Entry-level tech jobs are disappearing because the tools have moved up the value chain: LLMs and coding assistants now write the first-pass code and handle the routine debugging that used to be a junior developer’s first year on the job. Stanford SIEPR’s 2025 ADP-payroll study (Brynjolfsson, Chandar, Chen) shows employment for 22-to-25-year-olds in AI-exposed occupations dropped ~13% since late 2022, with software engineering and customer service down ~20% between late 2022 and mid-2025 — the first large-scale evidence that AI adoption is compressing entry-level demand, not just tech-cycle noise. The workaround, if you’re a 2026 grad or a career-changer, is not “learn to code harder”; it’s picking the segments and formats that still buy the pipeline.
TL;DR
- Down ~67% in entry-level developer opportunities since 2022 (hire volume, aggregated LinkedIn/Indeed data)
- Stanford SIEPR (2025): 22-25-year-olds in AI-exposed fields down ~13%; software engineering + customer service down ~20% between late 2022 and mid-2025
- UK: entry-level tech roles fell 46% in 2024, projected 53% by end of 2026
- 21% of US hiring managers have already frozen entry-level hiring due to AI; 36% will stop by end of 2026 (Resume.org survey of 1,000 managers)
- 66% of global enterprises plan to cut entry-level hiring due to AI (IDC / Deel surveys, 2026)
- Where hiring holds up in the EU: enterprise vendors, healthcare platforms, financial institutions, infrastructure companies — because the pipeline can’t be hollowed forever
- What still works: shipping a real project, using AI as a tool not a threat, applying where juniors are systematically hired rather than fighting for 800-applicant consumer-SaaS slots
What the numbers actually say
For most of the 2010s, “junior developer” was a well-defined slot in every reasonably-sized tech company: entry-level offer, 6-18 months of ramp-up, promoted or moved on. 2022 broke the pattern.
Aggregations of LinkedIn and Indeed postings across 2022-2026 show entry-level developer opportunities dropped roughly 67% by hire volume, while the number of postings for junior roles actually rose ~47% (per hakia.com aggregation, 2026). Postings rising while hires fall is not a paradox — it’s employers using junior-labelled listings to source cheap candidates for what turn out to be mid-level requirements. The postings-vs-hires gap has been called “seniorization”: the role is labelled junior but the responsibilities are not.
Stanford SIEPR’s 2025 working paper (Erik Brynjolfsson, Bharat Chandar, Ruyu Chen) was the first large-scale study to isolate AI’s contribution using ADP payroll data for 25 million US employees. The headline finding: employment for 22-to-25-year-olds in AI-exposed occupations dropped ~13% since late 2022 — even after controlling for other macro shocks — while older workers in the same sectors held steady. Software engineering and customer service specifically dropped ~20% between late 2022 and mid-2025 (per siepr.stanford.edu, 2026; coverage at Fortune and Time). That’s not the whole labour market; it’s the young end of the AI-exposed slice, which is the exact demographic entry-level developer roles draw from.
The Resume.org survey of 1,000 US hiring managers, published mid-2026, put a number on employer intent: 21% have already frozen entry-level hiring specifically because of AI, 36% expect to stop entry-level hiring by end of 2026, and 47% expect entry-level roles to be eliminated at their company by 2027 (per resume.org, 2026). IDC and Deel’s global enterprise survey landed at 66% of global enterprises planning to cut entry-level hiring due to AI.
Why AI is the cause, not the tech cycle
The 2022-2023 tech layoffs were real, and correlation between AI adoption and junior hiring collapse could plausibly be a tech-cycle artefact. Stanford SIEPR’s methodology is what rules that out: their model controls for firm-level, occupation-level, and macro shocks and still finds a significant relative decline for 22-to-25-year-olds in the MOST AI-exposed occupations vs. older workers in the SAME sectors. Same firms. Same industry cycle. Different age band. Different outcome.
The mechanism is straightforward: LLM output overlaps heavily with the skills that formal CS education produces — writing code from a spec, drafting boilerplate, doing first-pass debugging, writing test scaffolding. That’s the exact work profile a junior developer does in year one. Senior work — deciding what to build, negotiating requirements, spotting the load-bearing architectural decision — doesn’t overlap as cleanly, and AI hasn’t compressed demand there.
Some interpretations of the same data emphasise that AI is not eliminating junior work so much as changing what junior work looks like — the “seniorization” thesis. Junior postings now require judgment and stakeholder management (skills that used to define mid-level roles), because the routine work AI can handle no longer justifies a dedicated head count. Both readings agree on the shape of the effect: fewer entry-level offers, higher bar for the ones that remain.
Where the EU market splits
The EU-wide picture is patchier than the US, partly because national labour statistics agencies don’t break out “junior developer” as its own bucket. What is measurable:
| Market | Signal | Source |
|---|---|---|
| UK | Entry-level tech roles fell 46% in 2024, projected 53% decline by end of 2026 | Multiple UK tech-hiring reports, 2026 |
| EU-wide | 57% of EU firms report they can’t find qualified tech staff | index.dev EU tech-market report, 2026 |
| Germany / Netherlands | Skilled-worker shortage across the stack; junior hiring holds up better in enterprise / financial services than consumer SaaS | Multiple national IT-employer surveys, 2026 |
| Southern EU (Spain / Portugal / Italy) | Sharper junior contraction, offset by EU expat-tax regimes drawing mid-senior remote workers in | Regional hiring aggregators, 2026 |
The “57% of EU firms can’t find qualified staff” figure and the “junior hiring down 67%” figure are consistent, not contradictory. Employers cite unmet demand at the SENIOR end (5+ years of specific-stack experience) while cutting the pipeline that produces senior developers at the JUNIOR end. That gap is what breaks in 2-4 years if nothing changes.
