China's DeepSeek developing own AI chip to cut reliance on Nvidia, Huawei
Translated from English, summarized and contextualized by DistantNews.
At a glance
- Chinese AI startup DeepSeek is reportedly developing its own AI chip, aiming to reduce its reliance on Nvidia and Huawei hardware.
- The chip is designed for inference, the process of using a trained AI model to generate responses, rather than for training new models.
- This move aligns DeepSeek with global AI developers like OpenAI and Anthropic, and supports China's push for domestic technological alternatives amid U.S. export controls.
Chinese AI startup DeepSeek is reportedly embarking on a significant strategic shift by developing its own artificial intelligence chip, according to three sources familiar with the matter. This initiative aims to lessen the company's dependence on chips from Nvidia and Huawei, which have been crucial for training and running its widely acclaimed AI models.
The chip is designed for inference - the stage of AI computing in which a trained model generates responses for users - rather than for training new models.
The new chip is specifically designed for the inference stage of AI computing โ where a trained model delivers responses to users โ rather than for the more intensive process of training new models. If successful, this venture into semiconductor development would mark a major evolution for DeepSeek, a company already recognized in China as a leading AI innovator.
DeepSeek gained international attention over a year ago with the release of two highly efficient AI models that achieved global popularity. While the company has historically focused on AI model advancements rather than commercialization, this move into chip development places it in a growing race. Huawei, despite U.S. export bans impacting its access to advanced chips, has captured a significant portion of China's domestic AI chip market. However, competitors like Alibaba and Baidu are also developing their own chips, intensifying market competition.
US export controls bar Chinese companies from buying Nvidia's most advanced chips, and Beijing has been pressing its technology champions to build domestic alternatives.
The effort by DeepSeek is still in its early stages, involving discussions with external partners in chip design, foundries, and memory sectors. The company has also been discreetly increasing its hiring of chip-design engineers. This pursuit of an in-house chip mirrors a global trend among AI developers, including OpenAI and Anthropic, seeking greater control over their hardware infrastructure and reduced reliance on external suppliers like Nvidia. For DeepSeek, this is particularly strategic given U.S. export controls that restrict Chinese companies from acquiring advanced Nvidia chips, reinforcing Beijing's drive for domestic technological self-sufficiency.
chip export controls were a challenge for the company.
Originally published by CNA in English. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.