WASHINGTON: The United States has ordered its federal civilian agencies to remove products developed by Chinese artificial intelligence company DeepSeek from their information technology systems, escalating Washington’s technology restrictions on Chinese firms over national security concerns.
A notice from the Department of Homeland Security (DHS), published in the Federal Register, bars civilian agencies from acquiring DeepSeek products and third-party products and services built on DeepSeek implementations. Agencies must also remove existing deployments from their systems. The US DeepSeek ban applies to federal civilian agency IT systems and procurements.
US DeepSeek ban: what the order says
DHS said DeepSeek’s models present “multiple threat vectors”. It warned that use of the products by government employees or contractors “may result in the transfer of sensitive data to the PRC”, and that the data could produce “actionable intelligence” on US persons and government programmes. The department also said the models themselves “may contain malware or other security vulnerabilities”.
The order covers Hangzhou DeepSeek Artificial Intelligence Co and its affiliates and was issued under the Federal Acquisition Supply Chain Security Act of 2018. It applies to federal civilian agency IT systems and procurements. It does not extend to private businesses or individual users, and it does not remove DeepSeek from app stores.
The restrictions reach beyond the company’s own products to third-party tools developed using DeepSeek’s technology, a scope that is expected to complicate compliance, since agencies do not always know which tools were built on which models.
In a July memo made public with the notice, DHS also told federal agencies to monitor traffic flows in and out of government websites to address potential risks from DeepSeek.
A new step in a longer campaign
The move extends restrictions that were already in place. The US Navy and NASA had blocked the chatbot at their own discretion, the Commerce Department had restricted it, and states including New York, Tennessee and Virginia had barred it from government devices. Australia banned DeepSeek from government systems in February 2025, and South Korea’s foreign, trade and defence ministries blocked it on government computers the same month.
China responded on Tuesday. Its commerce ministry expressed “strong dissatisfaction and firm opposition”, saying the United States was using national security as a pretext to suppress Chinese firms, and pledged to take “all necessary measures” to defend its interests. The topic is expected to surface again when the two countries’ leaders meet next month at the Asia-Pacific Economic Cooperation summit.
One detail matters for readers: the order itself was signed on July 24, 2026. The October 20 Federal Register publication made it public rather than starting it.
The episode underlines how quickly AI governance is becoming a front line in the technology contest between Washington and Beijing, a contest with implications well beyond American agencies. Pakistan’s own institutions are waking up to the same questions: universities are building cybersecurity capacity with the HEC, the FBR is using AI to monitor tax returns, and the state is pushing a digital-nation agenda that will have to answer which models get to touch citizens’ data.
Frequently Asked Questions
Is DeepSeek banned in the United States?
No. The order applies only to federal civilian agency systems and purchases. The public, private companies and state governments outside their own rules can still use DeepSeek’s products.
Why did the US act?
DHS cited the risk of sensitive data reaching China, potential malware or vulnerabilities in the models, and the concentration of DeepSeek’s servers in China.
What has China said?
China’s commerce ministry expressed “strong dissatisfaction and firm opposition” and said it would take all necessary measures to defend its legitimate interests.
