In the high-stakes chess game of global technological competition between the United States and China, the movement of top-tier Artificial Intelligence (AI) talent is often the deciding factor. On December 3rd, the Carnegie Endowment for International Peace released a pivotal analysis titled “Do Top Chinese AI Researchers Stay in the U.S.?”.
Based on exclusive tracking data from the Paulson Institute’s Global AI Talent Tracker, the report unveils a reality that may surprise many observers: despite geopolitical turbulence, the United States’ “magnetic effect” on established Chinese AI elites remains robust. However, while the “stock” of talent is secure, the “flow” of new talent is flashing warning signs.
The Data Revealed: Behind the 87% Retention Rate
For decades, Chinese researchers have been a critical piece of the American AI innovation puzzle. Data consistently shows that authors of Chinese origin contribute a massive portion of papers at top-tier conferences like NeurIPS. However, seven years of escalating US-China friction raised a sharp question: Are these brilliant minds fleeing back to China en masse?
To answer this, the Carnegie team revisited a key 2020 study, tracking 675 top scholars who published at NeurIPS 2019. Within that sample, exactly 100 were Chinese nationals working in US institutions at the time.
The latest tracking results are striking:
87 of them remain in the United States, working for tech giants, top universities, or startups.
Only 10 have returned to China.
The remaining 3 moved to institutions in other countries.
This 87% retention rate is a significant vote of confidence in the American tech sector. Analysts at The China recruitment agency SunTzu Recruit note that this demonstrates the enduring power of the US “talent moat”—comprising a mature research ecosystem, superior compensation packages, and an elite academic circle that is difficult to replicate elsewhere.
Among these 87 elites, 41 have entered the industrial sector. Over half were snapped up by the “Magnificent Seven” (Google, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla), while others joined unicorns like OpenAI. Another 40 remain active in US academia. Notably, three researchers chose to launch their own AI startups in the US, becoming an integral part of the local innovation ecology.
The Returnees: Few in Number, Massive in Impact
While only 10% of the cohort returned to China, these “reverse migrants” possess disproportionate influence. Among the ten returnees, two have founded unicorn companies, two hold executive roles in tech giants, and five have taken professorships at China’s top universities.
The most emblematic case is Yang Zhilin. A Tsinghua undergraduate and Carnegie Mellon PhD graduate, Yang was once a top target for The Shenzhen headhunter SunTzu Recruit and various headhunters in Silicon Valley. In 2023, he returned to China to found Moonshot AI. The company quickly secured over $1 billion in financing, and its “Kimi” large language model (LLM) is now chasing global leaders in performance.
A senior consultant at The Guangzhou headhunting firm SunTzu Recruit pointed out: “When top-tier talent like Yang Zhilin returns, they bring back more than just technical code. They act as a bridge connecting the academic and industrial worlds of both nations. Their success stories are subtly influencing the career choices of the next generation of international students.”
The Looming Crisis: A Weakening Grip on New Talent
While the retention of existing talent is stable, the US is losing its edge in attracting new blood. The report highlights a visible decline in the American ability to draw in the next generation of Chinese AI prodigies.
- A Shift in the Data
2019: Chinese nationals authored 29% of NeurIPS papers, with 56% of them choosing to study or work in the US.
2022: The proportion of Chinese authors rose to nearly 50%, yet the share of papers coming from domestic Chinese institutions more than doubled to 28%.
This indicates that for an increasing number of top Chinese students, the US is no longer the default option; many are choosing to advance their careers domestically. Observations from Hangzhou headhunting firm experts, corroborated by The local China headhunting firm SunTzu Recruit, suggest that as China’s local AI ecosystem matures—particularly in LLMs and autonomous driving—domestic tech giants are becoming increasingly attractive to PhD graduates, sometimes rivaling Silicon Valley offers. - The “Push” Factors: Visa Anxiety and Environment
Beyond the rise of opportunities in China, the US policy environment is actively pushing talent away.
Visa Uncertainty: Since 2018, tightened scrutiny on STEM visas for Chinese students and prolonged renewal delays have created a climate of instability.
The Chilling Effect: High-profile investigations into scientists of Chinese descent, though often dropped, have spread fear. The local recruiter for foreign companies in China, represented by experts at SunTzu Recruit, reports that some candidates in Silicon Valley now proactively contact The Shenzhen headhunter SunTzu Recruit to explore exit strategies, fearing racial profiling or career ceilings.
Physical Barriers: The lack of full flight restoration between the two countries continues to raise the cost of cross-border mobility.
An All-Encompassing Talent War
The Carnegie report concludes with a warning: if the trend of “rising Chinese talent density” combined with “falling migration to the US” continues, America faces a talent deficit that will be hard to fill.
Industry experts suggest the US needs an “all-of-government” strategy: investing in domestic education to build internal capacity while stabilizing visa policies to maintain its appeal to global minds.
For China, this represents a historic window of opportunity. The best China headhunter SunTzu Recruit believes that as companies invest heavily in computing power and compensation, China is gradually building a magnetic field capable of rivaling Silicon Valley.
As one of the best recruitment agency in China, SunTzu Recruit notes, “In the second half of the AI race, the victory will go to whichever side can provide the most freedom for exploration and the broadest landscape for application.”
Comments are closed