TLDRocket
Sign in

US orders federal agencies to remove China’s DeepSeek AI from their systems

TechCrunch Kate Park ● Covered by 5 sources

The US told federal civilian agencies to stop using DeepSeek and strip it from their systems. Washington says China could get at government data through the AI startup’s tools.

Based on reporting by TechCrunch, Kate Park — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

The U.S. has ordered federal civilian agencies to stop buying DeepSeek products and services and to remove the Chinese AI startup from government IT systems. The Department of Homeland Security says the move is driven by national security concerns linked to DeepSeek’s operations in China.

This is broader than a simple product ban. It also covers third-party tools and services built on DeepSeek technology, though agencies can ask for exceptions.

The DHS warned that user data handled by DeepSeek could end up accessible to Chinese authorities under Chinese law. That’s the core fear here: if government users feed sensitive material into the system, the data may not stay where Washington wants it.

The order builds on earlier limits. A law passed in December 2025 already targeted DeepSeek’s use in intelligence national security systems, and this latest step extends the restrictions to civilian agencies, including both new purchases and existing IT systems.

China pushed back quickly. Foreign Ministry spokesperson Mao Ning said Beijing opposes “overstretching the concept of national security and politicizing technology and trade issues,” and said China would “resolutely safeguard the legitimate rights and interests of Chinese enterprises.” The clash sits inside the larger U.S.-China AI fight, where DeepSeek has been pressuring American rivals with low-cost open-weight models.

My take — AI-written commentary, not fact-checked reporting

This is what AI policy looks like once governments stop pretending data security is an abstract seminar topic. If a model vendor sits inside a geopolitical dispute, agencies are going to treat it like a risk surface, not a neat little productivity tool. The surprise is how long it takes for that lesson to arrive, every single time.

Read more about this at: TechCrunch

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.