The global map of AI research is changing. A new MacroPolo analysis of top researchers at NeurIPS shows China has become a central training ground for elite AI talent, while the United States still remains the leading destination for where that talent works.
The shift matters because AI progress depends on people as much as models, chips, or capital. The countries that educate, employ, and retain top researchers are better positioned to shape the next phase of the technology.
China is training a much larger share of top AI researchers
In 2019, researchers with Chinese origins made up one-tenth of the most elite AI researchers. By 2022, that share had risen to 26%, almost matching the US figure of 28%.
That change reflects a major expansion of AI education and industry in China. According to MacroPolo senior research associate Ruihan Huang, China has dramatically expanded AI programs across its university system, with some 2,000 AI majors, while also building an AI industry able to absorb that talent.
The result is a larger pipeline from university study into frontier AI research. More students in computer science and other STEM majors have moved into AI, helping make Chinese researchers a major force in cutting-edge work.
The analysis is based on researchers who gave presentations and had papers accepted at NeurIPS, a top academic conference on AI. MacroPolo, the think tank of the Paulson Institute, examined national origin, educational background, and current work affiliation. Its earlier Global AI Talent Tracker was based on the 2019 NeurIPS conference, while the updated analysis looked at the December 2022 NeurIPS conference.
Graduate school is becoming a powerful retention tool
One of the clearest findings is that AI researchers often stay in the country where they receive their graduate degree. The numbers are striking: 80% of AI researchers who went to a graduate school in the US stayed to work in the US, while 90% of those who went to a graduate school in China stayed in China.
That pattern gives graduate education strategic importance. A country that invests in strong graduate-level institutions can do more than train local students. It can also attract overseas students and, in many cases, retain them after graduation.
This is especially important in the US-China context. The source article notes that the worsening relationship between the two countries has affected academia. News reports have described Chinese graduate students being interrogated at the US border or denied entry in recent years, while a Trump-era policy persisted. Pandemic-era border restrictions added another barrier.
The practical effect is straightforward: when fewer Chinese AI researchers come to the US to study and work, more of that talent may remain connected to China’s own universities and companies. The source does not quantify how many people were affected, but it identifies the policy and border climate as a possible factor limiting movement.
The US still attracts the most AI talent
The United States remains the top country for where elite AI researchers work. In both 2019 and 2022, it led the ranking. But its margin has narrowed.
In 2019, almost three-fifths of top AI researchers worked in the US. By 2022, only two-fifths worked there. China, meanwhile, had 28% of the top AI researchers working inside the country in 2022.
That does not mean the US has lost its central role. MacroPolo senior research associate AJ Cortese said elite researchers generally want to work in the most cutting-edge and dynamic places, where they can do strong work and be rewarded for it. In his view, the United States still leads in the AI ecosystem that attracts top talent, from leading institutions to companies.
Still, China’s position has strengthened. Compared with 2019, three more Chinese universities and one company, Huawei, entered the top tier of institutions producing AI research. The source also notes that Chinese AI companies are still generally considered behind US peers, including in comparable generative AI models, but appear to have started catching up.
More researchers are working in their home countries
The broader trend is not limited to China. The MacroPolo data shows that top-tier AI researchers are now more willing to work in their home countries. Indian AI researchers, like their Chinese peers, are increasingly ending up in their home country for work.
Previously, more than half of AI researchers worked in a country different from their home. Now the balance has shifted toward working in their own countries.
For countries trying to reduce dependence on foreign AI ecosystems, this is significant. Retaining domestic researchers can help build local research communities, strengthen universities and companies, and create a deeper base for future AI development.
Cortese described the preference clearly: most countries would rather have “brain gain” than “brain drain,” especially in a technical and complex field like AI. The challenge is that retaining talent is not simply a matter of training more people. Countries also need environments where researchers can do ambitious work and find compelling career opportunities.
What the AI talent shift means
The central story is not that one country has replaced another. It is that the AI talent landscape has become less concentrated. The US remains the strongest magnet for elite AI researchers, but China is training far more of them than before and keeping a growing share at home.
That changes the competitive picture. AI leadership depends on the ability to train researchers, retain graduates, create strong institutions, and build companies that can use advanced skills. The MacroPolo analysis suggests China has made gains across several of those areas.
For the global AI sector, the takeaway is clear: talent flows are becoming more local, and national research ecosystems matter more. The next phase of AI competition will not be shaped only by who has the best models today. It will also depend on where the next generation of elite researchers chooses to study, work, and stay.