Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →In Carnegie Endowment for International Peace’s 2026 analysis of NeurIPS authors, the 2025 cohort showed 30 researchers of China origin working in the United States for every U.S.-origin researcher working in China. That is a measure of one specific bilateral flow in a sample of elite machine-learning researchers—not a count of all foreign AI talent in China. Carnegie also found that China had become the sample’s largest source of researchers and its leading workplace location.
What the 30:1 ratio measures—and what it does not
The ratio compares China-origin researchers working in the United States with U.S.-origin researchers working in China. Carnegie defines a researcher’s “origin” by the country where they completed undergraduate education. It does not mean that China has only one foreign researcher of any origin for every 30 Chinese researchers in the United States, nor does it measure all people who moved between the two countries.
As an Amazon Associate I earn from qualifying purchases.
The findings come from accepted-paper authors at NeurIPS, a major machine-learning conference. Carnegie’s 2026 analysis reports that the bilateral ratio was 46:1 in 2022 and 30:1 in 2025. The U.S. advantage in this particular flow measure therefore narrowed, but remained substantial. Carnegie’s analysis was published September 23, 2026.
Where the 2025 sample came from and where it worked
Origin and current workplace tell different stories. China was the most common origin in the sample, while the United States remained a major destination for researchers of Chinese origin. At the same time, China had become the most common current workplace in the 2025 results.
#1 Best Overall
| Measure | 2025 result |
|---|---|
| Origin by undergraduate education | 57% China-origin; 13% U.S.-origin |
| Current workplace location | 41% working in China; 34% working in the United States |
| Net balance in Carnegie’s flow measure | +2,145 researchers for the United States; -1,729 for China |
These are results for Carnegie’s defined conference-author sample, not whole-workforce totals. The net figures describe the study’s flow measure and should not be read as national counts of AI workers. The apparent contrast—China leading as a workplace while the United States has the stronger bilateral flow balance—comes from measuring different things: the location of the full sample versus the balance between researchers of U.S. and China origin working in the other country.
Attraction, retention, and workplace are separate questions
“Attracting foreign talent” asks how many researchers from other countries work in a country. Retention asks whether researchers originating there remain there. Workplace share asks where the sampled researchers are employed now. The 30:1 figure addresses a narrow origin-to-workplace comparison between the United States and China; by itself, it cannot answer all three questions.
For China-origin researchers, Carnegie reports that 69% worked in China in 2025, up from 57% in 2022. It reports U.S. retention at 89%. These retention percentages are not the same statistic as the 30:1 bilateral flow ratio: they describe where origin groups worked, rather than the count of one origin group in the other country relative to the reverse group.
Why China may be retaining more researchers
The 2026 study’s authors suggest that China’s growing AI industry may have made suitable jobs easier to find at home, while tighter U.S. visa restrictions—particularly affecting Chinese STEM graduate students—may have influenced decisions to stay or return. These are plausible explanations offered by the authors, not causal findings from the cohort analysis; the study does not isolate the effect of visas or job availability on individual careers.
A separate Carnegie analysis published in 2025 describes visa-processing delays, geopolitical suspicion and tensions, pandemic-era travel restrictions, and expanding Chinese research and industry opportunities as contextual factors that could shape career choices. Those factors likewise do not establish why any particular researcher moved. That analysis followed a different group: 100 China-origin researchers employed at U.S. institutions in 2019. Six years later, 87 remained at U.S. institutions, ten were at Chinese institutions, and three were elsewhere. It is evidence about retention within an already U.S.-based cohort, not a measure of the overall direction of new flows.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How broad—and how reliable—is the evidence?
Carnegie counted 5,823 accepted papers and 25,677 unique authors for the 2025 NeurIPS conference. Complete undergraduate, graduate, and employment histories were available for 10,280 authors, or 40% of the total; this subgroup was used for the final career-flow and paired comparisons. Current workplace was measured one year after each conference: 2026 for the 2025 cohort and 2023 for the 2022 cohort.
- It is not a census: NeurIPS authors are a proxy for elite machine-learning research talent, not all AI researchers, engineers, or workers.
- The complete-history group may differ from the full author pool: Carnegie notes slight overrepresentation of academia and underrepresentation of industry.
- The main country-level pattern is more secure than each profile: Carnegie says broader sample checks preserved the main U.S.–China findings, but manual verification was selective and no record-level accuracy rate is reported. Aggregate findings, including China’s position as the leading workplace location in the 2025 results, are more dependable than an individual’s inferred profile or physical work location.
The Carnegie authors summarize their interpretation this way: “The United States remains a magnet for global AI talent, particularly Chinese talent.” That observation can coexist with China’s growing retention and its leading workplace share in this sample. None of the measures alone establishes which country has greater overall AI capability.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
