Anthropic is reportedly in talks to use custom AI chips from Microsoft, a move that could reshape the fast-growing battle over AI infrastructure and reduce dependence on NVIDIA hardware.
According to a report by The Information (via Reuters), Anthropic is discussing renting servers powered by Microsoft’s in-house Maia AI chips to support growing demand for Claude and related AI services. The talks are still in the early stages, and there is no guarantee a deal will happen.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
What makes this interesting is the timing.
Anthropic has rapidly become one of the largest consumers of AI computing power. Its Claude chatbot, Claude Code tools, and enterprise AI products have seen strong adoption this year, especially among developers and businesses looking for alternatives to OpenAI models. That growth is pushing Anthropic to secure more compute capacity from multiple providers instead of relying heavily on Nvidia GPUs alone.
Microsoft’s Maia 200 chip enters the picture here.
Microsoft Wants This Deal
Microsoft has spent years building custom AI silicon to compete with Amazon and Google, both of which already rent their in-house chips to outside AI companies. Amazon offers Trainium chips through AWS, while Google rents its TPUs to partners, including Anthropic itself.
If Anthropic starts using Maia chips, it would become one of the biggest external validations of Microsoft’s AI hardware strategy so far.
Microsoft introduced the second-generation Maia 200 chip earlier this year. The chip is built using 3-nanometer technology from TSMC and is designed for inference workloads, which means running AI models efficiently at scale rather than training them from scratch.
That matters because inference has become one of the biggest costs in AI.
Every time users ask Claude a question, generate code, or process documents, servers must respond instantly. As AI usage grows, inference infrastructure becomes expensive very quickly. Microsoft reportedly optimized Maia 200 with a large SRAM memory capacity to improve chatbot responsiveness under heavy demand.
Anthropic Is Expanding Beyond One AI Supplier
Anthropic has already signed major infrastructure agreements with both Amazon and Google. Reports also suggest the company recently committed massive spending toward cloud and AI compute partnerships.
Adding Microsoft chips into the mix would show how major AI companies are now diversifying hardware providers instead of depending on one ecosystem.
That shift could have a bigger long-term impact on the AI industry than the deal itself.
For years, Nvidia has dominated AI computing because its GPUs became the standard for training and running large models. But rising costs, supply shortages, and exploding demand are forcing cloud providers to build alternatives faster than before.
Microsoft landing Anthropic as a Maia customer would signal that the AI chip market is finally becoming more competitive.
Discussion