The commitments are part of ‘new strategic partnerships’ the three companies jointly announced on Tuesday, with Microsoft and Nvidia committing to invest up to $5 billion and up to $10 billion, respectively, in Anthropic, an OpenAI rival backed by Amazon.
OpenAI competitor Anthropic has committed to spend $30 billion on Microsoft Azure compute capacity and use up to 1 gigawatt of data center infrastructure based on Nvidia’s latest rack-scale AI platforms.
The commitments are part of “new strategic partnerships” the three companies jointly announced on Tuesday, with Microsoft and Nvidia also committing to invest up to $5 billion and up to $10 billion, respectively, in Anthropic, a well-funded AI startup founded by former OpenAI employees that is known for its family of Claude LLMs.
[Related: AMD Sees ‘Very Clear Path’ To Double-Digit Share In Nvidia-Dominated Data Center AI Market]
The announcement marks a shift of sorts for Anthropic, which received $8 billion in investments from Amazon across deals that culminated last year with the AI model provider naming Amazon Web Services as its primary AI training partner. Google has also invested $3 billion into Anthropic, which uses Google Cloud as another cloud provider.
It’s also the latest in so-called circular spending by large and influential AI companies, where funds often flow back and forth between two or more parties, with suppliers in some cases investing in customers in exchange for buying their products. Recent examples include deals struck by OpenAI with Nvidia, Oracle and CoreWeave.
In Anthropic’s new partnership with Microsoft, the two companies are expanding upon their existing relationship to make Anthropic’s leading-edge Claude models available to customers of the Azure AI Foundry. The list of applicable models includes Claude Sonnet 4.5, Claude Opus 4.1 and Claude Haiku 4.5.
The three companies noted that this move will make Claude the “only frontier LLM model available on all three of the world’s most prominent cloud services.”
In another part of the deal, Microsoft is “committed to continuing access for Claude across Microsoft’s Copilot family, including GitHub Copilot and Copilot Studio,” they said.
Anthropic is expected to “contract an additional compute capacity” of up to 1 gigawatt from Azure in addition to its initial $30 billion commitment with Microsoft.
As for Nvidia’s end of the agreement, the AI infrastructure giant is forming a “deep technology partnership” with Anthropic for the first time, the three companies said.
With Nvidia and Anthropic expected to collaborate on design and engineering, their aim is to optimize Anthropic’s models “for the best possible performance, efficiency” and total cost of ownership. They also plan to optimize “future Nvidia architectures for Anthropic workloads.”
The 1 gigawatt of compute capacity Anthropic plans to use from Nvidia will consist of the vendor’s Grace Blackwell systems that debuted last year and its next-generation Vera Rubin systems that are expected to launch in the second half of next year.
