Key insights
- China's potential to rival the US in AI depends on compute power, which is linked to energy, data centers, and semiconductors. While the US currently leads in AI compute, China's energy production advantage could allow it to catch up by 2035, even with less efficient chips. However, bottlenecks in semiconductor technology and export controls remain risks. The US leads in chips and software, while China dominates in power and manufacturing scale.
Investing.com – China could potentially rival and even surpass the United States in artificial intelligence over the next decade, but doing so will depend less on chips and more on power.
According to a recent Bernstein analysis, AI leadership is ultimately a function of compute power, which scales with energy, data centers, and semiconductor capability. The U.S. currently holds a dominant lead with roughly 35 zettaFLOPS of AI compute versus China’s 5 zettaFLOPS, or about 15% of the U.S. level.
However, China’s structural advantage lies in energy. The country already generates more than twice as much electricity as the U.S. and is adding capacity at an unmatched pace, over 500 gigawatts annually, more than the rest of the world combined. This gives it the ability to scale data centers aggressively, even if its chips remain less efficient.
Bernstein estimates China could match U.S. compute capacity by 2035 if it compensates for weaker semiconductors with sheer scale. That would require massive investment, nearly $1 trillion in AI data center capex, alongside a rapid expansion in power infrastructure and battery storage.
In a more aggressive scenario where power remains the only constraint, China could even exceed U.S. compute capacity, potentially reaching more than three times the U.S. level by 2035.
Still, key bottlenecks remain. China lags in advanced semiconductor technology, with domestic AI chips currently operating at roughly one quarter the efficiency of U.S. counterparts, though that gap could narrow to more than 50% by 2035. Export controls and limited access to cutting edge manufacturing tools continue to pose risks.
The race, therefore, is asymmetric, the U.S. leads in chips and software, while China dominates in power, manufacturing scale, and cost efficiency. If energy proves to be the ultimate constraint on AI growth, China’s advantage could become decisive.
The outcome remains uncertain, but the analysis suggests that AI supremacy will be determined as much by megawatts as by microchips.