Chinese artificial intelligence startup DeepSeek is developing its own semiconductor for AI computing, according to a Reuters report, as the company looks to reduce its dependence on external chip suppliers, including Nvidia Corp (NASDAQ:NVDA, XETRA:NVD) and Huawei.
The chip is being designed primarily for inference workloads, which involve running already-trained AI models to generate responses and complete tasks.
Reuters reported that DeepSeek’s focus on inference reflects growing demand for hardware optimized for deploying AI applications rather than training large language models.
Specialized inference chips can offer lower costs and improved power efficiency compared with traditional graphics processing units (GPUs), making them an increasingly important area of development as AI adoption expands.
According to Reuters, DeepSeek has been working on the chip project for about a year and has held discussions with chip design firms, semiconductor manufacturers and memory suppliers. The company has also begun hiring engineers to support the effort, the report added.
The move comes as Chinese technology companies seek to develop domestic alternatives amid restrictions on access to advanced foreign semiconductors.
After initially falling on the report, shares of Nvidia were little changed at about $195 in the early afternoon on Tuesday.