Svmuu News: Sources familiar with the matter reveal that DeepSeek, a Chinese AI large-model company, is secretly developing an AI inference chip. The project was launched about a year ago and is primarily intended for model inference rather than training. Reports indicate that DeepSeek has been actively recruiting chip design engineers in recent months to advance the project’s development.
Currently, DeepSeek relies primarily on chips from NVIDIA and Huawei for model training and inference. If its in-house inference chip is successfully developed, it is expected to reduce dependence on external suppliers and improve the company’s ability to control hardware costs. However, the project is still in its early stages and faces restrictions on chip manufacturing and memory procurement due to U.S. export controls. (Reuters)