US orders federal civilian agencies to remove DeepSeek products and third-party derivatives

The US Department of Homeland Security issued a directive requiring federal civilian agencies to stop purchasing products and services developed by Chinese AI company DeepSeek and to remove them from government IT systems. The ban covers not only DeepSeek's own products but also third-party products and services built with DeepSeek technology; agencies may apply for exceptions. The directive cites concerns over DeepSeek's handling of user data, arguing that sensitive information of US government users could be obtained by Chinese authorities under Chinese law. A law passed in December 2025 had already restricted its use in intelligence-community national security systems; this move extends the restrictions to federal civilian agencies and covers both new procurement and existing systems. For system integrators and cloud providers taking on US federal projects, this means checking before delivery whether DeepSeek models or APIs are embedded in their solutions, and replacing and migrating existing systems as well. For teams operating in the US and offering AI capabilities to government and enterprise clients, factoring model provenance into compliance review will become a mandatory step when selecting models. China has objected, with Foreign Ministry spokesperson Mao Ning saying it opposes overstretching the concept of national security and politicizing science, technology and trade issues. The original text also notes that DeepSeek has faced regulatory scrutiny in multiple countries since releasing its R1 reasoning model in January 2025, and that the US has accused six Chinese AI companies, including DeepSeek, of extracting information from US AI systems through large-scale model distillation.