DeepSeek develops AI chip for inference, pressuring Nvidia shares

DeepSeek is developing its own AI chip for inference, a move that could reduce the Chinese startup’s reliance on Nvidia and Huawei hardware as US export controls tighten access to advanced processors. The effort is still at an early stage, with the Hangzhou-based company working with outside chip-design, foundry and memory partners and quietly hiring chip engineers, according to people familiar with the matter. If successful, the project would mark a strategic shift for DeepSeek, which built its reputation on efficient AI models rather than chipmaking. The report also rattled markets: Nvidia shares fell about 2% in premarket trading, helping push the Nasdaq lower at the open.