January 1, 1970

Tech Industry Layoffs and the CS Graduate Crisis: What the Data Shows

Empty tech office with rows of vacant desks symbolizing mass layoffs in the software industry

There's a cruel irony sitting at the center of the 2025-2026 tech job market. Computer science graduates earn the highest starting salaries of any college major — around $80,000 according to Federal Reserve Bank of New York data — yet their unemployment rate, 6.1%, now exceeds that of philosophy majors. Philosophy majors. If you graduated with a CS degree in the last two years and couldn't find a role, you're not imagining it. The writing was on the wall as early as 2023, but most universities kept enrolling students at record rates while the floor dropped out from under them.

The Scale of What Actually Happened

Between January 2022 and January 2025, entry-level software engineering job postings dropped 65%. That figure comes from Lightcast (formerly EMSI Burning Glass), which tracks millions of job postings across North America. Sixty-five percent. Not 10, not 20. Two-thirds of the entry-level market evaporated.

The supply side moved the opposite direction. CS bachelor's degrees more than doubled in a decade, rising from 51,696 awarded in 2013-2014 to 112,720 in 2022-2023. Enrollment continues climbing at 4.3% year-over-year, even as overall bachelor's degree enrollment declined 3.0%. More graduates. Far fewer jobs.

The layoff wave piled on. According to Crunchbase, Amazon disclosed 30,184 cuts in 2025, Intel 27,058, and Microsoft 15,347 — those three companies alone accounting for roughly two-thirds of all tracked layoffs. Experienced engineers flooded back into the candidate pool. Companies suddenly had their pick of five-year veterans applying for roles that, two years earlier, would have gone to someone straight out of college.

Year	Approx. Tech Layoffs	Entry-Level Postings vs. 2022 Baseline
2022	~80,000	100%
2023	~260,000	~65%
2024	~152,000	~45%
2025	~127,000	~35%

Why Entry-Level Jobs Got Gutted Specifically

Senior hiring rebounded at many companies. Entry-level hiring did not. That asymmetry is the critical detail most coverage misses.

SignalFire's talent data tells the clearest story: new graduate hiring at the 15 largest tech companies fell more than 50% since 2019. The share of entry-level roles in Big Tech dropped from 15% to 7% of all hires. The average hire age increased by three years since 2021. Companies didn't just pause junior pipelines — they quietly dismantled them.

Heather Doshay, SignalFire's head of talent, said it directly: "Companies are prioritizing experienced hires over junior talent, and we're seeing smaller funding rounds, shrinking teams, fewer new grad programs, and the rise of AI all contributing to this downturn."

Three structural forces converged to make these cuts permanent rather than cyclical:

Funding contraction: Smaller funding rounds meant smaller teams. Startups that used to hire one senior engineer plus two or three juniors shifted to senior-only models almost overnight.
Remote work pullback: Remote-friendly junior roles declined 71% since 2022, per Indeed's 2026 Tech Hiring Outlook. Junior roles went remote disproportionately during the pandemic; when office mandates returned, many of those positions vanished with the commute.
AI productivity tools: Companies found they could squeeze more output from senior engineers using AI coding assistants, quietly eliminating the workflow that once justified junior hires.

And there's a detail that doesn't get enough attention: over 80% of Bay Area entry-level postings now require a minimum of two years of experience, according to Lightcast data. The job is labeled entry-level. The requirements are not.

The AI Factor: Cause, Accelerant, or Scapegoat?

AI gets blamed for everything right now. Here the blame is partially earned, but the picture is messier than "AI took the jobs."

A Stanford study tracking millions of workers through mid-2025 found that software developers aged 22 to 25 in highly AI-exposed roles experienced a 13% employment decline since the emergence of large language model coding tools, while older developers in the same roles saw steady or rising employment. Employment among US software developers aged 22 to 25 has fallen roughly 20% from its late-2022 peak. That's a specific, age-stratified collapse — not a general tech slowdown.

