Microsoft has intervened in a legal dispute between artificial intelligence start-up Anthropic and the US Department of Defense, supporting the company’s challenge against a government decision to classify it as a supply chain risk.
The software group filed an amicus brief urging courts to issue a temporary restraining order that would prevent enforcement of the Pentagon’s designation while the case proceeds. Microsoft argued that the measure, described as “drastic” and “unprecedented”, could have wider consequences for the American technology sector and for the development of artificial intelligence used in national security contexts.
The dispute began after negotiations between Anthropic and the Pentagon collapsed over the potential military deployment of the company’s AI models. Anthropic’s chief executive Dario Amodei insisted on restrictions that would prohibit the use of its systems for lethal autonomous weapons and for domestic mass surveillance. Following the breakdown in talks, US defense secretary Pete Hegseth moved to remove Anthropic from the Pentagon’s supply chain, a step typically reserved for companies considered security risks.
The administration has also instructed federal agencies to stop using Anthropic’s Claude chatbot, escalating tensions between the government and parts of the technology sector. Microsoft’s intervention marks the first time a major technology company has publicly supported Anthropic’s legal challenge, highlighting divisions within Silicon Valley over the future role of artificial intelligence in military and intelligence operations.
In its filing, Microsoft argued that suspending the designation would allow time for negotiations and clearer policy discussions regarding acceptable uses of AI in defence. The company also warned that an abrupt exclusion of Anthropic could disrupt military capabilities, noting that the Claude system is currently the only AI tool used in classified military environments, although OpenAI recently secured a separate agreement with the Pentagon.
Microsoft is not a defendant in the case but submitted its legal arguments to courts in California and Washington DC. The company has commercial ties with Anthropic, including a $30bn cloud computing partnership and recent plans to integrate its AI coding models into Microsoft software used across parts of the US government.
