China’s DeepSeek is developing its own chip to help power artificial intelligence systems, Reuters reported, citing unnamed sources.
The chip is designed for inference — the stage in AI when models are trained and begin generating responses to users’ queries — rather than for building new models themselves, Reuters reported on Tuesday. DeepSeek didn’t respond to a request for comment on the report.
Nvidia Corp. shares fell 2.2 per cent in early trading on Tuesday.
Virtually every major AI developer is said to be developing its own chip in an effort to gain an edge by tailoring hardware that’ll work most efficiently with their systems. Just last month, OpenAI unveiled its first custom chip developed in partnership with Broadcom Inc. Amazon.com Inc. is in talks to sell custom chips for use in other companies’ data centers. Alphabet Inc. has seen significant demand for its own chips known as tensor processing units, or TPUs for short.
DeepSeek has been at the heart of China’s effort to compete globally on AI. It developed an AI model last year that stunned the tech world, largely by how few chips and resources the company said it expended to build a cutting-edge service. The breakthrough demonstrated that Chinese companies could compete with the best of Silicon Valley despite US export restrictions on advanced hardware.
Published on July 7, 2026