Computer-science enrollment is falling across U.S. colleges after a decade of rapid expansion, forcing academic leaders to reassess faculty plans, specialized facilities and the way they describe technology careers to students. New reporting published Tuesday found the reversal spreading from national datasets to individual campuses that built aggressively for continued growth.

The shift is substantial, but it is not evidence that computing work is disappearing. Instead, the data point to a mismatch between students’ near-term perceptions of an unsettled entry-level labor market and longer-term demand for software, artificial intelligence, cybersecurity and data skills.

A National Reversal After Years of Growth

The National Student Clearinghouse reported that undergraduate enrollment in computer and information sciences declined in every major institutional sector in spring 2026 even as overall undergraduate enrollment rose 1.3%. Its sector data showed declines of 8.4% at four-year institutions, 9.3% at primarily associate-degree baccalaureate institutions and 11.2% at two-year colleges.

A separate survey of computing departments reached the same conclusion. The Computing Research Association’s Taulbee report found that new bachelor’s enrollment fell 12.9% in 2025 within its longitudinal cohort, while new master’s and doctoral enrollment dropped 10.3% and 15%, respectively. Total undergraduate computer-science enrollment in that cohort declined 4.1%.

Those measures cover different samples and time periods, so their percentages are not interchangeable. Their direction, however, is consistent: fewer students are entering or remaining in traditional computer-science pipelines. The Clearinghouse also found that 85% of bachelor’s-seeking computer-science students who began in fall 2024 persisted into a second fall, the lowest rate among the 10 largest major fields.

Campus Numbers Show the Planning Problem

Institution-level records illustrate why the national change matters for budgets. Arizona State University’s official census shows enrollment in its computer-science bachelor of science program falling from 5,844 in fall 2024 to 5,008 in fall 2025, a 14.3% decline. Its computer-science master’s program fell from 1,897 to 1,463, while several adjacent fields—including electrical engineering and a data-science concentration—grew.

At the University of Tennessee, Knoxville, published program data show undergraduate computer-science enrollment declining from 672 in 2024 to 555 in 2025 after nearly a decade of uninterrupted growth. Those two universities do not represent every campus, but their primary records demonstrate how quickly program-level demand can turn even within institutions whose broader engineering enrollment remains strong.

For colleges, the operational risk is delayed adjustment. Faculty hiring, laboratory space, advising capacity and new buildings are planned years ahead. Enrollment can change much faster. Cutting too deeply after one downturn could leave departments short of expertise if demand rebounds; assuming the previous trajectory will resume could leave expensive capacity underused.

AI Is Reshaping Demand, Not Ending It

The enrollment decline is occurring alongside rapid growth in specialized offerings. A 2026 mapping study identified more than 350 undergraduate artificial-intelligence majors, minors, concentrations and certificates across 560 U.S. institutions. That expansion suggests some students may be moving toward more explicitly named AI, data-science, robotics or cybersecurity programs rather than leaving computing altogether.

Labor projections also complicate the idea that a traditional computer-science degree has lost its value. The Bureau of Labor Statistics expects employment for software developers, quality-assurance analysts and testers to grow 10% from 2025 to 2035—faster than the 3% projected across all occupations—with about 106,100 openings a year. The agency’s outlook specifically connects demand to AI, robotics, automation, connected devices and cybersecurity.

The more defensible conclusion is that the occupational signal has become harder for students to read. Highly visible technology layoffs and automation claims can shape major selection immediately, while federal employment projections describe a decade-long market that includes replacement hiring and new specialties.

What Academic Leaders Can Measure Now

A single enrollment total cannot reveal whether a department faces a temporary cycle, a shift into adjacent majors or a deeper loss of confidence. Leaders can separate those possibilities by tracking introductory-course demand, major declarations, transfers among related programs, persistence after the first year, internship placement and graduate outcomes.

Curriculum changes also require discipline. Adding “AI” to a program title may attract attention, but the mapping study found wide variation in requirements across AI degrees. Institutions will need to distinguish durable foundations—algorithms, data structures, statistics, systems and ethics—from tools that may change during a student’s time in college.

The Educator's Takeaway

The enrollment retreat is best treated as a planning signal, not a verdict on computer science. National and campus-level data agree that traditional programs are cooling, yet federal projections still show above-average employment growth and substantial annual openings. For faculty and academic leaders, the near-term task is to determine whether students are leaving computing or redistributing into AI, data science, engineering and cybersecurity. That distinction affects hiring, course scheduling, advising and capital commitments. Clearer communication will also matter: students need evidence about both the volatility of entry-level technology hiring and the continuing value of strong computational foundations. Programs that monitor movement across related majors and publish transparent outcomes will be better positioned to adjust without overreacting to one enrollment cycle.