"Right now, the only type of employee anybody's interested in hiring is a relatively heavyweight senior person who is very technical." — James O'Brien, UC Berkeley CS professor

Junior employment at AI-adopting companies declined 9-10% within six quarters of AI implementation, while senior employment stayed essentially flat, per available labor data. This makes structural sense. Junior developers typically handle the exact tasks that AI coding assistants do well: boilerplate code, unit tests, simple CRUD functions, documentation stubs. Senior engineers handle what AI still gets wrong — system architecture, cross-team dependencies, security trade-offs in legacy codebases.

Salesforce announced zero new engineering hires for 2025 while publicly crediting AI for team productivity gains. Shopify made similar statements. These aren't temporary freezes. They're strategic postures that signal a structural reduction in junior headcount across the industry.

One counterpoint worth taking seriously: a METR study found that AI tools actually decreased developer productivity by 19% for experienced engineers on complex tasks. AI doesn't replace senior developers. But it absolutely absorbs the specific workflow that once justified hiring junior ones.

Who Is Still Getting Hired

Not every CS graduate is struggling. The market didn't collapse uniformly across specializations.

Demand for AI skills in early-career job postings nearly tripled since fall 2025, appearing in 35% of early-career listings. Cybersecurity, cloud infrastructure, and AI/ML engineering remain at near-full employment. The candidates landing offers look measurably different from the ones grinding through month seven of rejections.

Handshake's 2025 data found that 78% of CS students who secured jobs within three months of graduation had either a strong GitHub portfolio, at least one internship, or a specialized certification. Only 31% of students relying on degree credentials alone succeeded in the same window. The credential isn't the differentiator anymore. Demonstrated output is.

Here's what the market is actually paying for:

AI/ML engineering (PyTorch, fine-tuning pipelines, retrieval-augmented generation): appears in 35%+ of new postings
Cloud infrastructure with AWS, GCP, or Azure certifications: consistent demand, especially with any security clearance eligibility
DevOps and platform engineering: companies are cutting headcount but keeping their infrastructure tight
Cybersecurity: 3.5 million open positions globally, largely unfilled — one of few genuinely undersupplied areas

The students getting hired built real things. Not class assignments. Deployed applications, open-source contributions, AI-integrated side projects with actual users. They shipped.

The Skills Gap Universities Won't Discuss

The Cengage Group's 2025 Graduate Employability Report has a finding that should embarrass every CS department chair in the country: only 40% of computing and tech instructors say their 2025 graduates are "very or sufficiently prepared" to enter the workforce. And 20% of students reported their program provided no help whatsoever fostering professional connections.

One in five CS students graduated with no professional network. In a market where, by Cengage's own data, 25% of students cite personal connections as the biggest factor in landing a relevant job. That's not a small gap. That's a structural failure.

The career readiness problem has four dimensions:

Building in public: GitHub activity, open-source contributions, and writing about technical problems are now active filters in hiring pipelines — many applicant tracking systems flag empty GitHub profiles.
Networking starting year one, not senior spring — personal connections outperform cold applications by a wide margin in the current market.
Two internships before graduation, ideally at companies outside the FAANG tier, where new grads actually get mentored rather than tracked through an automated pipeline.
Specialization over generalism — the "full-stack developer" label generates fewer callbacks than "ML infrastructure engineer" or "cloud security analyst."

Here's an uncomfortable footnote (and one universities rarely advertise): Microsoft applied for 4,712 H-1B visas in fiscal year 2025 while simultaneously conducting domestic layoffs. The skills gap isn't a myth companies invented to justify cheaper labor. They genuinely cannot find enough specialized talent domestically, and the generalist pipeline is oversupplied.

What CS Graduates Should Actually Do Right Now

The defeatist reading of this data leads to "don't study CS." That's wrong. It misses what's actually happening.

LinkedIn's chief economist Aneesh Raman compared the current disruption to the 1980s hollowing of manufacturing — painful at the entry level, not the end of the field. Manufacturing didn't disappear; it transformed. The same is underway in software. The graduates who position themselves for where the field is going, rather than mourning where it was, will be fine.

The bar for junior roles has permanently risen. You're not competing against other new grads. You're competing against AI tools, against experienced engineers who took pay cuts to stay employed, and against a global talent pool. Showing up with just a degree to that competition is like bringing a résumé to an audition.

