US Computer Science Enrollment Fell 12.4 Percent This Fall, the Steepest Drop of Any Field

Computer and information science lost more students this fall than any other field in American higher education, according to preliminary fall 2026 enrollment data published October 29 by the National Student Clearinghouse Research Center. Undergraduate enrollment in the field fell 12.4 percent at four-year institutions and 12.3 percent at two-year institutions, and graduate enrollment fell 10.4 percent, all year over year. Undergraduate enrollment across all fields rose 0.6 percent in the same period.

The decline compounds. Last fall's preliminary report already showed a 7.7 percent drop among four-year undergraduates in the field, the first reversal after years of growth, and a 5.9 percent dip at two-year colleges. Across two years, four-year undergraduate computer science enrollment is down 18.4 percent, two-year enrollment is down 17.5 percent, and graduate enrollment is down 27.1 percent. "Those are just remarkable declines for a field that until recently had been one of the fastest growing areas that we've seen in higher education," Matthew Holsapple, the Clearinghouse's senior director of research, said at a press conference.

The report supplies the numbers, not the cause. The attribution to artificial intelligence comes from the people reading it. George R. Boggs, president and chief executive emeritus of the American Association of Community Colleges, told The EDU Ledger that many formerly high-paying jobs, coding among them, can now be done more efficiently by AI, and that health careers look like safer bets. The Bureau of Labor Statistics said in August that computer and mathematics occupations would feel the largest impact from AI. The Federal Reserve Bank of New York has reported that unemployment among recent computer science and computer engineering graduates reached 7.0 percent and 7.8 percent in 2024, roughly three percentage points above the overall rate.

What makes the drop conspicuous is how little else moved. Undergraduate enrollment grew 0.6 percent, a fourth straight year of growth but far below the 2.4 percent gain in fall 2025. Graduate enrollment fell 1.6 percent, dragged by a 3.7 percent decline in master's programs, which many universities rely on as a revenue source; master's enrollment slipped only 0.4 percent last fall. Health professions gained 6.2 percent at four-year colleges, 4.3 percent at two-year colleges and 2.0 percent in graduate programs. Engineering rose more than 6 percent at the undergraduate level. Visual and performing arts and the liberal sciences each fell more than 4 percent.

Community colleges were the only bright spot in the headline numbers, adding 1.2 percent in undergraduate enrollment on top of a 3.5 percent increase last fall. "Students are identifying pathways that are right for them beyond the traditional pathway," Holsapple said. "And I think affordability is almost certainly one important piece of that." That reading cuts against a simple AI story: the field losing the most students is the one that historically promised the highest starting salaries, while the segment that grew is the cheapest to attend.

The preliminary data covers responses from 46 percent of institutions that report to the Clearinghouse, representing 7.9 million enrollments, and excludes international students, whose numbers will appear in the final January report. Other indicators point the same direction: new foreign student enrollment fell 17 percent in fall 2025 according to JB International and NAFSA, and international applications dropped about 10 percent this year, the largest decline in at least a decade, per the Common App.

The consequence runs in two directions. Four years from now, the graduating cohort in computer science will be roughly a fifth smaller than the one that entered before the reversal, at exactly the moment companies say they cannot hire enough engineers who can supervise AI systems rather than write boilerplate. The counter-argument is that entry-level coding was the part of the job AI absorbed first, so a smaller pipeline may simply be pricing that in. Neither side has data yet. What the Clearinghouse has produced is the first hard number showing that students, at least, have already decided.