The US Department of Homeland Security (DHS) issued an order on Monday, 19 October 2026, according to The Register, directing federal civilian agencies not to use technology products from the Chinese AI lab DeepSeek and to remove any existing implementations. The Register links the order as a PDF; the wording quoted here comes from that document.
What is banned
The order prohibits "all DeepSeek and DeepSeek-affiliate products, and third-party products and services developed using DeepSeek implementations". According to The Register, the second part is vague: it is not clear which third-party products were developed using DeepSeek, nor whether "developed using" means ongoing API access to DeepSeek models or any app coded with DeepSeek's help.
Open questions around AI gateways
The source points to a further problem: services such as OpenRouter and Vercel AI Gateway make DeepSeek models available via APIs. OpenRouter, by its own description, "routes requests to the best available providers for your model", and Vercel AI Gateway does the same by default. Agencies could therefore use DeepSeek models unintentionally if they have not constrained their AI providers. A DHS spokesperson could not immediately be reached for comment, according to The Register.
Background
The Register reports that US officials have considered a ban since DeepSeek-R1 was released in early 2025. In February 2025, Representatives Darin LaHood (R-IL) and Josh Gottheimer (D-NJ) introduced the "No DeepSeek on Government Devices Act"; according to The Register it never got out of committee. Other bodies acted earlier: the Defense Information Systems Agency (DISA) began blocking DeepSeek on Pentagon IT networks in January 2025, and the US Navy, NASA, the US House of Representatives and various US states took similar steps.
Security concerns
After concerns about DeepSeek's iOS app surfaced shortly after R-IL's launch, a December 2025 report on its Android and iOS apps cited, among other things, plaintext information leaks and links to servers run by Chinese tech firms. The source notes these are mainly common mobile app problems. Katie Moussouris, founder and CEO of Luta Security, told The Register that banning DeepSeek's own app and API is workable, but that the wording on third-party products goes too far because "developed using" is undefined and cannot be practically verified.
Assessment
According to The Register, the timing of the order could benefit US AI companies that hope to go public soon, since open-weight models from China have proven popular on capability and cost. For organisations in Germany the ban is not a direct intervention, but it could become relevant if their services or suppliers serve US agencies or source models through gateways. Whether and how this affects procurement rules cannot be derived from the available sources.
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
