Microsoft and Nvidia Bet $15B on Anthropic to Expand AI Power
The partnership boosts Anthropic’s valuation to $350 billion and secures billions in cloud and compute commitments from Microsoft and Nvidia.
For the last few years, Microsoft’s AI strategy has revolved around OpenAI. The partnership was huge, public, and central to everything from Copilot to Azure’s marketing pitch.
But as OpenAI grew into a commercial giant, it began signing huge cloud deals elsewhere and signalling that its long-term ambitions didn’t revolve around Microsoft alone, like the $38 billion cloud deal with Amazon.
Louis Eriakha
Add in OpenAI’s own restructuring, which gave it more freedom to choose how and where it scaled, and you can see why Microsoft began looking for a second major model partner.
That brings the story to this new three-way deal, one that ropes in Nvidia, Anthropic, and billions of dollars in fresh commitments. Microsoft is putting up to $5 billion into Anthropic, Nvidia is adding up to $10 billion, and Anthropic’s valuation has jumped into the $350 billion range.
The benefit to Microsoft is simple: it now has a second engine powering its AI future. The investment secures long-term access to Claude models, strengthens Azure’s relevance, and guarantees billions in future compute demand through Anthropic’s $30 billion commitment to Azure. Instead of being the company waiting for OpenAI’s next move, Microsoft now anchors two of the world’s top frontier model developers. And by making Claude available through Azure AI Foundry, Microsoft widens what it can offer enterprise customers who want alternatives or safeguards.
For Anthropic, the upside is structural. The company gains access to Azure’s massive compute pipeline, deeper engineering collaboration with Nvidia, and enough capacity to scale its next generation of Claude models without hitting infrastructure limits. With more than 300,000 enterprise customers and rising demand for Claude, the timing gives Anthropic breathing room and expansion room simultaneously.
Then there’s Nvidia. This deal deepens its influence over the next wave of foundation models. Anthropic’s commitment to train on up to a gigawatt of Nvidia’s Grace Blackwell and Vera Rubin systems ensures Nvidia’s architectures stay at the centre of frontier-model development. And by investing directly, Nvidia isn’t just selling chips, it’s embedding itself into another major AI ecosystem.
Seen in context, this partnership is part of a broader trend: AI alliances are no longer research collaborations; they’ve become infrastructure pacts. Massive, long-term, capital-heavy agreements designed to secure compute, model access, and strategic leverage.
We’ve already seen similar moves. In September, OpenAI, SoftBank, and Oracle announced the Stargate initiative, a project aiming to build AI-focused data centres across the US with a potential $500 billion commitment and 10 gigawatts of total capacity. That scale mirrors where the industry is heading: giant, multi-party alliances built around compute, not just models.
Ogbonda Chivumnovu
Against that backdrop, Microsoft’s decision becomes clearer. OpenAI is still enormously important, Satya Nadella, CEO of Microsoft, has emphasised that repeatedly, but no global platform can afford a single point of dependency in a race this capital-intensive and volatile. Bringing in Anthropic and Nvidia gives Microsoft strategic balance, a broader model portfolio, and more control over its long-term AI trajectory.
Put together, the deal resets the AI power map. Microsoft gains stability and diversification, Anthropic gains the infrastructure it needs to scale safely, and Nvidia gains deeper influence over the hardware backbone of future AI systems.
