US bans DeepSeek AI from government use as tech tension with China continues
DeepSeek is a powerful, popular and relatively inexpensive artificial intelligence model out of China.
US bars DeepSeek AI from government use as tensions with China in tech persist
The US has directed that all artificial intelligence (AI) technology from the Chinese company DeepSeek be removed from most government systems.
The Department of Homeland Security (DHS) has prohibited all use of DeepSeek across civilian government agencies, along with any technology developed with the help of its AI models.
As a powerful open-source AI model — meaning it is free to access and cheaper to use — DeepSeek has been viewed as a disruptive player in the continuing race over AI technology.
Although the DHS notice did not spell out the specific security worries tied to DeepSeek, the White House has this year criticised AI companies from China.
The DHS, DeepSeek, and the Chinese embassy in Washington DC have been approached for comment.
The DHS announced the order publicly on Monday, though the ban had already taken effect at the end of July.
The directive applies to civilian agencies, including the Department of the Treasury, the Department of State and the Department of Justice. Agencies within the Department of Defense are not included.
The ban comes after criticism of Chinese AI from the Trump administration and from US companies.
White House tech lead Michael Kratsios wrote in an internal memo earlier this year that "malicious exploitation" of American AI firms had been taking place, largely by rival companies in China.
Anthropic and OpenAI, both of which are expected to go public in the coming months, have also said their AI models were effectively copied by some Chinese AI firms, including DeepSeek.
US president Donald Trump has made it clear that he views China as the main US competitor in AI development, which he sees as vital to the country’s future.
DeepSeek has surged in popularity since its first public release in 2025, with demand for its models so high that the company’s servers have sometimes gone down.
Although DeepSeek’s free chatbot app is among the most downloaded apps in the Apple App Store rankings, its AI models are also used by US technology companies. Firms such as Snowflake, Vercel and OpenRouter all include DeepSeek in products they offer to customers.
Startups have been drawn to open-source models such as DeepSeek.
For a technology or AI developer, using DeepSeek in their work costs only a fraction of what a closed model like Google’s Gemini, Anthropic’s Claude or OpenAI’s Codex would.
Likewise, the cost DeepSeek has incurred to build its models has been vastly lower than what US AI companies have spent developing their own.
Last year, DeepSeek said it cost about $6m (£4.5m) to train the first version of its model, while the costs linked to OpenAI’s GPT-4 were around $100m. Those disclosures sent tech stocks falling.
The latest version of DeepSeek was launched last week. Users of the model have praised it online, saying it outperforms even OpenAI’s most advanced AI tools on common tasks for AI systems such as computer software coding.