Anthropic has agreed to purchase $30bn in compute capacity from Microsoft Azure as part of newly announced strategic partnerships with Microsoft and Nvidia.

The arrangement allows Anthropic to scale its Claude AI model on the Azure platform, which is powered by Nvidia ‘s technology.

Under the agreement, Anthropic will contract additional Azure compute capacity of up to one gigawatt (1GW), thereby increasing access to the Claude model for enterprise customers on the Azure platform.

Nvidia and Anthropic have also launched a technology partnership focused on co-designing and engineering AI workloads.

This collaboration aims to enhance the performance, efficiency, and total cost of ownership for Anthropic models while also optimising future Nvidia architectures for the AI safety and research company’s requirements.

The initial commitment will see Anthropic use up to 1GW of compute capacity equipped with Nvidia Grace Blackwell and Vera Rubin systems.

In addition to these developments, Microsoft and Anthropic are expanding their existing relationship to offer wider access to Claude AI models for businesses.

Models including Claude Sonnet 4.5, Claude Opus 4.1, and Claude Haiku 4.5 will become available to customers through the Microsoft Foundry platform.

This move is said to position Claude as the only frontier model accessible across what are described as the world’s three most prominent cloud services.

Azure customers are expected to benefit from an increased selection of AI models as well as features specific to Claude.

Microsoft noted that it will maintain support for Claude within its Copilot range, including GitHub Copilot, Microsoft 365 Copilot, and Copilot Studio products.

As part of these partnerships, Nvidia has agreed to invest up to $10bn in Anthropic, while Microsoft’s commitment stands at up to $5bn in the AI startup.

Separately, during the Microsoft Ignite conference this week, Nvidia broadened its collaboration with Microsoft with several technical integrations.

These include adopting next-generation Spectrum-X Ethernet switches for the new Fairwater AI superfactory based on the Nvidia Blackwell platform, as well as integrating Nvidia Nemotron technology with Microsoft SQL Server 2025.

The partnership includes public previews of new Azure NC Series virtual machines using Nvidia RTX PRO 6000 Blackwell Server Edition GPUs.

It also introduces features for onboarding AI agents into Microsoft 365, along with improvements in inference performance, cybersecurity, and physical AI applications.

Microsoft’s AI Superfactory project connects its Fairwater data centre in Wisconsin with a new facility in Atlanta, Georgia.

This infrastructure is set to integrate hundreds of thousands of Nvida Blackwell graphics processing units (GPUs) for training operations.

Additionally, Microsoft is deploying over 100,000 Blackwell Ultra GPUs in GB300 NVL72 systems worldwide for inference purposes.

Microsoft product management corporate vice president Nidhi Chappell said: “Our collaboration with Nvidia is built on driving innovation across the entire system and full stack, from silicon to services.

“By coupling Microsoft Azure’s unmatched data centre scale with Nvidia’s accelerated computing, we are maximising AI data centre performance and efficiency, which is of paramount importance for our customers leading the new AI era.”

Recently, Anthropic committed to investing $50bn to expand its computing infrastructure in the US, with new data centres planned for Texas and New York.