HANGZHOU, China: Chinese artificial intelligence startup DeepSeek is developing its own AI chip for inference, marking a major step toward reducing its dependence on Nvidia and Huawei as it expands beyond AI model development, according to three people familiar with the matter.

The chip is being designed for inference, the stage of AI computing in which trained models generate responses for users, rather than for training new AI models, the sources said.

If successful, the move would represent a significant strategic shift for DeepSeek, which has become one of China's leading AI companies, and could also intensify competition in the country's AI chip market.

Shares of U.S.-based Nvidia slipped about 1.6 percent in premarket trading.

"Nvidia is at zero in China and staying there. DeepSeek has almost no chance of selling silicon outside of China unless it gets access to leading-edge manufacturing," said analyst Richard Windsor of Radio Free Mobile, adding that the development does not affect the chipmaker.

DeepSeek gained international attention more than a year ago after releasing two highly efficient AI models that achieved global popularity, surprising many in Silicon Valley and Washington. The company has largely focused on advancing AI models rather than commercializing its technology.

Although Huawei's chips remain less advanced than Nvidia's top products, U.S. export restrictions have helped the Chinese company capture about half of China's US$50 billion AI chip market by supplying DeepSeek and other major domestic AI developers.

However, Huawei is facing growing competition as Alibaba and Baidu continue developing their own AI chips.

DeepSeek's chip project remains in its early stages. The company has spent about a year working with chip designers, foundries and memory suppliers while quietly increasing recruitment of chip-design engineers without advertising the roles publicly, according to the sources.

The people spoke on condition of anonymity because the project has not been publicly disclosed. DeepSeek did not respond to a request for comment.

The company is joining a broader industry trend of AI developers designing their own chips to reduce reliance on Nvidia. Last month, OpenAI unveiled Jalapeno, its first custom inference chip developed with Broadcom, while Anthropic has also explored building proprietary AI processors.

DeepSeek's efforts come as U.S. export controls continue limiting Chinese companies' access to Nvidia's most advanced chips, prompting Beijing to encourage domestic alternatives. It was reported in June that the company was preparing to raise $7 billion in its first external funding round, valuing it at between $52 billion and $59 billion.