The global artificial intelligence talent race has transformed into a high-stakes, dual-superpower contest. Research published by Carnegie China evaluating global research contributions reveals that the United States and China are decisively expanding their dominance in human capital. Analyzing 5,823 accepted papers and over 10,000 elite researchers from the 2025 Neural Information Processing Systems (NeurIPS) conference, the findings map human capital trends across the five core layers of technology development defined by Nvidia chief Jensen Huang: energy, chips, infrastructure, models, and applications. While the United States remains a primary destination for foreign researchers, China has established a commanding lead as the originating source of top-tier talent. This shift redefines the geopolitical balance of technology. Understanding the mechanics of this talent race requires examining underlying migration patterns, educational pipelines, and institutional shifts.
- Introduction to the Dual-Superpower AI Talent Race
- Methodology and the Elite AI Talent Cohort
- The NeurIPS Benchmark
- Data Collection Innovations
- China Dominates Undergraduate Origins
- The 57 Percent Share
- The United States as a Global Brain Gain Magnet
- Migration Dynamics and Net Gains
- Rising Retention Rates Within China
- Shifting Career Trajectories
- Global Shifts: Asia Ex-China and European Struggles
- South Korea and Singapore Ascend
- Europe Faces Persistent Headwinds
- India’s Unique Talent Dynamics
- The Intersection of Talent, Capital, and Infrastructure
- Private Versus State-Directed Funding
- Hardware Bottlenecks
- Strategic Implications for Technology Leaders
- Frequently Asked Questions
- Which country produces the most undergraduate AI talent?
- Why do researchers migrate to the United States?
- What is talent retention rate in China?
Introduction to the Dual-Superpower AI Talent Race
Nvidia chief Jensen Huang outlines five core layers governing modern technological supremacy: energy, chips, infrastructure, models, and applications. Each layer requires specialized human capital to function efficiently. The US leads in commercial deployment and brain gain, whereas China commands undergraduate origins, creating a complex geopolitical balance that shifts yearly.
Assessing this race requires deep analytical insight into how nations educate, attract, and retain elite professionals. The intersection of state policy and private enterprise investment dictates who wins the long-term technological war.
Methodology and the Elite AI Talent Cohort
The NeurIPS Benchmark
Quantifying elite research talent requires a standardized, longitudinal benchmark. Analysts look to NeurIPS as the definitive bellwether. The conference captures the highest caliber of machine learning research, encompassing core contributions in model capabilities, computational efficiency, and knowledge transfer.
The latest edition of the Carnegie China talent study analyzed the 2025 NeurIPS cohort, which featured 5,823 accepted papers and 25,677 total authors. Researchers processed complete three-stage career paths, tracking undergraduate origin, graduate education, and current workplace for more than 10,000 authors.
NeurIPS provides a consistent framework to measure talent flows over multi-year periods. Accepted authors represent the elite researchers capable of pushing artificial intelligence past current technological boundaries. This methodology extends historical datasets, creating the most comprehensive longitudinal study on elite AI human capital available today.
Data Collection Innovations
Modern analytical frameworks rely on advanced technology to process massive datasets. Automated tools like Claude Code and OpenAI Codex operate alongside rigorous human oversight to parse academic records accurately.
Establishing a longitudinal dataset allows analysts to measure multi-year talent flows across global institutions with unprecedented precision. This data-driven approach removes traditional biases and reveals concrete trajectories in global engineering migration.
China Dominates Undergraduate Origins
The 57 Percent Share
China has widened its lead as the world’s largest incubator for elite artificial intelligence researchers. In 2025, mainland institutions produced 57 percent of top-tier AI talent, defined as individuals who received their undergraduate degrees in the country. This represents an 11-percentage-point increase from 2022.
Conversely, the US share of undergraduate origin talent fell to 13 percent in 2025. This disparity highlights a structural divergence in foundational education. While North American universities continue to excel in advanced graduate training and commercialization, their domestic undergraduate pipelines supply a fraction of the raw engineering talent produced across East Asia.
| Country or Region | 2022 Share | 2025 Share | Net Change |
|---|---|---|---|
| China | 46% | 57% | +11 percentage points |
| United States | 16% | 13% | -3 percentage points |
| Europe | 15% | 12% | -3 percentage points |
| Other Regions | 23% | 18% | -5 percentage points |
The United States as a Global Brain Gain Magnet
Migration Dynamics and Net Gains
Despite China dominating undergraduate origins, the United States remains the primary global destination for working professionals. The American AI ecosystem functions as an unmatched magnet for international brain power, particularly researchers originating from China. Among artificial intelligence researchers currently employed in the United States, the proportion holding undergraduate degrees from China rose by 4 percentage points in 2025.
