawatts to gigawatts. The move reflects a broader shift in AI infrastructure: as models grow larger and inference demands explode, data centers must expand far beyond the scale of traditional cloud computing. NVIDIA says open standards are essential for making that growth faster, more efficient, and more reliable. The Vera Rubin architecture is designed to support the enormous computing needs of AI factories, which combine dense compute, networking, storage, power, and cooling into systems built for continuous AI workloads. By opening the architecture, NVIDIA aims to encourage ecosystem adoption and collaboration, allowing hardware and software partners to build compatible solutions rather than relying on closed, proprietary designs. That could matter for cloud providers, enterprises, and governments racing to deploy AI data centers. Open foundations can reduce integration friction, speed up innovation, and give customers more flexibility as they scale. The megawatt-to-gigawatt framing highlights the physical reality of modern AI. Training and serving frontier models require not only advanced GPUs but also efficient power delivery, advanced cooling, high-speed interconnects, and orchestration software. If AI factories are to become critical infrastructure, they need common standards that many vendors can support. NVIDIA's announcement positions Vera Rubin as a platform for that next phase. The ultimate test will be adoption. If partners embrace the open architecture, it could accelerate the buildout of AI capacity and make large-scale deployments more modular. For now, the message is clear: the AI boom is becoming an infrastructure race, and NVIDIA wants Vera Rubin to be a foundational layer.