Marc Andreessen's claim that international enrollment at top US universities rose from "2 or 3 or 4 percent" to "70%" is a significant exaggeration that conflates different enrollment levels, cherry-picks extreme cases, and misrepresents the scale of change. The real numbers tell a more nuanced story: undergraduate international enrollment at the eight elite universities studied ranges from 8% to 20% today, up from roughly 2–5% in the 1970s — a meaningful increase, but nowhere near 50–70%. The high percentages Andreessen invokes exist only in graduate programs, particularly STEM fields, where international students have constituted a substantial share for decades. His framing collapses undergraduate admissions, total university enrollment, and niche graduate programs into a single misleading statistic.
Andreessen made the claim at the Reagan National Economic Forum in July 2025, stating that "foreign enrollment rates of the top universities went from, you know, like, two or three or four percent 50 years ago to whatever, 27 or 30 or 50." When co-panelist Joe Lonsdale added "Columbia's over half, right?" Andreessen responded, "70% or whatever it is." This report examines institutional data for eight elite universities to assess these claims.
Current international enrollment data, drawn from official institutional fact books, Common Data Set filings, and IPEDS reports, reveals a consistent pattern across all eight universities: moderate international representation among undergraduates and substantially higher percentages among graduate students.
| University | Undergrad Int'l % | Graduate Int'l % | Total Int'l % | Year |
|---|---|---|---|---|
| Harvard | 14% | ~32% | 27% | 2023–24 |
| MIT | 12% | 40% | 29% | 2024–25 |
| Stanford | 12% | 36% | 26% | 2024–25 |
| Yale | 11% | ~34% | 24% | 2024–25 |
| Princeton | 13% | 45% | 25% | 2024–25 |
| Columbia | 20% | 46% | 39% | 2023–24 |
| U. of Chicago | 17% | 41% | 31% | 2023–24 |
| UCLA | 8% | ~26% | 13% | 2023–24 |
The total university percentages range from 13% (UCLA) to 39% (Columbia). Harvard's 27% matches Andreessen's low-end figure exactly, while Columbia at 39% is the highest among these eight schools. No institution approaches 50%, let alone 70%. Columbia's figure rises to approximately 55% only when including students on Optional Practical Training (OPT) — graduates working off-campus who are technically still in the visa system but not enrolled in classes. This OPT-inclusive number, drawn from DHS/SEVIS data analyzed by the Center for Immigration Studies, is the likely origin of the "over half" claim, but it fundamentally misrepresents campus enrollment.
Graduate programs drive the overall percentages upward dramatically. At Princeton, 45% of graduate students are international compared to just 13% of undergraduates. At MIT, the split is 40% versus 12%. Certain professional schools push even higher: Harvard's Kennedy School is 57% international, its Graduate School of Design 54%, and Yale's School of Engineering 65%. These program-specific figures in the 50–65% range are likely what gets laundered into claims about entire universities.
Andreessen's assertion that international enrollment was "2 or 3 or 4 percent" fifty years ago is approximately correct for undergraduate programs at most of these schools, but substantially understates the reality for universities as a whole. While specific 1970s institutional data is scarce — IPEDS only began in the 1980s, and most universities' online fact books don't extend back that far — multiple converging sources paint a clear picture.
Nationally, international students comprised roughly 1.7% of all US higher education enrollment in 1970–71, rising to 2.5% by 1980–81 (per IIE Open Doors data). However, elite research universities were already well above this average. The critical evidence comes from STEM doctoral programs: NSF data shows that 23% of science and engineering doctoral recipients were foreign-born by 1966, rising to 27% by 1973. This means that even in the mid-1970s, graduate programs at research-intensive schools like MIT and Stanford likely had international enrollment of 15–30% — far above the "2 or 3 percent" Andreessen claims.
Yale provides the best-documented historical undergraduate baseline. According to the Yale Daily News, citing institutional records, "foreign share of the undergraduate student population was only 3 percent in the late 1980s, about the same level as at the close of the 19th century." Working backward, Yale College likely had 2–3% international undergraduates in the mid-1970s. A Yale alum from that era described foreign students as "rare enough to be more of an oddity than an influence." Similar patterns likely held at Harvard College, Princeton, and Stanford's undergraduate programs.
MIT offers the most complete 1990s data point that helps triangulate: in 1994–95, MIT had 2,151 international students — 22% of the total student body — with 364 undergraduates (about 8%) and 1,787 graduate students (about 33%). Given that international enrollment grew steadily through the 1980s and 1990s, MIT's total in the mid-1970s was likely 15–20% — already far above Andreessen's claimed baseline. A university that was already one-fifth international in 1994 was certainly not at 2–4% twenty years earlier.
