At a recent conference in Beijing, artificial intelligence experts from China stated that it is unlikely the country will surpass the US in the coming years, even considering the progress made in this field. The main reason for this lag, experts say, is the lack of resources in the semiconductor sector and equipment for chip manufacturing, which are the foundation for modern AI models.
Lin Junyang, the technical leader of the Qwen team at Alibaba, noted that the chances of any Chinese company surpassing American giants like Google DeepMind or OpenAI in the next three to five years are less than 20%. He added that this is already an "extremely optimistic" scenario, as computational power in the US significantly exceeds that of China, and local companies are forced to operate at the limit of their capabilities to meet growing demand.
Tang Jie, co-founder and chief scientist of Zhipu AI (Z.ai), agrees with this viewpoint. He emphasized that the success of Chinese open AI models does not reflect the full potential of American developments, many of which remain closed. "The gap between China and the US may actually be widening, as there are models in the US that have not yet been made public," he added.
Experts believe that external tests confirm that Chinese AI models are gradually closing the performance gap compared to American ones. While open access to Chinese models promotes their international spread, it does not guarantee that China will be able to surpass the US overall.
On the other hand, Yao Shunyu, chief AI scientist at Tencent and former lead researcher at OpenAI, expressed a more optimistic view. He believes that in 3-5 years, a Chinese company could emerge as a leader in the AI field, considering China's experience in rapidly scaling technologies, such as in electric vehicles and manufacturing. However, he argues that to achieve this goal, the country needs to overcome several obstacles: develop its own EUV systems for advanced chip production, implement AI in the corporate sector, and increase investments in fundamental research.
Experts also noted that a new generation of researchers born in the 1990s and 2000s could make a significant contribution to the country's achievements if the government continues to create conditions for innovation. With increased transparency in competition between tech giants and startups, as well as reduced pressure on AI implementation, young talents will be able to focus on scientific breakthroughs.
At the end of last year, the US administration allowed the sale of Nvidia H200 chips to China, lifting previously imposed restrictions. However, Beijing has approached several companies with a request to temporarily suspend orders and is actively promoting a strategy to replace American chips with local alternatives.
