Why Math Majors Are Gaining as CS Grad Underemployment Hits 19%

The underemployment rate for computer science graduates stands at 19 percent according to 2024 Federal Reserve Bank of New York data, and undergraduate enrollment in CS programs has dipped by 8.4 percent per the National Student Clearinghouse Research Center, figures gathered by Fast Company on August 29. Meanwhile mathematics enrollment, which bottomed out in 2022, has begun climbing again. If you are choosing a major or considering a pivot, the useful question is not which subject is winning but what each one actually buys you. At Metaintro, we think that distinction is where most degree advice goes wrong.

How Did Computer Science Get Oversupplied?

Through a decade and a half of entirely rational decisions arriving all at once. After the 2008 recession, students moved toward a field with a visibly promising outlook, and between 2008 and 2024 the number of computer science degrees grew by almost 500 percent according to figures reported by The Washington Post. During that period computer science became the third most popular major behind nursing and psychology, making up almost 5 percent of early career college graduates.

A 500 percent increase in graduates does not require a collapse in demand to produce a problem. It only requires demand to grow more slowly than that, which it did. The result shows up as underemployment rather than unemployment, meaning graduates are working but in roles that do not require the degree. That distinction matters because it is easy to misread. We covered the wider measure in the U6 underemployment rate and the personal experience of it in being underemployed and overlooked.

Is Computer Science Actually a Bad Bet Now?

This is where the headline version breaks down, and the federal data is worth reading carefully before anyone changes a major. According to the Bureau of Labor Statistics Occupational Outlook Handbook, overall employment in computer and information technology occupations is projected to grow faster than the average for all occupations from 2025 to 2035, and the median annual wage for the group was 109,470 dollars in May 2025 against 50,980 dollars for all occupations.

That is not the profile of a dying field. It is more than double the median wage across the economy and above average projected growth. So the honest framing is that the computer science occupation is healthy while the computer science graduate pipeline is oversupplied, and those are different problems with different solutions. An oversupplied pipeline means more competition for entry level roles, which is exactly what a 19 percent underemployment rate among graduates describes.

Tech layoffs at large employers have reinforced the perception, and we have tracked those closely, including Oracle's headcount dropping by 21,000. Perception and projection are pointing in different directions right now, and students are responding to the perception.

What Is Actually Happening to Math Enrollment?

A modest recovery rather than a stampede, and precision matters here. Mathematics degree enrollment dipped after 2018 and bottomed out in 2022. Since then it has risen slowly, growing almost 2 percent by 2024 according to data from the Integrated Postsecondary Education Data System.

Two percent is a real change of direction but it is a small number, and it is worth holding that next to the 8.4 percent decline in CS enrollment. Students are not abandoning computer science for mathematics in large numbers. A modest math recovery is happening at the same time as a larger CS decline, and treating those as one migration overstates both.

It is also worth remembering how small the mathematics base is to begin with. Computer science accounted for almost 5 percent of early career graduates at its peak, while mathematics has never been anywhere near that share, so a 2 percent rise on a small base adds far fewer people to the market than an 8.4 percent fall removes from a large one. If you are worried about entering a field that is about to be flooded the way computing was, the arithmetic says that is not currently happening.

The occupational outlook for math is genuinely strong, though. The BLS Occupational Outlook Handbook for math occupations projects overall employment in the group to grow much faster than the average for all occupations from 2025 to 2035, with a median annual wage of 107,570 dollars in May 2025.

So Which One Pays More?

Almost the same, and this is the finding that should temper the entire discussion. Computer and information technology occupations carry a median annual wage of 109,470 dollars, and math occupations 107,570 dollars, both as of May 2025. The gap is under 2,000 dollars, and it favors computing.

Where math has the edge is in projected growth, described as much faster than average against faster than average for computing. That is a difference in trajectory rather than in current pay, and trajectory matters more the earlier you are in a career.

A caution on the figures circulating elsewhere. Some coverage cites a 28.7 percent growth projection for mathematics occupations covering 2021 to 2031. That is a genuine number from an older BLS projection cycle, but it has since been superseded twice, and the current cycle runs 2025 to 2035. When a projection is two cycles old, use the current one. Our guide to data scientist salary by experience and city covers one of the roles sitting at the overlap of both fields.

Why Are Quant Firms and AI Labs Bidding for Mathematicians?

Because the mathematics underneath modern AI has become commercially valuable in a way it was not a decade ago. The rise of AI has put a premium on skills like stochastic calculus, linear algebra and numerical methods, since people with advanced mathematical training understand the foundational logic of these systems and can spot errors and biases that are invisible to someone using the tools at the surface.

The same reasoning applies in finance. Quantitative firms use advanced mathematical modeling to analyze financial data, optimize portfolios and profit from market inefficiencies, and the ability to write new financial algorithms puts mathematicians in direct demand at hedge funds and trading firms. Fast Company reports, citing Forbes, that AI companies including OpenAI and Anthropic now offer salaries ranging from 180,000 dollars to 2 million dollars, and that competition has pushed quant firms to raise pay in response.

