Microsoft announced Thursday that it will continue using Anthropic’s artificial intelligence models in its products for clients, excluding the U.S. Department of War. The decision follows the Pentagon’s move earlier in the day to label Anthropic a supply-chain risk, a designation the startup plans to challenge in court. Microsoft confirmed that Anthropic’s Claude models will remain available through platforms such as Microsoft 365, GitHub, and AI Foundry for non-defense projects.

The announcement comes after U.S. President Donald Trump urged federal agencies to stop using Anthropic’s technology, and Secretary of War Pete Hegseth said the company’s Pentagon services would end within six months. Talks between Anthropic and the Department of War reportedly collapsed over issues related to mass domestic surveillance and autonomous weapons. Meanwhile, rival OpenAI reached an agreement for its models to be used in classified Pentagon workloads.

Microsoft’s stance makes it the first major company to maintain collaboration with Anthropic following the federal blacklist. The two firms have significant financial ties, including Anthropic’s commitment to spend $30 billion on Microsoft’s Azure cloud and Microsoft’s pledge to invest up to $5 billion in Anthropic.