The mid-1990s provide a useful midpoint that shows the trajectory of change. For the schools where reliable 1990s data exists:
| University | ~1994–95 Total Int'l % | 2023–24 Total Int'l % | Change |
|---|---|---|---|
| MIT | 22% (confirmed) | 29% | +7 pts |
| Harvard | ~18% (estimated from 3,410 int'l students per Open Doors) | 27% | +9 pts |
| Yale | ~11–14% (estimated) | 24% | +10–13 pts |
| UCLA | ~5% (1,668 non-immigrant students per SOCCIS) | 13% | +8 pts |
The Bound et al. (2021) study in the Journal of Economic Perspectives documents that undergraduate international enrollment at selective institutions barely grew before 2000. The surge in international undergraduate enrollment is primarily a 21st-century phenomenon, driven largely by the explosive growth in Chinese family incomes. Bound et al. estimate that the fraction of Chinese families able to afford US out-of-state tuition grew from 0.005% in 2000 to over 2% by 2013. The growth in master's-level and undergraduate enrollment from China and India accounts for the bulk of the increase at elite universities since 2005.
Columbia's trajectory illustrates this pattern clearly: international students rose from 25% in 2011 to 31.5% in 2015 to 39% in 2024. The Stand Columbia Society documented that between 2015 and 2024, Columbia's domestic enrollment grew just 4.7% while international enrollment grew 47.1%. Columbia is an outlier even among elite peers, partly because its massive graduate and professional school enrollment (26,168 graduate students versus 9,111 undergraduates) amplifies the graduate-driven international share.
Andreessen and Lonsdale's exchange — "Columbia's over half, right?" "70% or whatever it is" — is the most verifiable and most clearly false element of the claim. Columbia's own Office of Planning and Institutional Research reports 39% international enrollment. Columbia's ISSO (International Students and Scholars Office) counted 13,745 international students in Fall 2024 out of approximately 35,769 total students.
The 55% figure that circulates in immigration-restrictionist media comes from the Center for Immigration Studies, which used DHS/SEVIS data reporting 20,321 total foreign students at Columbia — but this count includes 6,483 students on Optional Practical Training who have graduated and are working, not attending classes. Including OPT participants as "enrolled students" is methodologically questionable and inflates the number by nearly 50%. The 70% figure has no documented basis in any dataset or methodology applied to Columbia. Northeastern University (~67% per DHS data including OPT) may be the actual institution behind this number, confused with Columbia.
For context, five of Columbia's schools do have majority-international student bodies at the graduate level: Engineering, Professional Studies, GSAS, SIPA, and GSAPP. A visitor to specific graduate classrooms at Columbia might indeed observe a majority-international environment. But extrapolating from individual graduate programs to "Columbia is 70% foreign" is a fundamental error of composition.
The most important analytical failure in Andreessen's framing is collapsing undergraduate and graduate enrollment into a single number, then presenting the combined figure as if it described the undergraduate admissions pipeline that most people picture when they hear "top universities."
When Americans debate elite university enrollment, they typically envision the undergraduate experience — Harvard College, Yale College, Princeton's undergraduate program. At this level, international students represent 8–20% of enrollment across the eight schools studied, with most clustered around 11–14%. This is a meaningful increase from the estimated 2–5% of the 1970s, roughly a tripling or quadrupling over fifty years. But it is categorically different from claiming these schools are "27 or 30 or 50 or 70 percent" foreign.
Graduate programs tell a different story, but one with important historical context. STEM doctoral programs were already substantially international decades ago. By 2003, 51% of all US science and engineering PhD recipients were from outside the US — 67% in engineering and 68% in economics. By the mid-2010s, 81% of full-time graduate students in electrical and petroleum engineering and 79% in computer science were international. These are striking numbers, but they represent a long-standing structural feature of American graduate education, not a recent transformation from "2–3%" to majority-foreign.
The growth in international graduate enrollment has been concentrated in master's programs (which are often revenue-generating for universities) and in specific STEM fields. Graduate computer science saw international enrollment increase 480% from 1995 to 2015 (from 7,883 to 45,790), while domestic enrollment grew only 45%. This reflects both enormous demand from abroad and universities' financial incentives to enroll full-tuition-paying international students.
Andreessen's claim operates through three layers of distortion. First, his 1970s baseline of "2–3%" is roughly accurate only for undergraduate programs and for the national average — but elite research universities as a whole were already 10–20% international by the mid-1970s, with doctoral programs far higher. Second, his current figures of "27–50%" are accurate only for total university enrollment (including massive graduate programs) at select schools, not for the undergraduate experience most listeners would envision. Third, the "70%" claim for Columbia is flatly unsupported by any standard enrollment measure.
The real story is significant but less dramatic: undergraduate international enrollment at elite universities roughly tripled or quadrupled over fifty years, from about 2–5% to 8–20%. Total university enrollment, heavily influenced by graduate programs, roughly doubled at most schools. And STEM graduate programs, which were already 20–30% international in the 1970s, have reached 40–80% in many fields. These are meaningful shifts worth debating, but the debate is poorly served by conflating them into a single alarmist figure suggesting that top universities went from 3% to 70% foreign. The actual data shows a more gradual, structurally differentiated trend that cannot be responsibly compressed into a single soundbite.