The AI circular economy just got an upgrade: Inside the Microsoft, Nvidia, Anthropic deal

This kind of tightly connected deal is becoming common in the industry, where cloud platforms, chipmakers and AI labs rely on one another to keep pushing the technology forward.

Even as US markets wobbled on AI bubble fears and the Dow logged its worst three-day slide since April on November 18, a new three-way partnership between Microsoft, Anthropic and Nvidia shows that the world's most valuable tech firms are doubling down on AI.

The deal expands the AI industry’s increasingly intertwined “circular economy” at a time when investors are questioning whether the sector’s infrastructure binge is running far ahead of monetisation.

What the deal actually means?

Anthropic, the company behind the Claude AI models, will run a significantly larger share of its workloads on Microsoft’s Azure cloud, which in turn is powered by Nvidia’s high-end AI systems. The company has also committed to purchase $30 billion worth of Azure compute capacity and contract access to additional capacity that could reach one gigawatt.

For the first time, Anthropic and Nvidia will also collaborate on chip and model design. Future Nvidia architectures such as Grace Blackwell and Vera Rubin will be optimised for Anthropic workloads, while Claude will be tuned to run more efficiently on Nvidia hardware.

Claude will also be offered more widely across Microsoft’s enterprise cloud tools and Copilot products. Customers of Microsoft Foundry will be able to access Anthropic’s frontier Claude models including Claude Sonnet 4.5, Claude Opus 4.1, and Claude Haiku 4.5. This partnership will make Claude the only frontier model available on all three of the world’s most prominent cloud services.

Both Microsoft and Nvidia plan to invest heavily in Anthropic as part of the arrangement, with commitments that could reach $5 billion and $10 billion, respectively, further tightening the financial and technical links among the three companies.

Also read: The trillion-dollar merry-go-round: Inside AI's circular economy

Why the Microsoft connection is significant?

The partnership also marks a shift in Microsoft’s AI strategy. For Anthropic, it provides deeper access to a cloud giant that invested more than $13 billion in OpenAI and helped turn ChatGPT into a global phenomenon. But in recent months, Microsoft and OpenAI have moved into more direct competition, especially in selling AI services to large companies.

Microsoft CEO Satya Nadella said on November 18 that the company and Anthropic are now becoming “customers of each other,” stating, “We will use Anthropic models, they will use our infrastructure, and we will go to market together.”

Nadella stressed that OpenAI “remains a critical partner,” but the new deal indicates Microsoft is reducing its dependence on the ChatGPT maker, despite holding a 27 percent stake in OpenAI’s for-profit business, currently valued at about $135 billion.

Why this matters

The agreement ties three important AI players together:

• Microsoft gets another major AI model running on Azure, strengthening its cloud advantage.
• Nvidia secures a long pipeline of demand for its GPUs, protecting its lead in AI chips.
• Anthropic gets the computing scale it needs to compete with OpenAI and others building frontier AI.

Also read: AI hits a speed bump: Why the next big thing isn't coming so fast

Why are these deals raising eyebrows?

The deal comes just as Wall Street is questioning whether the AI boom may be running ahead of real revenue. Nvidia’s upcoming earnings are seen as a major test of whether AI demand is durable or fuelled by hype.

Some worry the industry is spending first and figuring out business models later. Others argue AI is still in its earliest stage and that today’s infrastructure builders could end up dominating the next era of computing.

Bottom line

For now, the money is still moving, the chips are still selling and cloud capacity continues to expand. If demand for AI keeps rising, this deal could look like a smart move that shaped the future of computing. If growth slows, it may expose an industry that spent heavily on a future that did not arrive fast enough.