DeepSeek Withholds Details of Upcoming AI Model from US Companies, Showcases Prototype to Huawei

DeepSeek's Flagship Model Update Raises Concerns

DeepSeek, a Chinese artificial intelligence lab, has deviated from industry standard by not sharing its upcoming flagship model for performance optimization with leading U.S. chipmakers, Nvidia and Advanced Micro Devices (AMD). Instead, the lab has granted early access to domestic suppliers, including Huawei Technologies.

Pre-release Access Typically, AI developers share pre-release versions of major models with leading chipmakers to ensure software compatibility. However, DeepSeek has not followed this practice for its forthcoming model, which was expected to be released around the Lunar New Year holiday. The lab has provided access to Chinese chipmakers, including Huawei, several weeks ahead of the release date.

Impact on U.S. Chipmakers The decision may have minimal impact on Nvidia and AMD for general data accelerators, as most enterprises do not use DeepSeek models. However, the move is seen as part of a broader strategy by the Chinese government to keep U.S. hardware and models disadvantaged in China.

Read also: Treasury Yields Experience Largest Increase in Two Weeks Following Release of Labor Market Data

Export Control Concerns The development comes as a senior Trump administration official revealed that DeepSeek's latest AI model was trained on Nvidia's most advanced chip, Blackwell, using a cluster in mainland China. This move appears to violate U.S. export controls. DeepSeek may seek to remove technical indicators and publicly claim that it used Huawei's chips to train its model.

Market Impact DeepSeek's models have been downloaded over 75 million times on the open-source platform Hugging Face since January 2025, fueling the rise of Chinese open-source models competing with U.S. AI labs. The rapid growth of Chinese open-source models has intensified debate in Washington over exporting advanced U.S. AI chips to China.