BEIJING: DeepSeek’s AI chip development is underway as the Chinese startup seeks to reduce its reliance on Nvidia and Huawei hardware, Reuters reported.
The chip is designed for inference, the stage in which a trained artificial intelligence model generates responses for users. The project began about a year ago and remains at an early stage.
The Hangzhou-based company has increased private hiring of chip-design engineers in recent months, according to the report.
DeepSeek has also held discussions with chip-design, foundry and memory companies. The company did not respond to a request for comment. DeepSeek has used both Nvidia and Huawei chips for its models.
The company said the foundation model behind R1 was trained on Nvidia’s H800, a China-market chip later blocked by Washington.
DeepSeek later leaned more heavily on Huawei, releasing its V4 model adapted for Ascend chips in April, according to the report.
Read: China Allows DeepSeek to Buy Nvidia H200 AI Chips
Reuters said a successful in-house chip would place DeepSeek among AI developers seeking tighter control over computing hardware and lower dependence on Nvidia.
The project faces manufacturing and memory hurdles because US export controls restrict Chinese access to advanced overseas foundries and high-bandwidth memory.