If you're currently enrolled in a CS program:

Pursue internships aggressively starting sophomore year — apply to 50+ companies, not just the usual top ten
Pick a specialization by junior year: cybersecurity, cloud, or AI/ML have meaningfully stronger demand than general software engineering
Deploy one real project per semester — something used by actual people, not a grade-graded submission

If you recently graduated and are job hunting:

Target mid-size companies (100 to 2,000 employees) — Big Tech has the highest competition and the lowest new-grad conversion rate right now
Get an AWS Solutions Architect or Google Cloud Professional certification; each takes 6-12 weeks and consistently generates callbacks
Expect a longer search — budget 6-9 months, not the 6-week timelines that were common in 2021

If you're advising students or running a program:

89% of computing instructors agree students need hands-on generative AI experience before graduation, per Cengage's report; make it a core requirement, not an elective
Build industry partnerships that produce real internship pipelines — not annual career fairs that generate résumé drops into the void

Bottom Line

The tech layoff wave disproportionately destroyed entry-level opportunity. Not the field itself. CS graduates face a market where the credential signals capability but experience proves it — and companies are demanding proof. The 6.1% unemployment rate for CS graduates will not last forever, but it will persist for anyone who treats a degree as a destination rather than a foundation.

Specialize early. The "generalist software engineer" profile is the hardest sell in this market. AI/ML, cybersecurity, and cloud infrastructure have distinct, documented demand.
Internships are not optional. The gap between 78% placement (students with experience) and 31% (degree-only) is the single most important data point in this entire discussion.
Budget for a longer search. Six to nine months is realistic post-graduation. Plan financially for it rather than burning bridges by taking whatever pays in month two.
Mid-size companies are underrated. Smaller teams train junior engineers because they have to — there's no senior-heavy bench to absorb all the work. FAANG runs AI-assisted pipelines and hoards senior talent; the learning curve for a new grad is brutal.

The field will recover. The graduates who build for the recovery rather than grieve the boom that already ended are the ones who will be fine.

Frequently Asked Questions

Is a CS degree still worth it in 2026?

Yes — but with a significant caveat. The degree builds foundational knowledge that AI tools still can't replicate: system design, algorithm analysis, debugging complex multi-component systems. The problem is the degree alone no longer guarantees an entry-level role. Students who combine the degree with internship experience, a specialization, and demonstrable projects get hired. Those who don't, often don't.

Why do CS graduates have higher unemployment than philosophy majors right now?

Pure supply and demand mismatch. CS enrollment doubled in a decade while entry-level job postings dropped 65% between January 2022 and January 2025. Too many graduates are chasing too few junior openings at exactly the wrong moment in the cycle. CS skills remain valuable — the issue is that junior-level expression of those skills is currently oversupplied.

Is AI actually replacing junior developers, or is that overstated?

For specific workflows, it's real. AI coding assistants have absorbed much of what once justified hiring junior developers: generating boilerplate, writing tests, documenting functions. The Stanford study tracking developers by age found a 13% employment drop among 22-25 year olds since 2023, while older developers held steady. That age-stratified pattern is hard to explain any other way. Senior roles were not affected the same way because complex system work still requires human judgment.

What specializations give CS graduates the best shot in 2026?

Cybersecurity, cloud infrastructure (particularly with AWS, GCP, or Azure certifications), and AI/ML engineering are the three areas with the clearest demand signal. Data engineering and DevOps/platform engineering also show consistent posting volume. General "software engineer" or "full-stack developer" profiles face the most competition and the lowest callback rates in the current environment.

Should new CS grads focus on big tech or startups?

Mid-size companies — roughly 100 to 2,000 employees — are the underrated option right now. Big Tech has concentrated the most competition and reduced new-grad hiring the most sharply. Early-stage startups offer experience but carry real financial risk. Mid-size established companies have stable engineering teams that genuinely need junior talent to grow, and they typically provide more mentorship than a hyperscaler where you're one of 400 new hires in a cohort.

How long should recent CS graduates realistically expect their job search to take?

Plan for six to nine months. The 2021 market — where offers arrived within weeks of graduation — is not coming back soon. Graduates who landed roles in 2024 and 2025 typically applied to 200 or more positions, pursued certifications between applications, and frequently found their role through networking rather than a cold application. The financial planning matters as much as the job search strategy.