Chinese AI startup DeepSeek is reportedly planning to develop its own silicon for data center inference chips. This move, which has been in development for about a year, aims to lessen the company's dependence on external chip suppliers like Huawei and Nvidia. The United States export ban on advanced AI chips to China has significantly impacted Nvidia's market share there, creating an opening for domestic players.

DeepSeek is actively seeking hardware and silicon partners and hiring engineers for this initiative. While Huawei currently dominates China's data center chip market, other tech giants like Alibaba and Baidu are also exploring in-house chip development. This trend mirrors similar efforts by US AI companies such as OpenAI, which recently partnered with Broadcom on its first inference chip, Jalapeño, and Anthropic, which is also exploring custom chip designs.

For DeepSeek and other AI firms, designing custom chips offers a path to greater control over their technology stack and a potential advantage in securing compute resources. With demand for AI services increasing, access to dedicated hardware for inference is becoming a strategic necessity.