The salary movement in that niche is striking. Fast Company cites data from the finance careers site eFinancialCareers showing quant associate salaries grew 13 percent between 2024 and 2025, with quant vice presidents up 10 percent and directors up 18 percent, while over the same period tech analyst salaries fell 36.3 percent and associate compensation dropped 13.8 percent. Those quant numbers describe a small and highly selective corner of the market, not a general graduate outcome, and the reported starting range of 350,000 to 500,000 dollars for new math graduates at quant firms applies to a handful of firms hiring a handful of people. Read it as the ceiling, not the expectation. We covered one visible example in Susquehanna paying interns 86,000 dollars for a summer.

What Are Universities Doing About It?

Building a third option. Institutions including Carnegie Mellon, MIT and Northwestern have introduced AI majors, and MIT's Artificial Intelligence and Decision-Making major is now the second largest at the school. That is a fast structural response by the standards of higher education, and it tells you where those institutions think the demand is going.

Whether a dedicated AI major is better than mathematics or computer science plus AI coursework is genuinely unsettled, and anyone starting a four year degree now is making a bet on what the field looks like when they graduate. The safer read is that the mathematical and statistical foundations are the part least likely to be obsolete, because they underpin whatever the tooling becomes.

That view is consistent with what employers keep saying about the primacy of underlying capability, which we covered in why skills matter more than a computer science degree and in the surprising finding that philosophy majors are out hiring computer science graduates in parts of the AI job market.

Is Coding Still Worth Learning?

Yes, and the most useful line in the reporting is about combination rather than substitution. Daniel Hughes, director of quant research and trading at the recruitment firm Selby Jennings, said the biggest edge someone has is to be a mathematician who can code, describing candidates who have proven they can sit with a problem for a very long time, typically five years for a doctorate, and who then work in Python.

Note what that describes. Not a mathematician instead of a programmer, and not a programmer instead of a mathematician. The premium sits on the intersection, and that is the practical takeaway for anyone currently studying either subject.

The reporting also carries a useful corrective from a quant recruiter who spoke anonymously, saying that AI models at quant firms are operating at roughly the level of a median intern and cannot run financial analysis the way a seasoned mathematician can. Meanwhile computer scientists are needed to train those models and to code subagents for more granular tasks, and many funds have coding projects that still require a human in the loop. That is a specific and current description of where the automation frontier actually sits in one demanding industry, and it is more informative than the general claim that AI is taking coding jobs.

What Does This Mean for Your Career Right Now?

If you are choosing a major, choose on foundations rather than on this year's headlines. The 500 percent growth in CS degrees happened because a cohort read a signal correctly and then acted on it in such numbers that the signal stopped being true. Anyone piling into mathematics purely because of quant salaries risks repeating that exact mistake on a four year delay.

If you are a computer science student or recent graduate, do not read 19 percent underemployment as a verdict on the field. The occupation pays a median of 109,470 dollars and is projected to grow faster than average. What the number tells you is that the entry level is crowded, so differentiation matters more than it did for the cohort before you. Adding real mathematical depth, in linear algebra, statistics and numerical methods, is one of the clearest ways to do that, and the intersection is where the demand is.

If you are already working and considering a pivot, the intersection point applies to you too. You do not need a doctorate to benefit from stronger quantitative foundations, and roles in data analysis, risk and modeling sit between the two fields. Our guide to the quantitative skills employers want is the practical starting point, and a 90 day plan to build AI skills covers the tooling layer on top.

And keep the quant numbers in proportion. Starting salaries of 350,000 dollars exist, and they go to a very small number of people at a very small number of firms, most of whom hold doctorates. Using them to model your expected outcome is like planning a music career around a stadium tour. The realistic version of this story is a healthy median wage in both fields, a crowded entry level in one, and a genuine premium for anyone who can do both. At Metaintro, we would rather you build toward that intersection than chase whichever major the internet is currently excited about.

Related Articles

• Computer Science Jobs
• Software Engineer Salary in 2026 by Level and City
• AI Skills Appear in 73 Percent of Tech Job Postings
• The New Grad Job Market in 2026 and AI at the Entry Level
• The Old Entry Level Playbook Is Dead
• AI Is Coming for Entry Level Jobs, What New Grads Should Know
• Is a UK Degree Worth It When 1 in 4 Graduates Lose Money
• Kettering University Ranked First for Career Outcomes on a Co-op Model
• India's CS Graduates and the AI Skills Gap
• Best Employers for New Graduates

People Also Asked

Q: Is a computer science degree still worth it in 2026?

A: The occupation remains strong even though the graduate market is crowded. BLS data puts the median annual wage for computer and information technology occupations at 109,470 dollars in May 2025, with employment projected to grow faster than average from 2025 to 2035. The 19 percent graduate underemployment rate reflects oversupply at the entry level rather than a shrinking field.

Q: Do math majors earn more than computer science majors?

A: Not at the median, at least by occupation. Math occupations had a median annual wage of 107,570 dollars in May 2025 against 109,470 dollars for computer and information technology, so computing is slightly ahead. Math occupations are projected to grow much faster than average, compared with faster than average for computing, so the difference is in trajectory rather than current pay.

Q: What skills do quant firms and AI labs actually want?

A: Advanced mathematics applied to code. The specific areas named are stochastic calculus, linear algebra and numerical methods, alongside the ability to work in Python. One quant recruitment director described the strongest profile as a mathematician who can code, which points to the intersection of the two disciplines rather than to either one alone.

Ready to level up with the skills that actually transfer? Metaintro matches you to roles across data, engineering and quantitative finance based on what you can do rather than which department awarded your degree. Create your free profile and see where your foundations already qualify you.