The US Navy.

The US Navy has instructed its members to avoid the use of China’s artificial intelligence technology, as CNBC has learned.

In a warning issued by email to the “colleagues” on Friday, the Navy said that Deepseek’s AI should not be used “in any capacity” due to “potential security and ethical concerns associated with origin and use of the model “.

A US Navy spokesman confirmed the authenticity of the email and said he referred to the Department of the generative generative policy of the Navy.

The announcement followed the launch of Deepseek of its powerful new reasoning AI model called R1, which rivals OpenAi technology. The Deepseek model is open source, which means that any IA developer can use it. The Deepseek application has arising at the top of Apple App Store, Siempor’s Chatgpt of OpenAI and people in the industry have praised Your performance and reasoning capabilities.

Deepseek’s pronouncements He sprayed the capital markets on Monday due to the concerns that future products will require a less expensive infrastructure than Wall Street has assumed. Deepseek said at the end of December that his large language model took only two months and less than $ 6 million to build despite US chips exports. UU. To China three times in three years. That is a small fraction of the amount spent by Operai, Anthrope, Google and others.

AI chips manufacturers actions Nvidia and Broadcom Each one fell 17% on Monday, a route That eliminated the $ 800 billion combined in market capitalization. These actions led a 3.1% drop in Nasdaq.

The Navy warning landed days before.

“We would like to get your attention a critical update about a new model of the call Depseek,” said the email. The memorandum said that it is “imperative” that team members do not use the AI ​​DEPEEEK “for any task related to personal work or use.”

The email was sent on Friday morning to the OPNAV distribution list, which represents the operational navy, indicating that it was a memorandum of all hands. The warning was based on an advice of the division of the Cyber ​​Force Manger Naval Aircraph Division.

He said the recipients had to “refrain, install or use the Deepseek model in any capacity.”

Depseek said Monday that he would Temporarily limit user records “Due to large -scale malicious attacks” in their services, before resuming operations as usual.

President Donald Trumpwho assumed the position last Monday, saying The sudden increase of Deepseek “should be a call for attention” for the technology companies of the United States. Trump is currently trying to keep the application of Chinese social networks Tiktok alive in the United States, after legislators determined that the service must be forbidden or sold due to national security concerns. Trump was in favor of prohibiting Tiktok in his first administration before turning the matter.

The capitalist of risk, David Sacks, Trump’s AI and the cryptographic tsar, Posted in x On Monday, that Deepseek R1 “shows that the career of AI will be very competitive”, and added that it is “safe in the United States, but we cannot be complacent.” Goalwho has developed his own open source models called Llama, began four related to Depseek “War rooms“Within its generative department of AI, the information reported.

Alexandr Wang, CEO of Scale AI, He told CNBC Last week, that last model of Deepseek was “messy” and that its launch of R1 is even more powerful. He said he is “approximately with the best American models” and described the race between the United States and China as a “AI war.” The Wang company provides training data to Key OpenAi, Google and Meta.

Trump’s first great movement in AI arrived last week, when his administration announced a folded joint company Stargate Between OpenAi, Oracle and SoftBank to invest billions of dollars in AI infrastructure in the United States

LOOK: Depseek is another way that China is undermining US players