Chinese AI developer DeepSeek is designing its own inference chip in a bid to reduce reliance on Nvidia and Huawei hardware. The early-stage project focuses exclusively on inference acceleration rather than model training, and the company has been quietly poaching hardware engineers and holding closed-door talks with chip designers and memory suppliers without posting public job openings. DeepSeek joins OpenAI and Anthropic in the growing trend of AI labs building custom silicon tailored to their workloads.

Reuters: China’s DeepSeek developing its own AI chip, sources say