Anthropic announced that its Claude family of AI models is now generally available in Microsoft Foundry on Azure, running on Nvidia’s GB300 Blackwell Ultra GPU systems. This marks Anthropic’s first deployment on Nvidia hardware.
The deployment provides Azure customers access to Claude models for building autonomous and domain-specific AI agents as enterprises increase investments in agentic AI applications. The models utilize Nvidia GB300 NVL72 systems with Quantum-X800 InfiniBand networking, enabling the creation of advanced AI agents and specialized sub-agents that can function across various business operations.
Anthropic stated that the deployment aims to enhance inference performance and efficiency while lowering the total cost of ownership for enterprise AI workloads. This announcement follows an indication from Nvidia and Microsoft in June that Claude models would “run natively on NVIDIA GB300 Blackwell Ultra systems on Azure, with customer availability in the weeks ahead.”
The launch builds on a strategic partnership announced in November 2025, where Anthropic committed to purchasing $30 billion of Azure compute capacity. Nvidia and Microsoft agreed to invest up to $10 billion and $5 billion, respectively, in Anthropic.
At the time of the partnership, Anthropic’s compute commitment was set to reach initially up to one gigawatt of capacity, utilizing Nvidia Grace Blackwell and Vera Rubin systems. The partnership also made Claude available across Microsoft’s Copilot family, which includes GitHub Copilot and Microsoft 365 Copilot.
Microsoft initially deployed a large-scale Azure cluster with over 4,600 Nvidia GB300 Blackwell Ultra GPUs in October 2025, designed primarily for OpenAI’s inference workloads. The cluster, connected through Nvidia’s next-gen InfiniBand network, was developed to reduce AI model training times from months to weeks. Claude now becomes the second major AI model family to operate on this Azure infrastructure.
