Chinese AI startup DeepSeek is developing its own AI chip, according to three people familiar with the matter, as reported by Reuters and echoed by multiple outlets. The reported chip is intended for inference, the stage where a trained AI model generates responses for users, rather than for training new models. Sources say the move could reduce DeepSeek’s reliance on external chip suppliers, including Nvidia and Huawei, which it has depended on to train and run its widely used models.
DeepSeek rose to global attention more than a year ago after releasing highly efficient AI models that gained widespread popularity. The company is known for focusing on AI model performance and breakthroughs rather than commercializing chip technology. If DeepSeek’s chip development succeeds, it would represent a strategic expansion into semiconductors. Several reports also note the potential competitive implications in China’s AI chip ecosystem, including the possibility of adding pressure to established players.
The reports cite unnamed sources and do not provide timelines or details on the chip’s specifications or manufacturing partners.
