Industry·
NVIDIA and Microsoft Extend Accelerated Computing from Azure to Windows PCs
At Microsoft Ignite, NVIDIA and Microsoft unveiled deeper integration of Vera Rubin NVL72 systems on Azure, new confidential‑computing protections, and tools that let enterprises customize open models and run AI workloads on local Windows PCs.

NVIDIA and Microsoft are deepening their full‑stack partnership to let organizations build, customize, and deploy AI workloads from the cloud down to the desktop.
At Microsoft Ignite the companies announced that Azure now hosts Vera Rubin NVL72 systems, delivering rack‑scale performance for reasoning, agentic applications, and large‑model training. Recent MLPerf runs show Azure scaling to more than 10,000 GPUs on the DeepSeek‑V3 model and leading cloud inference throughput on DeepSeek‑R1.
The collaboration also extends confidential computing to bare‑metal and rack‑level infrastructure, using NVIDIA attestation so customers can verify the environment before loading proprietary models or data. Joint work on MaxLPS aims to squeeze more performance out of every watt in the data center.
On the software side, Microsoft Foundry adds serverless Programmed Training and Managed Compute, letting enterprises fine‑tune open models such as Nemotron and deploy custom weights without managing the underlying hardware. Early adopter EY reported a 55 % accuracy gain and lower cost after adapting Nemotron for purchase‑record classification.
Why it matters for GPU and AI infrastructure
By unifying cloud‑scale Vera Rubin hardware, confidential‑compute guarantees, and a managed customization pipeline, the partnership reduces the friction of moving models from training to edge deployment. For GPU buyers and AI‑ops teams this means a clearer path to right‑size infrastructure, protect IP, and accelerate time‑to‑value across hybrid environments.
- aigpu
- ai gpu
- ai gpu cloud
- aigpu dubai
- nvidia
- microsoft
- azure
- vera rubin
By AiGpu Editorial · Editorial rewrite based on public reporting (NVIDIA Blog)
← All articles