When measuring overall migration dynamics, the United States registered a net gain of 2,145 researchers, while China registered a net loss of 1,729 researchers. American laboratories offer unmatched compensation packages, massive computational resources, and close proximity to frontier model developers.
Proximity to industry leaders like OpenAI, Anthropic, and Google DeepMind gives American employers a distinct competitive edge. These advantages ensure that the US continues to attract elite international minds, even amidst tightening regulatory environments and geopolitical friction.
Rising Retention Rates Within China
Shifting Career Trajectories
Although China experiences a net outward flow of human capital, an increasing share of domestic talent is choosing to stay home. In 2022, 57 percent of Chinese-origin researchers who graduated locally ended up working in China. By 2025, that retention figure climbed to 69 percent.
Two primary catalysts drive this elevated retention. China’s commercial artificial intelligence sector offers abundant, high-paying opportunities across application layers, smart manufacturing, and autonomous systems. Simultaneously, tighter immigration controls and strict visa scrutiny in Western nations make relocation increasingly difficult for sensitive STEM fields.
China’s talent retention rate of 69 percent still trails the United States, which retains 89 percent of its domestic talent, but it represents a significant structural strengthening compared to previous decades.
Global Shifts: Asia Ex-China and European Struggles
South Korea and Singapore Ascend
While the spotlight remains fixed on the US and China, other Asian economies are securing prominent positions in the global talent hierarchy. South Korea and Singapore have posted notable gains across research productivity and talent retention.
South Korea boasts the second-highest talent retention rate in the world at 77 percent. Leading institutions like KAIST have climbed twelve places since 2022 to rank among the highest globally for producing top-tier AI researchers. Singapore has experienced a similar ascent, with NUS and NTU securing top twenty status through aggressive national strategies and robust digital infrastructure.
Europe Faces Persistent Headwinds
Europe faces persistent headwinds in the global artificial intelligence talent race. While the continent remains a viable destination for academic research, top European institutions have lost ground in producing and retaining elite talent.
Retention rates hover around 60 percent, edging closer to historical brain-drain levels. Strict regulatory compliance frameworks and lower private venture capital funding hinder institutional competitiveness across the continent.
India’s Unique Talent Dynamics
India maintains a massive workforce of technology professionals and a rapidly expanding startup ecosystem, yet its representation among elite NeurIPS authors reveals a different trend. India’s share of top AI talent dropped by 2 percentage points in 2025, and the proportion of Indian-origin researchers working in the United States declined as well.
No Indian academic institution ranks among the global top twenty for elite paper contributions. This gap highlights a strong focus on enterprise IT and digital public infrastructure over foundational machine learning research.
The Intersection of Talent, Capital, and Infrastructure
Private Versus State-Directed Funding
Private AI investment shows a stark disparity between superpowers. The United States attracted $285.9 billion in private investment compared to China’s $12.4 billion in 2025. This vast gap reflects differing financial ecosystems and venture capital availability.
China counters this deficit with massive state-directed guidance funds and a $138 billion state venture capital fund. These state injections keep domestic research laboratories well-funded despite lower private sector participation.
Hardware Bottlenecks
The US hosts over 5,400 specialized AI data centers, yet the global hardware ecosystem relies heavily on TSMC in Taiwan. Both superpowers face shared structural vulnerabilities in energy grids, high-bandwidth memory chips, and custom silicon clusters.
Securing physical infrastructure remains just as important as acquiring human capital. Without reliable power and advanced chips, even the brightest minds cannot train frontier models.
Strategic Implications for Technology Leaders
Technology leaders must calibrate their human capital strategies against deepening superpower bifurcation. Securing localized talent pipelines is essential for maintaining a competitive advantage in global markets.
Managing hardware and regulatory dependencies will dictate organizational survival over the coming decade. Enterprises must remain adaptable to shifting geopolitical realities.
Frequently Asked Questions
Which country produces the most undergraduate AI talent?
China leads the global landscape, producing 57 percent of top-tier artificial intelligence talent based on undergraduate degree origins in 2025.
Why do researchers migrate to the United States?
The United States attracts elite international researchers through competitive compensation packages, abundant venture capital, massive compute resources, and proximity to frontier labs.
What is talent retention rate in China?
China retains 69 percent of its domestic-origin artificial intelligence researchers, showing steady improvement driven by local commercial expansion.
