Anthropic, Microsoft in talks for AI chip deal after $5 billion investment

Anthropic is currently in discussions with Microsoft to adopt the tech giant’s Maia artificial intelligence chip, although no formal agreement has yet been reached. Microsoft introduced the Maia 200 AI processor earlier this year but has not made it available to Azure cloud customers. The talks come as Anthropic faces growing computational demands amid the rising popularity of its AI products, including the Claude assistant and Claude Code programming tool. Microsoft’s Maia chip, which CEO Satya Nadella claims offers over 30% improved efficiency compared to existing silicon, is already operational in some Microsoft data centers. The potential deal would mark a significant win for Microsoft, which trails Amazon and Google in supplying specialized AI hardware to cloud clients. Anthropic has historically relied heavily on Nvidia GPUs for training and running its generative AI models but has recently diversified its chip suppliers. In April, Anthropic announced a 10-year, $100 billion agreement to use Amazon Web Services’ custom Trainium chips, and in October, it revealed plans to incorporate Google’s tensor processing units. Additionally, SpaceX disclosed that Anthropic will pay $1.25 billion monthly through May 2029 for computing power, underscoring the company’s urgent need for scalable infrastructure. Microsoft’s $5 billion investment in Anthropic last November and Anthropic’s commitment to spend $30 billion on Azure services highlight the deepening partnership between the two companies. However, Anthropic continues to leverage multiple cloud providers to meet its compute requirements. The adoption of Microsoft’s Maia chip could enhance Anthropic’s operational efficiency and reduce costs, potentially strengthening Microsoft’s position in the competitive AI cloud market. Both companies have declined to comment publicly on the ongoing negotiations.