News

Is the gap between China and the U.S. in AI significant?

In recent years, both China and the U.S. have invested hundreds of billions of dollars in AI development. 

The U.S. has focused on corporate-led investments, with tech giants like Meta, Google, OpenAI, Intel, Apple, Oracle, 

and Amazon actively purchasing high-end chips, building computing power centers, and committing hundreds of billions annually. 

These efforts have produced landmark large models such as ChatGPT, Gemini, Grok, and Claude, which have set global 

industry benchmarks.  

China’s recent massive investments in AI computing power stem from two key directions: first, government-led initiatives to 

establish computing  power centers across regions; second, enterprises actively developing AI large models. Notable examples 

include Huawei’s Pangu model,Alibaba’s Tongyi Qianwen, Baidu’s ERNIE Bot, 

and the globally renowned DeepSeek developed by Hangzhou’s DeepSeek in 2026, 

as well as Kimi 3.0 launched by Moonshot AI in July 2026. These achievements have demonstrated China’s extraordinary progress 

in the AI field.  

The scale of both nations has endowed them with unparalleled advantages in AI technology, making them the two dominant powers 

far ahead of other countries and regions. This has naturally sparked debates about which country holds the upper hand in AI.  

Recently, I reviewed an interview with Jensen Huang, CEO of NVIDIA, whose insights into AI and leadership in high-end chips offer 

compellingperspectives. Huang decomposed the AI industry into five layers—energy, chips, infrastructure, large models, 

and applications—and compared China and the U.S. from a full-industry-chain perspective:  

Energy: The “underlying fuel” of AI

China’s electricity generation is 2.5 times that of the U.S. Huang emphasized: “The essence of AI competition is energy competition; 

electricity is the foundationof computing power.” 

China not only has abundant energy supply but also uses energy subsidies to reduce operating costs for AI data centers. 

In contrast, the U.S. faces energy costs 4–8 times higher than China, giving China a clear advantage.  

Chips

The U.S. possesses the world’s most advanced chips, excelling in both cutting-edge manufacturing processes and high-end chip integration. 

It can leverage global top-tier technology, such as ASML (Netherlands) and TSMC (Taiwan), to ensure access to the most advanced chips. 

NVIDIA’s GPU and CUDA ecosystem remain globally leading, outpacing China by several generations. Meanwhile, the U.S. continues to 

restrict China’sdevelopment in high-end chips, including limiting ASML’s sales of the most advanced EUV lithography machines to China 

and prohibiting NVIDIA from selling its highest-end chips in China. This has forced China to pursue breakthroughs independently.  

Infrastructure

Undoubtedly, China is the “infrastructure giant,” with faster and more efficient project implementation—completing infrastructure builds

in one-third of the time required by the U.S.  

Large Models

U.S. frontier models (e.g., OpenAI, Google) lead China by approximately 3–6 months, with clear advantages in general intelligence and 

multimodal capabilities. 

However, China’s open-source model ecosystem far surpasses the U.S., with models like DeepSeek, Alibaba’s Tongyi Qianwen, and 

Baidu’s ERNIE Bot gaining global influence. 

China’s open-source ecosystem is also significantly more active.  

Applications

The U.S. primarily focuses on general-purpose models, with applications concentrated in finance, arts, film/TV, and services. In contrast, 

China’s large models have taken a distinctly different path, 

fully embracing industrial integration. AI is widely applied in smart manufacturing, medical diagnostics, port operations, and 

mining development, continuously creating value and profoundly transforming humanity’s ability to harness nature.

Thus, in terms of comprehensive AI capabilities, the gap between the two nations is negligible.  

There is another point Huang did not mention: the cultivation of AI talent. Chinese-speaking talent holds a clear advantage—most top 

AI experts globally are Chinese or of Chinese descent. China’s 

annual production of millions of STEM graduates continues to strengthen its technological prowess across various fields.  

It is hoped that the U.S. will one day recognize that China and the U.S. can coexist harmoniously in the world, using their combined 

technological strengths to benefit humanity, rather than persisting in zero-sum games to maintain hegemony and 

suppress China’s development.