Story
Oct 20, 2026

Washington’s DeepSeek Ban Turns the AI Race Into a Security Standoff

The U.S. has ordered civilian agencies and contractors to remove DeepSeek technology over data-security fears, extending the ban to third-party tools. China’s defense of the startup underscores how quickly the AI contest is becoming a geopolitical clash.

DeepSeek’s rise was swift. The Hangzhou-based startup drew global attention in early 2025 after releasing models seen as competitive with leading U.S. systems — and immediately brought a Chinese AI challenger into Washington’s security sights.

By February, U.S. cybersecurity firm NowSecure had identified flaws in DeepSeek’s iPhone app, including unencrypted transmission of some information and insecure storage of sensitive data, according to the later Homeland Security account. A House Select Committee investigation in April separately accused the company of sending U.S. user data to China and tailoring some responses to Beijing’s censorship rules.

The administration’s formal crackdown followed in July. A Homeland Security order, signed July 24 and disclosed Oct. 19, warned that DeepSeek servers are located in China, where authorities can demand company information. The department said material submitted by U.S. employees or contractors could be transferred to China and used to collect intelligence on personnel and federal operations.

On Oct. 16, the secretaries of homeland security and war, alongside the director of national intelligence, issued orders requiring executive agencies to remove the software. The prohibition reaches further than DeepSeek’s own chatbot: civilian agencies cannot use outside services that integrate its models, either. CISA said the orders “prohibit the installation, accessing, or running of a Deepseek application” and require agencies to find and remove existing installations.

Washington’s case is thus not merely about one app, but about the less visible spread of its underlying models through government suppliers and contracts. SAM.gov lists DeepSeek under an active exclusion with no scheduled end date.

China pushed back Tuesday, defending DeepSeek after the federal civilian-agency ban and casting the action as part of the broader friction in the two countries’ heated technology race. The divide is now plain: U.S. officials see a data-security exposure embedded in an AI supply chain; Beijing sees another front in a strategic contest for technological power.