U.S. Navy bans use of DeepSeek due to ‘security and ethical concerns’

The U.S. Navy has instructed its members to avoid using artificial intelligence technology from China’s DeepSeek, CNBC has learned.

In a warning issued by email to “shipmates” on Friday, the Navy said DeepSeek’s AI was not to be used “in any capacity” due to “potential security and ethical concerns associated with the model’s origin and usage.”

A spokesperson for the U.S. Navy confirmed the authenticity of the email and said it was in reference to the Department of the Navy’s Chief Information Officer’s generative AI policy.

The announcement followed DeepSeek’s release of its powerful new reasoning AI model called R1, which rivals technology from OpenAI. The DeepSeek model is open source, meaning any AI developer can use it. The DeepSeek app has surged to the top of Apple’s App Store, dethroning OpenAI’s ChatGPT, and people in the industry have praised its performance and reasoning capabilities.

DeepSeek’s pronouncements rocked the capital markets on Monday due to concerns that future AI products will require less-expensive infrastructure than Wall Street has assumed. DeepSeek said in late December that its large language model took only two months and less than $6 million to build despite the U.S. curbing chip exports to China three times in three years. That’s a tiny fraction of the amount spent by OpenAI, Anthropic, Google and others.

Shares of AI chipmakers Nvidia and Broadcom each dropped 17% on Monday, a route that wiped out a combined $800 billion in market cap. Those stocks led a 3.1% drop in the Nasdaq.

The Navy’s warning landed days earlier.

“We would like to bring to your attention a critical update regarding a new AI model called DeepSeek,” the email said. The memo said it’s “imperative” that team members do not use DeepSeek’s AI “for any work-related tasks or personal use.”

The email was sent on Friday morning to the distribution list OpNav, which stands for Operational Navy, indicating it was an all-hands memo. The warning was based on an advisory from Naval Air Warcraft Center Division Cyber Workforce Manger.

It said recipients were to “refrain from downloading, installing, or using the DeepSeek model in any capacity.”

DeepSeek said on Monday it would temporarily limit user registrations “due to large-scale malicious attacks” on its services, before later resuming operations as usual.

President Donald Trump, who took office last Monday, said the sudden rise of DeepSeek “should be a wake-up call” for America’s tech companies. Trump is currently trying to keep Chinese social media app TikTok alive in the U.S., after lawmakers determined that the service must be banned or sold due to national security concerns. Trump was in favor of banning TikTok in his first administration before flip-flopping on the matter.

Venture capitalist David Sacks, Trump’s AI and crypto czar, posted on X on Monday that DeepSeek R1 “shows that the AI race will be very competitive,” adding that he is “confident in the U.S. but we can’t be complacent.” Meta, which has developed its own open-source models called Llama, started four DeepSeek-related “war rooms” within its generative AI department, The Information reported.

Alexandr Wang, CEO of Scale AI, told CNBC last week that DeepSeek’s last AI model was “earth-shattering” and that its R1 release is even more powerful. He said it’s “roughly on par with the best American models” and described the race between the U.S. and China as an “AI war.” Wang’s company provides training data to key OpenAI, Google and Meta.

Trump’s first big move in AI came last week, when his administration announced a joint venture dubbed Stargate between OpenAI, Oracle and SoftBank to invest billions of dollars in AI infrastructure in the U.S.