STEM and healthcare degrees claim 10 of the top 20 positions in a new employment ranking of 74 college majors with mathematical sciences projected to grow 28.4% through 2033, the fastest rate of any occupational category tracked by the Bureau of Labor Statistics. Nursing ranked first overall with a 1.42% unemployment rate, followed by mathematics, computer science, and four engineering disciplines. The analysis, published by National University in November, combines Federal Reserve Bank of New York unemployment data with BLS employment projections to score majors on job security and growth potential.
The “STEM dominance” narrative is being carried by two very different engines: patient-facing healthcare roles anchored by irreversible demographics, and data science positions riding a wave of AI investment that may or may not pay off. Meanwhile, the laboratory and R&D workforce, chemists, microbiologists, biological technicians, is projected to grow at 2% to 5%, barely outpacing plumbers and lawyers.
The healthcare ceiling
A large share of the job growth lumped into some “STEM and healthcare” rankings comes from nursing, which, depending on who’s counting, may or may not qualify as STEM directly. (The BLS excludes nursing from its core STEM tables; other federal agencies treat it differently.) Regardless of classification, the numbers are significant: the BLS projects roughly 189,100 registered nurse openings per year through 2034, and 35% growth for the combined category that includes nurse practitioners, nurse anesthetists and nurse midwives.
Meanwhile, medical scientists, essentially PhD-track research roles, show 9% growth. For the sake of comparison, that is roughly the same pace as electricians, a field that Nvidia CEO Jensen Huang told the World Economic Forum last week will see “six-figure salaries” as data center construction accelerates. Other scientists aren’t in the same growth category even if they are areas in or adjacent to medicine. Biological technicians (+3%), microbiologists (+4%) and clinical laboratory technologists (+2%) are growing at or below the national average. In other words, the healthcare boom is concentrated in direct patient care rather than in research.
The data science paradox
At the other end of the spectrum, data scientists are projected to grow 36% through 2033, with operations research analysts close behind at 23%. The BLS projections reflect genuine demand: 51% of AI-related job postings now originate outside traditional tech sectors, and industries from agriculture to logistics have increased AI hiring budgets by 42% since 2023.
But there’s a catch. The projections assume that current AI investments will produce sustained productivity gains. If the billions flowing into generative AI don’t deliver ROI beyond the current 2.7% productivity baseline, demand could correct sharply. And on the ground, the labor market is already bifurcating by experience level.For years, FAANG and the broader “Big Tech” cohort set the market on pay and prestige. That’s not as true as it once was. SignalFire’s 2025 State of Tech Talent report found that new graduate hiring at major tech companies has dropped 50% from pre-pandemic levels. Now, new grads now accounting for about 7% of hires. A Stanford Digital Economy Lab analysis of ADP payroll data documented a 20% employment decline for software developers aged 22-25 from late 2022 to September 2025. Meanwhile, workers over 30 in the same roles saw 6% to 9% growth.
Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla has fallen by more than half since 2022, according to SignalFire’s 2025 State of Tech Talent report. Fresh grads now account for less than 6% of new hires at startups, down more than 30% from pre-pandemic levels. The average age of technical hires has increased by three years since 2021.
“Times have changed, and lean is in,” said Heather Doshay, SignalFire’s head of talent. “Companies are prioritizing experienced hires over junior talent.”
The lab workforce: Automated at the Bench
The pattern in software is repeating in laboratory science. Routine tasks like titration, sample preparation and basic data entry are increasingly handled by automated systems and data operations workflows. A single lab can now process higher volumes with fewer technicians.
This isn’t a sign of a dying field; it’s a maturing one. But it means that entry-level lab roles face the same “junior purge” pressure as entry-level coding jobs. The output of the life sciences sector is growing, but the labor intensity of entry-level positions is declining.
What Actually Looks Resilient
The BLS data points to a few pockets of durable growth, and they share a common thread: work that’s hard to automate because it requires physical presence, adversarial thinking, or judgment under uncertainty.
Information security analysts (+29%): The BLS projects 29% growth through 2034, driven by what the agency calls “the frequency and severity of cyberattacks and data breaches on U.S. businesses.” The U.S. currently has roughly 500,000 unfilled cybersecurity positions, according to CyberSeek. Unlike data science, where entry-level roles are getting squeezed, cybersecurity has a genuine supply problem, 90% of hiring managers won’t consider candidates without prior IT experience or certifications, which creates a barrier but also protects incumbents from the bootcamp-grad glut.
Mechanical engineers (+9%) and embedded systems: Mechanical engineering is projected to grow 9% through 2034, with the BLS citing “automation and innovation” as drivers, specifically, the need to integrate complex automation machinery into existing systems. Work that bridges software and physical hardware (sensor integration, real-time constraints, production troubleshooting) resists virtualization. When a device is miscalibrated or a factory line is behaving strangely, progress depends on hands-on diagnosis with oscilloscopes and multimeters, not prompts.
Medical scientists (+9%): The PhD-track research roles show 9% growth, faster than the technician positions that support them. The premium is shifting toward experimental design, interpretation, and translating results into new hypotheses. Execution-oriented work is getting automated; the ability to ask the right questions is not.
The unemployment paradox
The New York Fed data reveals some counterintuitive comparisons. Recent computer science graduates face a 6.06% unemployment rate, higher than fine arts graduates (6.1%), and not far from philosophy majors (5.2%) or political science grads (5.0%). Nursing holds at 1.42%, civil engineering at 1.05%.
But unemployment rates only tell part of the story. Fine arts graduates who do find work earn significantly less than their CS counterparts, the median early-career wage for CS is roughly double that of most liberal arts fields. The comparison isn’t “CS is as bad as art school.” It’s that CS no longer guarantees the easy on-ramp it once did, even if the ceiling remains high.
And that ceiling has never been higher, for an extraordinarily small slice of the field. Meta upended traditional pay structures by offering pay packages worth as much as $300 million over four years in some cases. The heat is concentrated in a narrow band: researchers pushing the frontier of foundation models, engineers who can ship production ML systems at scale and the senior architects who can orchestrate it all.
The STEM labor market is bifurcating. On one side: patient-facing healthcare roles with demographic tailwinds and physical-world requirements that resist automation. On the other: data science and software positions with strong projected growth but high volatility, especially for entry-level workers.
Tell Us What You Think!
You must be logged in to post a comment.