America’s AI Success Relies on Chinese Experts as US Races to Beat CCP - Newsweek

President Donald Trump has made keeping China out of America's most advanced technology central to his strategy to win the AI race.

But as he prepares to meet Chinese President Xi Jinping in Washington, the United States is facing a problem of its own making: Some of the researchers helping drive America's AI advantage came from China.

Xi arrives at the White House on September 24 for a state visit where artificial intelligence will be among the dominant subjects. Washington and Beijing have spent weeks discussing whether to fold a long-delayed intergovernmental AI dialogue, agreed to at the two leaders' May meeting in Beijing, into a broader economic session held in the days beforehand.

Trump has described the competition in absolute terms. "WHOEVER WINS AI, WINS!" he posted to Truth Social this week, arguing that guardrails on the technology would serve Beijing's interests.

That advantage relies heavily on foreign-born talent, and China is one of the countries clawing some of it back.

Yao Shunyu, a Tsinghua- and Princeton-trained researcher who left OpenAI in 2025 to join Tencent, was named the company's chief AI scientist three months later. He is now pushing for China to build up its own long-term AI research.

"My personal goal is that in China we should establish a long-term AGI organization," Yao said in remarks CNBC translated from Mandarin.

Another case is Yang Zhilin. He finished his undergraduate degree at Tsinghua University in 2015, earned a doctorate at Carnegie Mellon University, and returned to China in 2023 to found Moonshot AI. The company has since raised more than a billion dollars and released the Kimi model series, now among the strongest open-source models available, with American startups building on Kimi and Alibaba's Qwen for reasons of cost and performance.

Years of friction preceded moves like theirs. The Carnegie Endowment for International Peace, in a study by senior fellow Matt Sheehan and research assistant Sophie Zhuang, describes visa uncertainty, indictments that later collapsed and a broad sense of suspicion as conditions that accumulated after 2018.

"The reasons why any individual researcher returns home are varied and often personal," Sheehan and Zhuang wrote. "But the recent geopolitical tumult has created new obstacles and pressures for Chinese researchers who want to remain in the United States."

Carnegie tracked 100 top-tier Chinese-origin AI researchers working at U.S. institutions in 2019 and found that 10 had moved to Chinese companies or universities by 2025. The 10 were not minor figures. Two founded startups, two took leadership roles at Chinese technology companies, and five became professors at leading Chinese universities.

Retention stayed high, the authors said. Eighty-seven of the 100 remained at American institutions, and three went to institutions abroad. Of those who stayed, 41 work for U.S. companies, 40 hold professorships at U.S. universities, and six are finishing doctorates or postdoctoral appointments.

More than half of the ones in the private sector work for Google, Amazon, Apple, Meta, Microsoft, Nvidia or Tesla. Three started companies of their own in the United States.

Sheehan and Zhuang call that an enduring strength for U.S. AI competitiveness, built substantially on Chinese talent.

"Studies of the top AI research papers have shown the authors originally from China contributing as much, if not more, to American AI output than authors from the United States," they wrote.

Recruiting the next cohort is a harder problem for the U.S., and the numbers on where top researchers end up have started to shift.

The Paulson Institute, which built the dataset Carnegie later updated, found that researchers with Chinese undergraduate degrees made up 47 percent of its 2022 sample of top-tier AI authors, up from 29 percent in 2019. Over that same period, the share of those researchers working at U.S. institutions fell from 59 percent to 42 percent, while the share working in China rose from 11 percent to 28 percent.

"There are some signs that America's ability to attract Chinese talent is waning," Sheehan and Zhuang wrote.

Carnegie lays out the accumulated friction behind that shift. Proposed visa restrictions beginning in 2018, including public discussion of barring Chinese students outright, left researchers unsure whether they could finish degrees and stay to work. Indictments of Chinese researchers made news and then, in many cases, fell apart.

A 2021 survey of university researchers who identify as Chinese, cited in the study, found 42 percent said they felt racially profiled by the U.S. government. Daily flights between the two countries are still running below 30 percent of pre-pandemic levels.

Departures are being counted. Princeton researchers tallied about 50 tenure-track scholars of Chinese descent leaving American universities for China in the first half of 2025, on top of more than 850 since 2011. A Stanford Center on China's Economy and Institutions paper examining 1.2 million scholars from 2003 to 2022 found the annual number of ethnically Chinese scientists shifting their affiliation to China rose from 31 to 1,409 over that span.

The pattern has reached elite researchers who are not Chinese, too. Omar Yaghi, a chemistry Nobel laureate born in Jordan, left his faculty position at the University of California, Berkeley, for Tsinghua University, where he is to lead an institute using AI in materials discovery.

Newsweek reached out to the White House for comment.

Nobody studying the competition agrees on how it would be scored.

Chen Yixin, China's minister of state security, recently wrote that American AI models function as a "propaganda war and a cognitive war" against his country. Beijing's foreign ministry, quoted by the BBC, said last week that framing the technology around threat and confrontation does nothing for global AI governance.

American officials have been just as direct about the stakes. Speaking at a Washington event this month, Treasury Secretary Scott Bessent said, "There is no day after tomorrow if China wins at this."

"I think winning means first having the most advanced AI companies, which the U.S. currently does," Chris Miller, a Tufts University professor and author of Chip War, told Newsweek. The second piece, he said, is deploying the technology fast enough to capture what it produces.

David Lin, a former State Department and Central Intelligence Agency analyst now at the Special Competitive Studies Project, put the test elsewhere, in whether a country turns technical capability into productivity gains, scientific discovery and functioning public services.

"The goal then should not be dominance at any cost," he told Newsweek.

Carnegie's own recommendation runs shorter. Sheehan and Zhuang call for investment in American high schools, research grants and visa certainty for graduate students, and conditions under which the best AI researchers in the world, Chinese ones included, want to build their careers in the United States.

None of it is on the schedule for September 24.

Contact Newsweek editors on this story: Jason Lemon and Sam Wilson.