The Daily Byte
Breaking
Welcome to The Daily Byte — real stories from across Pakistan.
Saturday, 21 November 2026 · Pakistan

The Daily Byte

Stories from Pakistan that deserve to be heard
Minor World

Nvidia Vera Rubin Systems Now Running on Azure

By The Daily Byte ·21 Nov 2026 ·4 min read
Server racks inside a Microsoft Azure data centre

Nvidia's Vera Rubin NVL72 systems are now running in Microsoft Azure data centres, the two companies announced on 19 November at Microsoft Ignite in San Francisco. Access to Azure virtual machines powered by the systems is expanding through a preview, as the companies deepen an AI partnership that stretches from cloud data centres to Windows PCs.

Vera Rubin is Nvidia's newest AI computing platform — the hardware built to train and run the largest AI models. The announcement does not make the systems generally available: it is a preview, and neither company has given a price, a virtual-machine series name, a list of regions, or a date for general availability. Customers can try the systems, but cannot yet plan a budget or a rollout around them.

First to validate, still in preview

Microsoft says it was the first cloud provider to bring up a Vera Rubin NVL72 system for validation, in March 2026, and the first to run production systems, in August. Azure's offering is nevertheless still a preview, and reporting on the announcement notes that Oracle reached customers first, with Vera Rubin generally available on its cloud since 27 October.

The systems are designed for demanding reasoning and agentic applications, as well as large-scale model training and inference. Nvidia points to recent benchmark results as evidence of the joint engineering: in MLPerf Training v6.1, Azure scaled to 10,240 Nvidia GPUs using GB200 NVL72 systems on the DeepSeek-V3 model, the largest scale in that round. In MLPerf Inference v6.1, Azure's GB300 NVL72 systems led cloud submissions on DeepSeek-R1 interactive throughput by more than 10%, according to Signal65.

The companies are also working to extend confidential computing to bare-metal and rack-scale AI infrastructure, so that sensitive models and data stay protected while they are being processed. That work is aimed at organisations handling proprietary or regulated information. They are collaborating on power efficiency too, using Nvidia MaxLPS to help data centres get more performance from the power available to them.

Open models on Microsoft Foundry

Beyond raw computing, the expanded partnership covers the open models and tools available through Microsoft Foundry. The expanded Nvidia Nemotron 3.5 lineup on Foundry Managed Compute includes Super, for long-running agents working across coding, research and business tasks, and Lightning, for applications that need fast responses and tool use. Embed and Rerank support information retrieval, Parse extracts structured information from documents, and Content Safety detects harmful content. Bolt, for high-throughput, low-latency reasoning, and Transcribe, for speech recognition, are coming soon.

Microsoft's serverless Programmed Training offering gives developers control over post-training workflows while Foundry manages the underlying infrastructure, and Foundry Managed Compute adds support for deploying custom model weights. Nvidia reported that EY used Programmed Training to adapt Nemotron open models for classifying corporate purchasing records, improving accuracy on its evaluation dataset by 55% and lowering cost by 15 times compared with proprietary frontier models.

Microsoft is also using Nvidia technology in its own operations. Microsoft AI uses Nvidia Dynamo inference software for reinforcement-learning rollout during post-training, and Bing search models run on RTX PRO 6000 Blackwell graphics processors.

Beyond the data centre

The collaboration extends to systems closer to where data is generated. Expanded Nemotron 3.5 support in Foundry Local brings reasoning, coding and tool-using agents to Azure Local environments, and Azure Local will support RTX PRO 6000 Blackwell GPUs from hardware partners. Azure Local is also expanding to Jetson Thor and Jetson Orin edge AI platforms for industrial uses such as machine vision, robotics and autonomous operations, with Nvidia's Cosmos physical AI models being brought to Foundry Local.

On Windows, the companies will demonstrate local large-language-model experiences on RTX Spark-powered PCs at Ignite. For organisations using Cloud PCs, Direct Stream capabilities are coming to GPU-accelerated Windows 365 Cloud PCs, supporting up to two 4K streams at 60 frames per second with low latency. A limited preview is open for sign-ups, with a public preview planned for early 2027.

Nvidia's BioNeMo Agent Toolkit and ALCHEMI Toolkit are also expanding into the Microsoft Discovery catalogue, adding models for life sciences, drug discovery, and materials and chemical science to Microsoft's agentic research platform. The picture is of a partnership that now covers both the supply of AI computing and its use — but for customers waiting on Vera Rubin in Azure, the practical point is that the preview has begun, and the full launch date is still to come.

Sources: NVIDIA Blog | WindowsForum