Computer science is no longer the automatic safe bet it looked like after ChatGPT launched. The first real enrollment drop is showing up at the same moment junior software work is getting harder to find.
The boom was supposed to last. For years, computer science looked like the major you chose if you wanted the cleanest route from a campus lab to a high-paying job. Then students started looking at the entry-level market. They noticed something ugly. The first rung is missing.
Business Insider reported on July 22, 2026 that research by Stanford economist Jacob Light found computer science course enrollment fell 4.6% in the 2025-26 academic year across 1,019 institutions. That's the first national decline in nearly two decades by his measure. It isn't a collapse. It's still a break in the story colleges and parents have been telling students since the early 2000s.
The University of California numbers point the same way. TechSpot, citing UC system data and earlier reporting by the San Francisco Chronicle, reported that 12,652 UC undergraduates were majoring in computer science in the latest count, down 6% from 2024 and 9% over two years. The National Student Clearinghouse found undergraduate enrollment in computer and information science at four-year colleges fell 8.1% in fall 2025, from about 659,700 students to 606,100. Graduate enrollment dropped 14%.
That's not students walking away from technology. It's students walking away from the old promise that learning to code, by itself, gets you hired.
The entry-level door is narrowing
The job data gives them a reason. Erik Brynjolfsson, Bharat Chandar and Ruyu Chen at Stanford's Digital Economy Lab studied ADP payroll data and found that early-career workers ages 22 to 25 in the most AI-exposed occupations had a 16% relative employment decline since generative AI took off, even after controlling for firm-level shocks. Software developers are one of the clearest cases. Fortune reported in April that employment among developers ages 22 to 25 had fallen nearly 20% from its late-2022 peak.
That's the pressure point. Older developers aren't seeing the same damage. Stanford's Canaries Dashboard, updated July 1, 2026, shows the pattern weakens and eventually disappears as you move into older age groups. The reason is not mysterious. AI is strongest at the written, codified part of work: documentation, examples, boilerplate, tutorials, basic debugging. That overlaps with what a new CS graduate has just spent four years learning.
Experience still counts. Senior engineers carry product judgment, old outages, architectural scars, codebase history and the ability to know when a plausible answer is wrong. You don't get that from a textbook. You get it by doing the job.
Hiring managers know this, even if they don't always say it cleanly. SignalFire's 2025 State of Tech Talent report found that new graduates accounted for just 7% of Big Tech hires, down from 15% before the pandemic. Big Tech cut new-grad hiring 25% in 2024 compared with 2023. Startups cut it 11%. When a seed-stage company can put one strong engineer beside Cursor, Copilot or Claude Code, the old argument for hiring three juniors gets harder to make.
Frankly, that logic is dangerous. It works for this quarter. It does not build the next staff engineer.
UC San Diego shows where some students are going instead. TechSpot reported that it was the bright spot in the UC data after launching a dedicated undergraduate AI major, with enrollment rising roughly 20%. The signal is plain enough: students still want the technology career. They want the version that looks less likely to be automated before they graduate.
The shortage comes later
If you're a founder, the near-term answer looks obvious. Hire fewer engineers. Hire more senior ones. Pay for output, not headcount. For the next two or three years, that may even be right.
The harder part arrives after that. The mid-level developers startups rely on in 2028 and 2029 are supposed to come from the junior developers companies hire in 2026. If those jobs disappear now, the pipeline doesn't pause politely and restart when the market wants it. It thins out. Then it gets expensive.
The Computing Research Association's 2025 Taulbee Survey adds another warning sign. New CS bachelor's enrollments fell 12.9% in its longitudinal cohort, while master's enrollments fell 10.3% and doctoral enrollments fell 15%. Separate CRA pulse survey coverage found 62% of computing programs reported declining undergraduate enrollment, while only 13% reported growth. This is not just one California system cooling off after a hot decade.
Students are making a rational calculation from messy evidence. They see layoffs, fewer junior roles, AI coding tools improving quickly and employers asking entry-level candidates to show mid-level judgment. Some will choose AI programs. Some will choose electrical engineering, data science or cybersecurity. Some will leave computing entirely.
The irony is sharp. The technology that made coding look easier may make trained developers scarcer. If companies stop hiring beginners because AI can do beginner tasks, they shouldn't be shocked when there are fewer experienced people later who know which AI-written code is worth shipping.
The CS rebound may come. It probably will. But the next boom won't look like the last one, and students can already see that from the application form.
Also read: CuspAI raises $450 million to let AI design the next generation of chip materials • Synthesia turns its AI avatars into performance coaches with Roleplay Sessions • Omen AI raises $31M to detect the bacterial threat quietly killing GPU clusters
