The United States government has taken its most sweeping action yet against a Chinese artificial intelligence platform, DeepSeek, that has been under scrutiny for months.

A new federal order moves to strip the tool from every civilian agency in the country, citing concerns that go well beyond just data privacy.

Homeland Bans DeepSeek AI on All US Gov't Tech

The Department of Homeland Security published a formal notice in the Federal Register ordering the exclusion and removal of all DeepSeek products from Federal Civilian Executive Branch information technology systems and procurements. The order applies to the full breadth of civilian government agencies, from the Environmental Protection Agency to the Federal Communications Commission, and requires them to both stop using DeepSeek and actively remove any existing installations of its products.

The DHS order also prohibits DeepSeek from participating in any federal government contracts going forward, effectively cutting the company off from any role in federal procurement. The order was published earlier today and comes just days after DeepSeek released its latest AI model, V4.1-Flash-1017, which the company described as smarter, faster, and more efficient than its predecessors.

The DHS order is the most sweeping federal action taken against DeepSeek to date, moving the ban from a patchwork of individual agency restrictions into a unified government-wide mandate covering all federal civilian systems. This is not, however, the first time the US government has moved against the platform. Individual agencies, including the Pentagon, NASA, the Navy, and the Department of Commerce, had already restricted or banned DeepSeek on their own devices prior to this order.

Homeland Security: DeepSeek Brings Security Threats

DHS characterized the risks posed by DeepSeek in stark terms in its Federal Register notice. The agency stated that DeepSeek's models could contain malware or other security vulnerabilities, making them a direct risk to government systems and the personnel who use them, according to Reuters.

TechTimes

The department also warned that using DeepSeek on government systems could result in the transfer of sensitive data belonging to US government employees and vendors to the Chinese government. That concern is rooted in a broader legal reality: Chinese law grants the Chinese government broad authority to compel domestic companies to cooperate with intelligence agencies and hand over data upon request.

DeepSeek's own privacy policy acknowledges that personal data it collects is stored on servers located in China. That data includes not only text and uploaded files but also more granular information such as keystroke patterns, IP addresses, and device identifiers.

Security researchers also found hidden code in DeepSeek's platform that appeared to link user data to CMPassport.com, a registry operated by China Mobile, a company that has itself been banned by the Federal Communications Commission from operating in the United States.

DeepSeek AI Collects Sensitive US Data

The range of data that DeepSeek's platform is capable of collecting has been a central concern for security experts and lawmakers since early 2025. The AI tool first broke into global consciousness by releasing a capable model at a dramatically lower cost than its American competitors.

Within weeks of that release, Texas became the first state to ban DeepSeek on government devices in January 2025, followed by New York, Virginia, and Tennessee in February 2025. At the federal level, lawmakers moved quickly as well. A bipartisan group of senators introduced the No DeepSeek on Government Devices Act, identified as S.765, which sought to prohibit DeepSeek's use across all federal government devices and networks.

Senators Bill Cassidy and Jacky Rosen separately introduced the Protection Against Foreign Adversarial Artificial Intelligence Act of 2025, which aimed to prohibit federal contractors from using DeepSeek to fulfill government contracts.

Those legislative efforts now appear to have informed the DHS order, which enacts a version of the same restrictions through regulatory rather than legislative action. The timing also places the order alongside an existing federal approach to Chinese technology, following earlier restrictions placed on companies like Huawei and ZTE, and mirroring the kinds of bans previously applied to TikTok on government devices.

Frequently Asked Questions

What Is DeepSeek AI?

DeepSeek is a Chinese artificial intelligence platform that gained global attention in early 2025 when it released a series of large language models that rivaled the capabilities of leading American AI systems at a significantly lower cost. The platform offers a conversational AI chatbot and a set of underlying models that developers can use to build AI-powered applications.

Its flagship model, R1, drew particular attention for its reasoning capabilities and its ability to match or exceed the performance of more expensive American counterparts including ChatGPT and Google Gemini across a range of benchmarks. The company also permanently reduced the price of its V4 model by 75 percent in 2026, making it an attractive option for businesses and developers looking to cut AI costs.

Who Created DeepSeek AI?

DeepSeek was founded in 2023 by Liang Wenfeng, who also co-founded the Chinese quantitative hedge fund High-Flyer. The company is headquartered in Hangzhou, China, and operates as a subsidiary of High-Flyer. Unlike many AI companies that raise venture capital, DeepSeek has been funded primarily through High-Flyer's own resources, which has allowed it to operate with a level of independence unusual for a startup of its scale.

Liang Wenfeng has publicly stated that DeepSeek is focused on fundamental AI research, and the company has released several of its models as open-weight, meaning that the underlying model weights can be downloaded and used by third parties. That open availability has meant DeepSeek's technology has spread far beyond its own app and is embedded in numerous third-party products and services globally.

Is DeepSeek AI a Danger to Americans?

For government employees and federal contractors, the DHS order has answered that question clearly: the risks are serious enough to warrant a complete removal from all federal civilian systems. For ordinary American consumers, the picture is more nuanced. DeepSeek's own privacy policy confirms it stores user data, including chat content, uploaded files, and device information, on servers in China.

Under China's national security laws, that data can be accessed by Chinese authorities at any time without user notification. Security researchers have also identified code in the DeepSeek app that routes data to infrastructure tied to China Mobile, which US regulators have already determined poses a national security risk.

Cybersecurity experts advise that using any version of DeepSeek, including third-party deployments of its open-weight models, should come with a careful review of how data is handled and whether it is being transmitted back to DeepSeek's own servers.