The number of computer science graduates at UC Berkeley is expected to decrease to 851 for the 2025-26 academic year, down from 1,029 graduates in 2024-25. According to Electrical Engineering and Computer Sciences Chair Jelani Nelson, as of late March, the CS department is slated to graduate approximately 350 students in 2027. These figures represent a 59% decrease in CS enrollment from the 2025-26 to 2026-27 school years.
The decline in campus computer science graduates mirrors a trend across the UC system, with CS major enrollment across the university decreasing in 2025 for the first time since the early 2000s. It also contributes to a larger nationwide decline in CS majors, with an 8.1% drop at four-year colleges in fall 2025.
Over the past 10 years, the number of campus computer science graduates has increased dramatically from 422 in 2015-16, despite slightly lower numbers in the 2020-21 and 2023-24 classes. The current expected number of computer science major graduates for 2026-27 will increase due to some juniors changing their majors in the new comprehensive review process, but “not by much,” Nelson said in an email.
Since the release of ChatGPT in 2022, employment for computer science and math majors aged 22 to 27 has fallen by 8%. However, campus spokesperson Janet Gilmore noted that student interest in CS-related majors is “still strong” despite the rise of AI. Gilmore cited rising instructional costs, campus budget constraints and faculty availability as contributing factors in the reduction in enrollment.
“Lest there be any confusion, the collapse of the CS major is not due to a lack of student demand,” Nelson said in an X post.
In an X post, Nelson identified the high cost of instruction as the primary cause of campus’s decision to reduce CS major enrollment. Undergraduate teaching assistants now cost the department between $71.95 and $80.51 per hour. Since winning a grievance in January 2020, campus EECS and data science undergraduate TAs receive proportional tuition waivers depending on how many hours they work. According to Nelson’s post, this change significantly increased department costs, which led campus to reduce undergraduate CS enrollment and decrease the number of undergraduate TAs.
In the email, Nelson also called on campus leadership to “bargain harder” on TA compensation, noting that it is currently “multiples higher” than the market average. He added that more than 80% of undergraduate TAs do not qualify for need-based financial aid but are still receiving partial or full tuition coverage, much of which is paid using the other students’ tuition and state funding.
“As a state university I do not think it is aligned with our public mission to use such funds to subsidize tuition for primarily more privileged students,” Nelson said in the email.
As campus has cut back on the number of undergraduate TAs in the CS department, some computer science and data science classes have begun adopting AI tools such as Pensive AI Grader in the grading process, reducing TAs’ workload.
Senior CS major Daniel Wang, who serves as a head undergraduate TA for CS 61B, “Data Structures,” shared that the course staff considered using AI for grading, as well as implementing AI bots to answer questions on Ed Discussion, but ultimately decided against it.
“I think the innate value of having staff members is that you can interact with students,” Wang said. “Human interaction is the best way to intimately understand the course material.”