Microsoft announced several AI infrastructure and platform expansions at NVIDIA GTC 2026. Key highlights include: Foundry Agent Service and Observability reaching general availability for production-scale AI agents; NVIDIA Nemotron models now available through Microsoft Foundry; Azure becoming the first hyperscale cloud to power on NVIDIA's next-gen Vera Rubin NVL72 systems; and deeper integration between Microsoft Fabric and NVIDIA Omniverse for Physical AI and digital twin workflows. The announcements also cover sovereign/regulated environment support via Azure Local, a Voice Live API integration for multimodal agents, and a public Physical AI Toolchain GitHub repository connecting robotics workflows with Azure services.
Table of contents
What’s new at NVIDIA GTCFrom Frontier models to production-ready agentsScaling AI infrastructure for the world’s most demanding workloadsBringing AI into the physical worldFrom innovation to impactSort: