Computer Science Enrollment Just Had Its First Drop in Two Decades

For the first time in nearly two decades, fewer American students are majoring in computer science. Undergraduate enrollment in computer and information sciences and support services fell 8.1% in fall 2025, according to the National Student Clearinghouse Research Center. The drop is small by historical standards, large for a pipeline that expanded almost every year since the mid-2000s, and the first sustained reversal of the post-recession CS boom.

The signal reached the public through Stanford economist Jacob Light's new research, first reported by Business Insider. Light's dataset tracks computer science majors specifically and shows the first sustained decline in nearly 20 years. The 8.1% Clearinghouse number covers a broader bucket, computer and information sciences and support services, which includes adjacent IT degrees. The two figures point the same direction but measure different things, and the gap between them is part of the story: students appear to be moving within the IT-adjacent category rather than out of it entirely. The timing is hard to miss. ChatGPT launched in late 2022, generative AI tools reached students by 2023, and the 2025-26 academic year is the first to reflect a post-AI cohort applying to college.

Light himself is careful about what the data does and doesn't show. He frames the enrollment decline as descriptive, not a causal estimate of how much AI exposure changes student demand. The drop is consistent with several overlapping drivers: AI tools reshaping how introductory coursework gets done, a weaker tech labor market weighing on expected returns, and students making substitutions across closely related majors. "Computer science" as a category may be losing share to degrees that didn't exist a decade ago, even as overall interest in tech-literate work holds.

That substitution is already visible. A reallocation analysis of NSC data surfaced by Futurism's writeup of the trend finds that some disappearing CS enrollments are being routed into data analytics and data science, adjacent fields that share a programming core but pitch themselves around applied statistics, business decision-making, and the AI tooling stack. The CS major looks more unbundled than abandoned.

The labor market adds weight, and a caveat. The Layoffs.fyi tracker counts more than 120,000 tech jobs cut in 2026 year-to-date, on top of a 2025 in which employers cited AI in announcements covering more than 54,000 positions, as Business Insider noted. The AI label in layoff notices is a soft signal. "AI-driven restructuring" has become a useful announcement line for cost cuts that would have happened anyway, and reading the layoff count as a clean measure of AI displacement overstates the case.

The second-order effect is on the major itself. Universities that built data science and analytics tracks over the last five years are now absorbing the students the CS major is shedding. The NSF NCSES STEM Talent 2026 indicators still point to long-run demand for technically trained workers across the economy, even as the undergraduate CS label stops being the obvious single bet. The credential is unbundling faster than the curriculum.

For a student weighing a CS degree in fall 2026, the practical picture is more textured than the headline. Programming remains the entry point. What is changing is which wrapper around it, whether AI engineering, data science, applied statistics, or domain-specific analytics, gets a first job fastest. The next enrollment release from the Clearinghouse, due in fall 2026, will show whether the 8.1% dip is a one-year correction or the start of a sustained redistribution.