TLDR
- Microsoft has entered a strategic partnership with Anthropic to integrate Claude AI models into its Foundry platform.
- Anthropic will purchase $30 billion worth of Azure compute capacity from Microsoft for AI model operations.
- Microsoft Foundry customers will gain access to Claude Sonnet 4.5, Claude Opus 4.1, and Claude Haiku 4.5 models.
- Despite the Microsoft deal, Amazon will remain Anthropic’s main cloud and training partner for model development.
- Nvidia is collaborating with Anthropic to optimize Claude models for next-generation GPU architectures.
Microsoft has announced a new strategic partnership with Anthropic, expanding its AI model portfolio across Azure and Foundry. The deal allows Microsoft Foundry customers to use Anthropic’s Claude models for the first time. Anthropic has also agreed to purchase $30 billion of Azure compute capacity from Microsoft.
Anthropic Brings Claude Models to Microsoft Foundry
Microsoft Foundry will now support Claude Sonnet 4.5, Claude Opus 4.1, and Claude Haiku 4.5 models from Anthropic. These models will be hosted on Microsoft’s AI infrastructure and integrated into enterprise services. Microsoft stated this move helps broaden customer options within the Foundry ecosystem.
Anthropic will continue using Amazon as its primary cloud and model training provider despite the Microsoft integration. This separation allows Anthropic to operate across platforms while benefiting from Microsoft’s enterprise tools. The company is positioning itself for broad AI adoption across ecosystems.
Anthropic has agreed to contract up to one gigawatt of compute power using Microsoft’s Azure infrastructure. “Anthropic will access scalable compute from Microsoft as part of our long-term AI development goals,” a spokesperson confirmed. Microsoft aims to strengthen its Copilot platform with expanded support for this model.
Nvidia Joins to Optimize Models for New Hardware
Anthropic is also collaborating with Nvidia to optimize its Claude models for new hardware. The company plans to use Nvidia’s upcoming Blackwell and Vera Rubin systems. Nvidia’s AI chips will support up to one gigawatt of compute capacity for Anthropic workloads.
Nvidia is investing up to $10 billion in Anthropic through this partnership. The investment will be phased as Anthropic ramps its hardware usage. Microsoft is also contributing $5 billion to support this joint venture.
Both companies will benefit from aligning model training and performance optimization on next-gen hardware. This tri-party cooperation aims to accelerate AI deployment at massive scale. Anthropic’s models are being tuned to deliver high performance on Nvidia’s new architecture.
Microsoft Adjusts OpenAI Terms to Enable New Deals
Microsoft revised its existing OpenAI partnership before moving ahead with the Anthropic deal. The updated agreement now includes extended IP rights through 2032. Microsoft also secured rights to post-AGI models under the new terms.
The companies agreed to remove exclusivity restrictions in the revised partnership. OpenAI can now work with other partners and release open-weight models. Microsoft no longer has first-refusal rights to OpenAI compute services.
These changes cleared the way for Microsoft to pursue other AI deals. Microsoft has now started integrating Anthropic models in several products. These models are prioritized in tools such as Visual Studio Code and Microsoft 365 Copilot.
Microsoft has increasingly used Claude 4 models in its auto AI selector. The company has also onboarded Claude Sonnet 4 and Opus 4.1 into enterprise tools. This partnership further expands Microsoft’s platform diversity.
