The US Department of Homeland Security published through the Federal Register an order excluding and removing products and services of China's DeepSeek from federal civilian agencies' IT systems. The order also covers third-party products built on DeepSeek implementations and bans future purchases. It was signed on 24 July 2026 under the Federal Acquisition Supply Chain Security Act of 2018 (FASCSA) and was already in force before publication. The Department of Defense, the intelligence community and national security systems are excluded.

What it means: The order publicly formalises a ban that had already applied since the summer and shows how artificial intelligence is now treated as a supply chain security issue. According to the reported reasons, the concern covers storage of user data in China, transfer of prompts and responses to servers there, and vulnerabilities found in forensic analyses. The most important detail is that the ban is not limited to the DeepSeek app but also covers third-party products that use its models. That means software suppliers to the US government will need to know and disclose which models they use behind the scenes. Because many DeepSeek models are released with open weights and are easy to embed, checking is not simple. In Europe, some data protection authorities have already restricted the app, while the AI Act and the GDPR set transparency and data protection requirements. For Greek companies selling software to the public sector or to regulated clients, keeping a record of which AI models are used and where data is processed is becoming more and more necessary.

Read the original on Federal Register

Related service: myDATA integrations

Related stories