Mapping the Institutional Pipeline for Global AI Talent

Immigrants, particularly those from Asian Low- and Middle-Income Countries (LMICs), play a central role in U.S. frontier AI firms, yet the pathways that bring them into these positions are not well understood. We examine how U.S. universities serve as institutional intermediaries linking global talent to domestic firms. Using data on 1,757 AI graduates from top U.S. PhD programs alongside surveys and interviews, we develop a three-stage process model tracing how immigrant students move into, through, and beyond doctoral programs. We find that, compared to native students, immigrant students are more likely to match with diaspora faculty at entry, that advisor industry ties are more strongly associated with immigrants’ participation in industry internships, and that internships are a common pathway through which immigrant graduates, particularly those from Asian LMICs, enter top AI firms. Our findings reframe skilled immigration into elite firms as a product of institutional intermediation, and reveal a pipeline whose functioning depends on specific relational structures that are vulnerable to disruption.