NVIDIA and Microsoft widen accelerated computing from Azure to on-premises, edge and Windows PCs
The partnership now reaches Azure Local, edge hardware and Windows 365 Cloud PCs, not just cloud data centers.

NVIDIA and Microsoft announced at Microsoft Ignite that they are widening their joint work, from Azure data center infrastructure to on-premises, edge and Windows PC deployments. Azure now runs NVIDIA Vera Rubin NVL72 systems, and the companies are extending confidential computing and power-efficiency tooling to rack-scale AI hardware. Developers can customize and deploy NVIDIA Nemotron 3.5 open models through Microsoft Foundry, where EY reported a 55% accuracy gain and 15x lower cost than proprietary frontier models on a corporate purchasing classification task. Azure Local is gaining support for NVIDIA RTX PRO 6000 Blackwell GPUs and Jetson Thor and Orin edge platforms, while Windows 365 Cloud PCs are getting Direct Stream, which is in limited preview with a public preview planned for early 2027.
Why it matters
The work now spans cloud, on-premises and edge hardware, so accelerated AI computing is reaching more places than the data center. For developers, open Nemotron models can be adapted through Microsoft Foundry, and EY's reported results suggest open models can compete on accuracy and cost with proprietary ones on specific tasks. Windows 365 Cloud PCs are getting Direct Stream, though it is still in limited preview with public availability planned for early 2027, so its effect on users is not yet settled.
Matters or hype?
Does this story matter, or is it hype? Vote before you see what everyone else said.
Sources · 1
- NVIDIA Blog ↗ Primary source