Is China Winning the AI Race?
▼ Summary
– China is positioned to become the 21st century’s AI superpower, leveraging its ability to mobilize whole-of-society resources for AI development and deployment.
– The US maintains a lead in top-tier AI research, with 50 of the 100 most cited publications in 2023, but its share has been declining over time.
– China leads in AI publications and patents, accounting for 22.6% of citations and 69.7% of AI patents globally as of 2023.
– The gap in top AI research talent between the US and China is narrowing, with the US share dropping from 59% in 2019 to 42% in 2022, while China’s rose from 11% to 28%.
– Chinese AI models, such as DeepSeek-V3 and Qwen 2.5-Max, surpass top US models in algorithmic efficiency, demonstrating their ability to achieve more with fewer resources.
When considering the global competition in artificial intelligence, many international observers believe China is steadily positioning itself as the dominant AI superpower of this century. While the United States has historically held a strong advantage, China’s focused strategy and resource mobilization present a formidable challenge that cannot be overlooked.
Western perspectives often emphasize America’s deep expertise in semiconductors, its advanced AI research, and massive data center investments. Warren Buffett famously advised never to bet against the United States, highlighting its unparalleled track record as an incubator for innovation. Still, betting against China’s technological ascent may be equally unwise. The country possesses the resources, motivation, and strategic coordination to deploy AI at a scale that could reshape global leadership.
Recent data underscores this shifting balance. China leads in both AI publications and patents, accounting for 22.6% of all citations in 2023, compared with Europe’s 20.9% and the United States’ 13%, according to Stanford University’s Artificial Intelligence Index Report. Chinese institutions also filed 69.7% of all AI patents worldwide. Although the U.S. still excels in the most influential research, with 50 of the top 100 cited publications versus China’s 34, its relative share has been declining steadily.
Talent distribution tells a similar story. The U.S. continues to attract more top-tier AI researchers, but the gap is shrinking. In 2019, 59% of the world’s leading AI experts worked in the United States, while only 11% were based in China. By 2022, those figures had shifted to 42% and 28%, respectively. Tighter U.S. visa policies under the Trump administration may further encourage Chinese researchers abroad to return home, accelerating this trend.
In terms of breakthrough models, American organizations produced 40 of the world’s most significant AI systems in 2024, compared to 15 from China. However, Chinese researchers have demonstrated remarkable efficiency, achieving high performance with fewer resources. Models like the open-source DeepSeek-V3 and Alibaba’s Qwen 2.5-Max now rival or exceed top U.S. models in algorithmic efficiency, proving that China’s approach, emphasizing optimization and practical application, is yielding tangible results.
(Source: Technology Review)
