Anthropic made its Claude models available through Microsoft Foundry on Azure, expanding enterprise access to its AI systems for building autonomous agents.

The integration marks the first time Anthropic has deployed its models on Nvidia hardware, running Claude on the chip maker's GB300 Blackwell Ultra systems designed for advanced AI workloads.

Azure-native businesses can now access Claude for building domain-specific AI agents through Microsoft's cloud platform. The move addresses growing enterprise demand for more autonomous, intelligent technologies beyond traditional chatbots.

Enterprise deployment features

Customers can deploy Claude agents using Nvidia's Secure Agent Workspace Reference Design, which provides blueprints for governed AI environments. The framework includes controls for identity management, network access, credentials, and runtime policies.

Anthropic currently lists 11 Claude models in Microsoft Foundry, including Opus 4.8 and Sonnet 4.6 variants.

The launch builds on a strategic partnership formed by Microsoft, Nvidia, and Anthropic in 2025 to expand enterprise Claude access and accelerate AI workloads on Nvidia infrastructure.

Under that agreement, Anthropic committed to purchasing approximately $30 billion in Azure compute capacity, with options to contract additional capacity up to 1 gigawatt. Microsoft invested up to $5 billion in Anthropic, while Nvidia committed up to $10 billion.

The partnership represents a significant shift toward enterprise-focused AI deployment, as companies increasingly seek autonomous agents capable of handling complex, domain-specific tasks rather than general-purpose conversational AI.

Nvidia said the project demonstrates how Claude can scale for enterprise and agentic AI workloads, positioning the collaboration as a blueprint for future enterprise AI deployments across cloud platforms.