SemisDecoded
Nvidia · Business ·

Nvidia and Microsoft expand Azure, Foundry, edge and Windows PC AI offerings at Ignite

Nvidia and Microsoft announced a set of expansions at Microsoft Ignite. Nvidia Vera Rubin NVL72 systems (rack-sized AI computers) are running in Azure data centers in a preview. Nvidia's Nemotron 3.5 open AI models are being added to Microsoft Foundry, and Azure Local, Microsoft's service for running Azure on a customer's own hardware, is adding support for Nvidia RTX PRO 6000 Blackwell GPUs and Jetson Thor and Jetson Orin edge platforms. Microsoft also said it uses Nvidia Dynamo software for post-training and RTX PRO 6000 Blackwell GPUs for Bing search models.

What it means

Companies can use Nvidia-based AI hardware and models through Microsoft's cloud, on their own premises, and on edge devices such as robots and factory machines. Confidential computing, which keeps data and models protected while they are being processed, is also being extended to large AI server racks.

THE FINE PRINT
  • The Vera Rubin NVL72 systems are described as a preview; no pricing, general availability date, or capacity figures were given.
  • The EY result (55% higher accuracy and 15x lower cost than proprietary frontier models) applies to one task, classifying purchasing records, and is measured on EY's own evaluation dataset.
  • Nemotron 3.5 Bolt and Transcribe are labeled "coming soon," the Windows 365 Direct Stream feature is in limited preview with public preview planned for early 2027, and confidential computing for rack-scale systems is described as ongoing collaboration.

More news

Nov 19 · AMDAMD and Unsloth enable NPU inference for 100+ agentic AI apps on Ryzen AI PCsNov 22 · SamsungSamsung promotes three chip veterans to lead memory, System LSI, and chip technology

← All news