Hal Brands explores the prospects for maintaining America’s artificial intelligence dominance.
“The future of American warfare is here, and it’s spelled AI,” Defense Secretary Pete Hegseth declared in early December. It wasn’t hyperbole.
Just weeks later, the midnight raid that snagged Venezuelan dictator Nicolás Maduro was guided by artificial intelligence. The Pentagon reportedly used Anthropic’s Claude model, paired with Palantir’s command-and-control software, to map the physical and political landscape of that daring mission. …
… American success in the age of AI is not inevitable. It depends on the ability of the U.S. to adapt quickly and respond strategically to a transformative technology. …
… America’s lead at the technological frontier is sizable but insecure. The top U.S. AI models, produced by firms such as Anthropic and OpenAI, have been months ahead of Chinese rivals DeepSeek and others, not least because of America’s massive edge in computing power. That advantage, in turn, rests partly on an architecture of semiconductor export controls built to hold back Beijing.
But Chinese firms are striving to close that gap. …
… A second challenge: keeping the world from being conquered by Chinese AI. U.S. models are more advanced, but China’s cheaper, easier-to-adapt models are often more attractive in the Global South. Beijing aggressively exports these models as part of a larger tech bundle that features everything from hardware to financing. The U.S. could lose the fight for global market share — and global influence — even if its innovation remains unsurpassed. …
… The third challenge requires securing the global AI stack amid surging threats and disorder. The UAE and other Gulf states, which are making generational bets on AI, will require massive investments in hardening data centers and other infrastructure, by burying it or surrounding it with counter-drone and counter-missile defenses. The UAE’s deepening security ties with Israel and the U.S. look especially valuable in this light.
