Microsoft is in early talks to supply Anthropic with artificial intelligence chips, a potential deal that would give the Claude maker more computing power while offering Microsoft a major test case for its own custom AI hardware.
The discussions center on Anthropic renting servers powered by Microsoft-designed AI chips, according to The Information. The talks are still in preliminary stages and may not result in a final agreement, but the negotiations show how aggressively major AI companies are searching for alternatives in a market still dominated by Nvidia.
The possible chip arrangement would follow a major November partnership among Microsoft, Nvidia, and Anthropic. Under that deal, Microsoft committed to invest up to $5 billion in Anthropic, while Nvidia committed up to $10 billion. Anthropic also agreed to buy $30 billion in Azure compute capacity from Microsoft.
For Microsoft, supplying Anthropic with AI chips would be more than a cloud contract. It would be a chance to prove that its custom silicon can support one of the most important AI labs in the world. Microsoft introduced its Maia 200 chip in January, calling it an inference accelerator built to improve the economics of AI token generation.
The company said Maia 200 is built on TSMC's 3-nanometer process and includes 216GB of HBM3e memory. Inference has become one of the highest costs in artificial intelligence. Training large models requires enormous computing power, but running those models for millions of users can also be brutally expensive.
Every Claude response requires chips, servers, electricity, and cloud capacity. If Microsoft can offer Anthropic efficient Maia-powered servers, it could lower costs for Anthropic and give Microsoft a stronger position in the AI infrastructure race. The talks also show Anthropic's growing leverage.
The company, founded by former OpenAI employees, has become one of the leading AI model developers through its Claude family of products. Its cloud partnerships already include Amazon and Google, and the Microsoft agreement adds another layer to its compute strategy.
Reuters reported that Anthropic has positioned Claude as available across major cloud platforms, making it a valuable customer for companies trying to sell AI infrastructure. Microsoft has long been closely tied to OpenAI, but its Anthropic partnership signals a broader strategy.
The potential deal comes as cloud giants race to reduce their dependence on Nvidia. Google has built its own Tensor Processing Units, while Amazon has developed Trainium and Inferentia chips. Neither Microsoft nor Anthropic has confirmed a deal. Reuters said both companies declined or did not immediately comment on the report.
