The executives cited limited computing capacity and U.S. export controls on advanced chips as key obstacles. U.S. infrastructure, they said, is one to two orders of magnitude larger, while Chinese companies must devote most of their resources to ongoing projects rather than large-scale expansion.
A more optimistic view came from Yao Shunyu, a former OpenAI researcher and now chief AI scientist at Tencent. He believes China could catch up within three to five years, identifying the lack of advanced chip manufacturing equipment as the main technical bottleneck.
The cautionary remarks follow a strong week for Chinese AI startups in the capital markets. Zhipu AI and MiniMax Group together raised more than $1 billion in Hong Kong, with MiniMax shares doubling on their first day of trading.
Outlook
China’s AI development is increasingly constrained not by a lack of talent, but by physical factors. As long as large-scale compute infrastructure, access to advanced chips, and critical manufacturing equipment remain concentrated in the United States, China’s progress in 2026 and 2027 is likely to be evolutionary rather than breakthrough-driven.
