China Struggles to Attract Talent Despite Easing Visa Rules, While U.S. Builds AI Talent Hubs in Asia on Strength of Research and Compensation

China struggles to attract overseas talent despite relaxed visa rules and research funding
Language barriers and internet controls highlight U.S. advantages in research and compensation
Singapore hubs expand recruitment and talent development pathways

As competition between the United States and China for leadership in artificial intelligence (AI) intensifies, China’s efforts to recruit overseas AI talent through an array of incentives appear to be falling short of expectations. Despite relaxed visa requirements and increased research grants, language barriers, internet controls and conditions for settling in the country continue to deter foreign researchers. The United States, meanwhile, retains an advantage in attracting talent through its advanced research environment, extensive career opportunities and exceptional compensation packages. As U.S. companies extend their recruitment networks to Singapore, Asian researchers are also gaining more opportunities to participate in American companies’ research and development (R&D) locally.

China’s Talent Recruitment Efforts Yield Limited Results

According to research from the Carnegie Endowment for International Peace, a U.S. think tank, on October 7, China is struggling to attract foreign researchers. While U.S. President Donald Trump’s administration raised barriers to foreign talent last year, including increasing the H-1B visa fee for skilled workers to $100,000, China has made recruiting overseas talent in advanced science and technology a critical priority. Alongside substantial research grants designed to attract foreign talent, China introduced the “K visa” last year. The measure aims to expand exchanges with overseas talent by allowing applications without an invitation from a Chinese employer or institution and offering greater flexibility over entries and lengths of stay. These incentives, however, have yet to produce clear results.

There has also been little evidence of a pronounced shift to China among Chinese-origin AI talent working in the United States. The Carnegie Endowment analyzed 10,280 authors of papers accepted at last year’s Conference on Neural Information Processing Systems (NeurIPS) and found that, for every 30 researchers with undergraduate degrees from China working in the United States, just one researcher with an undergraduate degree from the United States was working in China. Although the ratio narrowed from 46 to one in the 2022 sample, the United States retained its advantage in attracting talent. The share of researchers working in the United States who had completed their undergraduate studies in China also rose by four percentage points over the same period. Despite U.S.-China tensions and tighter visa restrictions, AI talent educated in China continues to account for a substantial share of researchers working in the United States.

As China Expands Publications and Patents, the U.S. Retains an Edge in Frontier Model Development

A major factor in America’s ability to attract AI talent is an environment that allows researchers to participate in frontier research while competing with and learning from world-class peers. According to the “AI Index 2026” published by Stanford University’s Institute for Human-Centered Artificial Intelligence (HAI), China surpassed the United States in the number of AI publications, citations and granted patents, but the United States retained its lead in notable AI model development and patent impact. The United States developed 59 notable AI models last year, roughly 1.7 times China’s 35. Beyond its research base for generating publications and patents, the United States thus demonstrated its strength in translating research into working models. In patents, too, China led in the volume of grants, while the United States showed an advantage in securing high-impact technologies. These achievements are likely to encourage researchers seeking experience in frontier model development and opportunities to work on core technologies to choose the United States.

Underpinning this research activity are extensive computing infrastructure and private capital. According to a study published by the Carnegie Endowment in June this year, approximately 75% of the world’s advanced AI computing clusters were concentrated in the United States last year. Stanford also put U.S. private investment in AI last year at $285.9 billion. With training and validating large models requiring enormous expenditure, abundant funding and equipment are considered essential foundations for repeated experimentation and long-term research. The United States also has well-developed research and startup networks connecting universities, companies and venture capital firms. University research feeds into joint development with industry, while research teams with proven technical capabilities raise investment to launch startups. Through this process, researchers move between academia and industry, building research experience and exploring commercialization opportunities.

Table 1. Examples of Major U.S. AI Companies Recruiting Talent from China

Research Opportunities and Exceptional Compensation Draw Talent to the U.S. AI Industry

With its established research and startup ecosystem, the United States offers AI talent abundant employment options and opportunities for professional growth. Researchers can also gain recognition for development experience acquired at startups and move into core projects at major technology companies. Shengjia Zhao, appointed chief scientist of Meta Superintelligence Labs in July last year, is a prominent example. After completing his undergraduate studies at Tsinghua University in China and earning a PhD from Stanford University in the United States, he contributed to the development of ChatGPT and GPT-4 at OpenAI. Upon joining Meta, he took on a role working directly with CEO Mark Zuckerberg and Chief AI Officer Alexandr Wang to define the new lab’s research priorities and direction. His experience developing major models brought him into a position leading next-generation AI research at a major technology company. Meta also recruited Hongyu Ren, Jiahui Yu and Shuchao Bi, researchers from China who were working at OpenAI at the time, to join its superintelligence development organization.

