What happened
Stanford HAI and the Hoover Institution released a first-of-its-kind assessment measuring America's STEM talent pipeline end-to-end — from K-12 through graduate training, employment, and immigration — and warns that 'its advantages over China are narrowing rapidly.' The report aggregates decades of data from federal surveys, visa filings, publication and patent records, and company histories covering hundreds of thousands of students and workers. Key findings: America ranks 34th in the world in mathematics despite per-pupil spending 48% above the OECD average; international students earn 51.3% of STEM master's degrees and 38.7% of research doctorates (64.9% of CS master's degrees); the US share of top-cited emerging-technology publications fell from 14.2% to 8.4% over ten years while China's rose from 11.5% to 20.8%; and 44.2% of Forbes AI 50 companies had at least one immigrant founder, with every firm worth $100B+ having one. It warns that weak domestic preparation, narrowing research advantages, and immigration bottlenecks (per-country caps from 1965) create a compounding risk where 'none can safely be treated as permanent.'
Why it matters
Executives and policy leads deciding where to build AI teams should read this as the clearest current evidence on the US talent supply, its dependence on international graduates, and the immigration-policy exposure that underpins AI R&D and engineering headcount.
Action needed
Reference the report's data when planning AI talent strategy and immigration-dependent hiring; brief HR and government-affairs leads on the international-student concentration and research-share trends.