[Bloomberg Column] Why the US Is Falling Behind in the AI Power Race Against China

AI's Expanding Influence from Military to Economy
US Leads China in Cutting-Edge Semiconductor Development

The competition among great powers ultimately comes down to a battle for technological dominance. A nation that holds the lead in the core technologies of an era often ends up setting the rules for that era. In recent weeks, both the United States and China have unveiled their strategies to win the race in artificial intelligence (AI). Their goal is to secure breakthroughs in AI and leverage them strategically. This could become a factor that shakes the future balance of international power. For now, the United States holds the lead, but President Donald Trump has undermined existing policies that could contain China's rise, putting America's position at risk.

AI is already transforming military combat methods and economic systems as a whole. Especially when considering the potential of artificial general intelligence (AGI) capable of human-level thinking, the geopolitical stakes of the AI supremacy race are becoming increasingly significant.

Last month, the Trump administration released a 23-page report detailing its approach to this technological revolution. According to the report, the Trump administration's plan focuses on deregulation to accelerate domestic innovation, expanding infrastructure such as building data centers, and strengthening America's technological superiority and influence through export controls and other measures. President Trump emphasized that the ultimate goal of this strategy is to secure "global technological dominance that is beyond doubt or challenge" for years to come.

Beijing's position is different. Just three days after the U.S. released its report, China immediately announced a counter-strategy consisting of 13 provisions. China has long made AI a core part of its economic and industrial strategy. The newly announced plan focuses less on the technology itself and more on AI diplomacy. The emphasis is on China taking the lead in designing the norms and order of the global environment in which AI is developed and utilized.

Overall, the United States is ahead in developing cutting-edge AI semiconductor models. U.S. government officials estimate that the country maintains a technological lead of up to two years. American AI companies such as OpenAI, Google, Anthropic, and xAI are still regarded as "world-class" in the field of advanced innovation. While Chinese companies like DeepSeek are rapidly closing the gap, the U.S. is still considered to have the upper hand for now.

OpenAI recently released its next-generation model, ChatGPT-5. CEO Sam Altman described the model as "like carrying a team of PhD-level experts in your pocket." However, the AI race is not simply a speed contest to see who crosses an imaginary technological finish line first. From this perspective, America's position does not seem particularly solid.

China has strengths in the broad practical application of AI, that is, integrating AI across its economy and national governance. Colin Kahl and Jim Mitre wrote in a Foreign Affairs article that China has begun to introduce robots and "extreme automation" on a large scale to enhance its world-class manufacturing capabilities. In addition, China is actively using AI to strengthen internal regime control.

In this context, China's authoritarian system actually works to its advantage. Beijing can easily control and utilize private sector innovation, whereas the U.S. government faces much greater constraints in doing the same. Furthermore, China is spreading its technology and standards to developing countries through initiatives like the "Digital Silk Road" and various national programs. Chinese companies are excelling in "open-weight" based AI models rather than "proprietary" ones, and this open structure allows startups to more easily modify and utilize the models. Even if China's AI models do not lead in cutting-edge technology, the country is rapidly building a foundation to win the competition for market share and AI utilization.

The second challenge is that America's AI strategy is based on a risky bet in the Middle East. Last spring, President Trump signed major agreements with Saudi Arabia and the United Arab Emirates (UAE) to bring these countries into the U.S. AI ecosystem and build large-scale data centers locally. While this bet could yield significant returns, it could also backfire severely if it provides China with an "indirect access" route to the latest semiconductor technology or if the critical infrastructure becomes a target for Iranian missile and drone attacks.

An even bigger problem is that these risks are intertwined with another decision by President Trump. He has dismantled safeguards designed to prevent core technologies from falling into Chinese hands. Obsessed with a trade deal with China, he has put on hold plans to strengthen and update U.S. semiconductor export controls.

More recently, President Trump has also struck deals that could create perverse incentives. He has effectively forced semiconductor companies such as Nvidia and AMD to pay a portion of their sales to the U.S. government if they sell advanced chips to China, which could actually encourage more business with China.

President Trump's aides argue that these measures will further strengthen U.S. AI leadership by maximizing America's influence and market share. However, it is also possible that these actions could erode America's innovation advantage.

Even Chinese companies admit that U.S. export controls have slowed their technological progress. And as Ben Buchanan, former White House National Security Council (NSC) official, pointed out, President Trump is now effectively opening the door for China to access the chips it needs to refine its capabilities in key areas such as cyber warfare and drone attacks.

President Trump is right that the race to dominate AI could become the most consequential contest of the 21st century. But if he believes the U.S. can win that fight while voluntarily giving up some of its most important weapons, he is clearly mistaken.

Hal Brands, Professor of International Relations at Johns Hopkins University and Bloomberg Opinion columnist

This article is a translation by Asia Economy of the Bloomberg column "How the US Could Lose the AI Arms Race to China."