Alongside abundant research opportunities, high salaries and exceptional compensation packages offered to entice recruits are bolstering the U.S. AI industry’s ability to attract talent. Ruoming Pang, who completed his undergraduate studies at Shanghai Jiao Tong University, spent 15 years at Google before moving to Apple in 2021, where he led its AI foundation model development team. When Meta recruited him in July last year, its offer reportedly exceeded $200 million in salary, stock and other compensation payable over several years. The substantial package was designed to secure his experience managing a large research organization and his model development expertise. Rival companies continued pursuing Pang, and OpenAI hired him in February this year. Coming approximately seven months after his move from Apple to Meta, the appointment illustrates how a researcher who completed his undergraduate education in China became the target of successive recruitment efforts by leading U.S. AI companies.

Language and Settlement Barriers Hold China Back

The willingness to spend heavily to secure key researchers reflects the view that a small number of individuals can determine AI model performance and the pace of development. The Carnegie Endowment likewise characterized U.S.-China competition in frontier models as, in effect, a contest to secure the very best researchers. Industry figures involved in global AI recruitment estimate that the pool of top-tier talent ranges from several dozen to around 1,000 people. With companies simultaneously pursuing a limited pool of researchers, even one individual’s move can affect a rival’s development capabilities. Ariel Herbert-Voss, CEO of RunSybil and a former OpenAI researcher, explained that AI companies are seeking combinations of talent with complementary expertise to accelerate development. Alongside an individual researcher’s ability, whether that person can fill gaps in an existing team’s capabilities is a key consideration in recruitment.

For China to attract overseas talent to its research institutions and companies, it must address constraints arising from language barriers and the conditions for settling in the country. Many Chinese technology companies use Chinese as their working language, limiting employment options for researchers who do not speak it. The “Great Firewall,” which restricts access to foreign internet services, is also cited as a deterrent for foreign technology professionals considering a move to China. Losing access to information search and communication tools used abroad can make it harder to obtain research materials and collaborate with existing colleagues. Institutions supporting long-term settlement are another major consideration. When the K visa was introduced last year, critics pointed to a lack of detailed guidance on permanent residency and accompanying family members, while foreign nationals were understood to be permitted to acquire Chinese citizenship only in highly exceptional cases.

U.S. Tech Giants Seek Talent in Asia

While China’s efforts to recruit overseas talent face language and institutional barriers, U.S. companies are extending their recruitment networks to Singapore and accelerating local hiring and research collaboration in Asia. In May this year, OpenAI signed an agreement with the Singapore government and announced plans to establish its first applied AI lab outside the United States. The company plans to invest more than $234.4 million in Singapore’s AI ecosystem and build a workforce of more than 200 forward-deployed engineers and technical specialists over the coming years. These employees will work alongside customers to build and deploy AI systems in finance, healthcare and public services. OpenAI also plans to introduce a training program to develop experienced software engineers into AI deployment specialists. Denise Dresser, OpenAI’s chief revenue officer, cited Singapore’s strong technical talent and proactive environment for AI adoption as reasons for choosing the country.

Companies are also actively expanding their engagement with next-generation talent through joint research with local universities. Microsoft opened its first Southeast Asian research lab in Singapore in July last year and launched nine new research projects with the National University of Singapore and Nanyang Technological University during its first year, covering areas including healthcare AI, robotics and multi-agent systems. University and corporate researchers jointly supervise PhD students, while internships and visiting researcher programs provide opportunities to participate in research. Qiming Huang, a National University of Singapore student participating in the joint doctoral training program, had his research on robotic video analysis selected for an oral presentation at this year’s Conference on Computer Vision and Pattern Recognition (CVPR). These initiatives have created pathways for local talent to collaborate with researchers at U.S. companies and build a record of achievement while still pursuing their degrees.