NVIDIA and Microsoft Expand AI Partnership From Azure Data Centers to Windows PCs


nvidia-microsoft
Image credit: NVIDIA | Microsoft

NVIDIA and Microsoft are taking their AI partnership beyond Azure data centers by bringing accelerated computing to Windows PCs, industrial systems, and enterprise infrastructure. The announcement, made at Microsoft Ignite 2026 yesterday, covers new hardware, open AI models, and tools for businesses. It also puts NVIDIA’s latest Vera Rubin systems at the center of Microsoft’s expanding AI infrastructure plans.

NVIDIA Vera Rubin systems are coming to Azure

The announcement centers around the expansion of NVIDIA Vera Rubin NVL72 systems across Azure data centers. These systems are designed to handle demanding AI reasoning, agent-based applications, and large-scale model training and inference.

Microsoft and NVIDIA are also working to extend confidential computing across bare-metal servers and rack-scale AI infrastructure. This technology helps protect sensitive models and data while they are being processed. NVIDIA’s attestation tools also let customers verify the computing environment before sharing sensitive information or encryption keys.

The companies are additionally working on power efficiency through NVIDIA MaxLPS. Microsoft says its Azure infrastructure scaled to 10,240 NVIDIA GPUs using GB200 NVL72 systems in MLPerf Training v6.1. Azure’s GB300 NVL72 systems also led cloud submissions for DeepSeek-R1 interactive throughput in MLPerf Inference v6.1, according to Signal65.

The partnership is also expanding NVIDIA Nemotron models through Microsoft Foundry. Businesses can customize these models using Microsoft’s Programmed Training service and deploy their own model weights through Foundry Managed Compute.

AI is moving beyond Azure data centers

NVIDIA and Microsoft are also expanding AI deployment options through Azure Local, Jetson platforms, and Windows PCs. Azure Local is gaining support for NVIDIA RTX PRO 6000 Blackwell GPUs, alongside Jetson Thor and Jetson Orin platforms for industrial AI workloads.

Speaking of Windows, the companies will showcase local large language model experiences on NVIDIA RTX Spark PCs at Ignite. This could give developers more options to run AI applications directly on their devices. Microsoft is also bringing GPU-accelerated Windows 365 Cloud PCs Direct Stream capabilities. The technology supports up to two 4K streams at 60fps, with a public preview planned for early 2027.

The collaboration also extends to scientific research through NVIDIA BioNeMo and ALCHEMI toolkits in Microsoft Discovery. Meanwhile, Microsoft AI is using NVIDIA Dynamo to accelerate reinforcement learning during model training.

More about the topics: AI, microsoft, nvidia

Readers help support Windows Report. We may get a commission if you buy through our links. Tooltip Icon

Read our disclosure page to find out how can you help Windows Report sustain the editorial team. Read more

User forum

0 messages