Particle.news
Download on the App Store
3 ARTICLES
·
last mo.

Microsoft, NVIDIA Expand AI Partnership With Foundry GA as Azure Powers On Vera Rubin Systems

The collaboration focuses on moving enterprise AI from pilots to production through a single stack that reaches cloud, hybrid, and sovereign environments.

An employee standing between server racks at a Microsoft AI datacenter.
Image

Overview

  • Microsoft makes the next‑generation Foundry Agent Service and Observability generally available, refreshes the Foundry portal, and opens a Voice Live API public preview.
  • NVIDIA Nemotron open‑weight models are now available in Microsoft Foundry via NIM microservices, with examples including Nemotron Nano 9B v2 and Nemotron Super variants, and serverless APIs plus an expanded lineup planned later in 2026.
  • Azure is the first hyperscale cloud to power on NVIDIA Vera Rubin NVL72 in its labs, with rollout to liquid‑cooled Azure datacenters planned over the coming months.
  • Microsoft reports deploying hundreds of thousands of liquid‑cooled Grace Blackwell GPUs across its global datacenter footprint to support inference‑heavy, reasoning workloads.
  • Foundry Local and Azure Local extend advanced AI into sovereign and on‑premises environments with initial Vera Rubin support on Azure Local, and the companies release an open Azure Physical AI toolchain and deepen Fabric–Omniverse integration for digital‑twin and robotics workflows.