The global technology corporation Microsoft and NVIDIA, an American technology company, have announced an expanded partnership at NVIDIA GTC to integrate advanced hardware, software, and cloud infrastructure for production-ready AI. Central to this collaboration is Microsoft Foundry, which now serves as an operating system for building and operating AI agents at enterprise scale. This includes the general availability of the Foundry Agent Service and Observability in the Foundry Control Plane, alongside the addition of NVIDIA Nemotron models to the platform’s model catalog.
In their statement, they say that the partnership also introduces significant advancements in AI infrastructure optimized for inference-heavy and reasoning-based workloads, with Microsoft being the first hyperscale cloud provider to power on the next-generation NVIDIA Vera Rubin NVL72 systems, which will be rolled out to liquid-cooled Azure datacenters in the coming months. Additionally, Microsoft will extend these capabilities to sovereign and regulated environments through initial support for the Vera Rubin platform on Azure Local, managed via Azure Arc and Foundry Local.
Beyond digital applications, the companies are collaborating on Physical AI systems to bridge simulation and real-world operations. This includes the introduction of an Azure Physical AI Toolchain and deeper integration between Microsoft Fabric and NVIDIA Omniverse libraries. These tools allow organizations to connect live operational data with digital twins to coordinate AI-driven actions across physical facilities.
