US Orders Federal Agencies to Remove DeepSeek From Their Systems, Including Tools Built on It

TechCrunch adds that the stated concern is that data from US government users could be reachable by Chinese authorities under Chinese law. A July memo, released alongside the notice, had already told agencies to watch data moving in and out of government websites for DeepSeek-related risk.

How did we get here?

The order is the latest step in a longer run of restrictions. The Commerce Department and the Navy had already banned DeepSeek on staff devices, and New York, Virginia and Tennessee restricted it on state government hardware, according to OfficeChai. A law passed in December 2025 limited its use in intelligence national security systems, TechCrunch says.

In September, TechCrunch notes, a CISA advisory accused DeepSeek and five other Chinese AI firms, including Alibaba and Moonshot AI, of large-scale distillation, meaning the extraction of capabilities from US models. OpenAI and Anthropic have both said they saw suspected attempts by DeepSeek to collect their models' outputs.

How has China responded?

Foreign Ministry spokesperson Mao Ning said on Tuesday that Beijing opposes stretching the idea of national security and politicising technology and trade, and that China would defend its companies' legitimate rights, TechCrunch reports. Bloomberg reported on 16 October that Beijing is weighing export controls on 23 AI-related technologies, including chip designs and some advanced models.

The practical test now falls on federal IT and procurement teams. They must find every product with DeepSeek inside it, including software that quietly calls its models, and replace it, while the exceptions process shows how strict the order will be in practice.