Microsoft Foundry Integrates Anthropic Claude Models with NVIDIA Blackwell GPUs to Power Secure Autonomous AI Agents Across Enterprise Networks
Azure native enterprises are now able to access Anthropic Claude models directly from Microsoft Foundry. This implementation runs on Microsoft Azure hardware leveraging the NVIDIA GB300 Blackwell Ultra Graphics Processing Units. The roll out allows organizations to build and scale domain focused autonomous artificial intelligence agents at a high level of computation performance. From the Microsoft Foundry documentation this rollout offers companies a direct way to launch autonomous software systems across specialized administrative tasks.
The computational infrastructure supporting this integration is built upon NVIDIA GB300 NVL72 computing systems. These are interconnected via NVIDIA Quantum X800 InfiniBand infrastructure to increase throughput on the data while the system runs large scale calculations. In turn, running Claude on the NVIDIA GB300 NVL72 should be able to optimize the execution of each inference for a faster and more cost effective performance within a given production. This reduces the total cost of ownership while increasing the automation of your domain.
The partnership between Microsoft, NVIDIA, and Anthropic is an extension of the three companies initiative to expand enterprise access to state of the art cognitive models. This synergy allows developers to plug predefined skills directly into the Anthropic stack utilizing approved NVIDIA components. As a result, enterprises will be able to run automated sub agents that can autonomously function within different parts of the organization to perform intricate administration functions.
For security organizations deploying autonomous systems in the enterprise, the NVIDIA Secure Agent Workspace Reference Design offers an architecture to adapt to tightly regulated corporate infrastructure. Security controls of user identity, network access, and runtime policies are implemented within the infrastructure to prevent exposing corporate data to the outside world. This ensures that the corporate data remains secure even as the autonomous agent performs high level workflows across the enterprise network.
