Microsoft, Nvidia back Anthropic, bring Claude to Azure

Both companies intend to participate in Anthropic’s next funding round – with Nvidia committing up to $10 billion and Microsoft up to $5 billion – according to Tuesday’s announcement and a person familiar with the matter.

“We’re increasingly going to be customers of each other. We will use Anthropic models, they will use our infrastructure and we’ll go to market together,” Microsoft CEO Satya Nadella said in a video, adding that OpenAI remains a critical partner.

Microsoft will make the latest Claude models available to customers through Azure AI Foundry, positioning Claude as the only frontier model available across all three major clouds. Amazon will remain Anthropic’s primary cloud provider and training partner as the startup expands work on Azure.

Anthropic will also work with Nvidia on chips and model performance and commit up to 1 gigawatt of compute using Grace Blackwell and Vera Rubin hardware. Industry executives estimate that 1 GW of AI computing can cost $20–$25 billion, underscoring the scale of resources involved.

Analysts framed the pact as spreading dependencies across the AI stack and reducing reliance on OpenAI. “Microsoft has decided not to rely on one frontier model company. Nvidia was also somewhat dependent on OpenAI’s success and is now helping generate broader demand,” said Gil Luria of D.A. Davidson. Jacob Bourne at eMarketer called the investments further consolidation around a few key players.

Founded in 2021 by former OpenAI staff, Anthropic has become a major rival to the ChatGPT maker, driven by enterprise adoption. Reuters has reported the company was recently valued at $183 billion, projected an annualized revenue run rate of around $26 billion next year, and counts 300,000+ business and enterprise customers. The announcement also comes weeks after OpenAI restructured away from its non‑profit roots and announced a $38 billion cloud deal with Amazon.com as it reduces reliance on Microsoft; CEO Sam Altman has said OpenAI aims to spend $1.4 trillion to develop 30 GW of computing capacity.

As we covered previously, Microsoft signed a multi-year deal with Nebius for dedicated AI computing capacity, initially valued at $17.4 billion and potentially reaching about $19.4 billion over five years. Deliveries are slated from a New Jersey data center beginning later this year, with 300 megawatts planned and growth to 400 megawatts. Microsoft also procures AI compute from CoreWeave. Nebius finances equipment using contract cash flow and loans backed by Microsoft’s credit.