Which segments still hire juniors systematically
Consumer SaaS and fast-moving startups are where junior hiring collapsed hardest — those employers over-index on shipping speed and can substitute AI + one senior for two seniors + three juniors of headcount. The segments where the pattern is different, in the EU as elsewhere:
- Enterprise software vendors (SAP, Software AG, Sage, Amadeus, Sopra Steria) — long deployment cycles mean today’s junior is next year’s mid-senior on the same customer install. Hollowing the pipeline breaks the account book.
- Financial institutions with in-house engineering (Nordea, KBC, Erste Group, BNP, Santander, ING) — regulated environments where credentialing takes 12-24 months anyway; a junior is a 2-year investment they’ve already planned around.
- Healthcare platforms and med-tech (Doctolib, Kry, Medtronic, Boston Scientific) — same credentialing dynamic as finance; also less exposed to generative AI’s early-productivity gains because the code is safety-critical.
- Infrastructure / dev-tools vendors (JetBrains, Elastic, GitLab EU offices, Storyblok) — companies that sell to developers still hire developers with real ramp-up runway.
- National champion / gov-tech — Digital agencies of France, Germany, the Netherlands still hire juniors at scale, though the process is slower and lower-paid than the private sector.
Xeito’s own EU-workable listings feed skews toward the first four segments — that’s where remote-first junior hiring is still happening across the EU.
What to do differently as a 2026 candidate
Apply where juniors are hired, not where you’d rather work. A 2019 CV strategy (apply to 60 consumer-SaaS startups, cast a wide net) does not work in 2026 — you’re competing against 800+ applicants for each of those roles. Applying to 15 enterprise-vendor postings in your city + region is a materially different funnel.
Ship one project a hiring manager can clone and run in under 5 minutes. Portfolio depth matters more than portfolio count. One complete side project with a real README, a test suite, and CI beats fifteen half-finished tutorial forks. This is the single highest-leverage investment for a junior candidate in 2026.
Learn the AI tools rather than avoiding them. The “seniorization” thesis says junior roles now expect judgment; the practical implication is that you’re expected to close the gap between what you know and what the role demands with AI in the loop. Being credible on “I ship with Copilot / Cursor / Claude Code and know when to override them” is now table stakes, not a differentiator.
Weight remote-first vs. hybrid honestly. Remote-first employers over-index on demonstrated autonomy — they can’t compensate for a junior’s ramp-up with hallway mentorship. If your CV is specific and your portfolio is deep, remote-first is a fair fight. If it’s generic, hybrid or on-site will land you a role faster. Not romantic advice, but true.
Don’t skip cover letters at junior level. At the senior end, cover letters get skimmed; at the junior end, they carry more weight because you don’t have the résumé signal yet. A short, specific note (three sentences: why this company, what you’d contribute in the first 90 days, what you built that’s relevant) moves you from the reject stack to the interview stack.
The medium-term picture
Both the “seniorization” and “AI replaces juniors” theses agree on the direction: fewer entry-level roles, higher bar per role, faster expected ramp-up. They disagree on whether this stabilises or continues down. The signal to watch is not the aggregate hiring number but the ratio of postings to actual hires — right now, postings up 47% and hires down 67% says employers are testing the market at low prices, not investing in the pipeline.
The pipeline break, if the current pattern continues into 2027, is what forces a correction. Senior developers don’t spontaneously appear; they’re former juniors with 3-5 years of on-the-job learning. When mid-2020s hiring rebuilds the senior end of the market and finds the middle empty, the incentive to hire juniors returns — but only for the employers who kept hiring them through the down-cycle. Which brings the choice back to the individual candidate: apply where the pipeline is still open.
If you’re looking for EU-workable remote roles from employers with real junior on-ramps, Xeito filters listings from 130+ sources to isolate the “junior + remote + EU-workable” slice specifically — not “remote (US only)” noise, not senior-only postings dressed as junior.
Sources
- Stanford Digital Economy Lab — Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (Brynjolfsson, Chandar, Chen — 2025 working paper)
- Fortune — Stanford economist Erik Brynjolfsson: ‘It’s not going away’ (2026 follow-up coverage)
- Time — Who’s Losing Jobs to AI? New Stanford Analysis Breaks It Down
- Resume.org — 2026 Entry-Level Hiring Survey (1,000 US hiring managers)
- Hakia — The Junior Developer Extinction: 67% Hiring Collapse (LinkedIn / Indeed posting aggregation, 2026)
- Index.dev — Europe’s Tech Job Market: Statistics and Trends for 2026
- CACM — Junior Developer Extinction: What the ADP Numbers Actually Say (ACM commentary on the SIEPR methodology)