ChatGPT 
Summary
Key takeaways
- AI is reshaping software engineering rather than eliminating it: entry-level hiring is contracting sharply while demand for experienced engineers remains comparatively strong.
- AI was cited in 87,714 announced US job cuts through May 2026, representing 22% of all layoffs, but these figures describe employer-stated layoff reasons rather than a complete measure of jobs directly replaced by AI.
- New-grad hiring at major tech companies is down by roughly 65% versus 2019, while early-stage startups show a decline of about 76%.
- Employment for workers aged 22-25 in AI-exposed occupations is about 19% below trend, while experienced workers in the same occupations show no comparable decline.
- Senior software roles are taking a larger share of the market: 71% of the rebound in software-development postings from May 2025 to May 2026 came from senior positions.
- AI coding tools do not produce uniform productivity gains. Studies show strong improvements on routine or greenfield work, but near-zero or negative effects on complex, mature codebases in some cases.
- More AI-generated code does not automatically mean more productive engineering: independent telemetry measured 26.9% of merged production code as AI-generated while overall productivity gains remained much smaller.
- Code quality and verification remain major constraints, with 45% of AI-generated code in one security study introducing an OWASP Top 10 vulnerability.
- AI skills command a meaningful wage premium, with US postings requiring AI skills paying about 28% more on average and roles requiring multiple AI skills carrying an even higher premium.
- Long-term forecasts point toward job transformation rather than wholesale elimination: the WEF projects a net gain of 78 million jobs globally by 2030 and continues to rank software developers among the fastest-growing occupations.
When this applies
This applies when a company is planning engineering headcount, deciding how much to invest in AI-assisted development, budgeting for junior versus senior talent, or evaluating how AI may change its hiring strategy over the next several years. It is especially relevant for CTOs, engineering leaders, recruiters, founders, and workforce planners who need to distinguish between actual layoffs, reduced entry-level hiring, productivity effects, and wage premiums. It also applies to software engineers assessing which skills remain valuable as routine implementation work becomes increasingly automated.
When this does not apply
This does not apply as proof that AI has already replaced a fixed percentage of software engineers or that all software jobs face the same level of risk. Layoff announcements, payroll data, job postings, productivity trials, and executive statements measure different things and should not be combined into one universal displacement figure. It is also not a reliable basis for predicting exactly how many jobs will disappear by 2030, since the article explicitly separates measured data from forecasts and treats round-number predictions as scenarios rather than established facts.
Checklist
- Separate AI-attributed layoffs from broader technology-sector layoffs when analyzing workforce reductions.
- Track entry-level, mid-level, and senior hiring separately instead of relying on total software-job counts.
- Compare new-grad hiring with historical baselines rather than only with the previous year.
- Use official employment projections alongside private job-posting datasets.
- Distinguish reduced hiring from actual layoffs when assessing AI displacement.
- Measure AI productivity separately for greenfield, routine, legacy, and high-complexity work.
- Compare perceived productivity gains with observed delivery metrics.
- Track time-to-pull-request together with code-review and rework time.
- Measure production defects and security findings in AI-generated changes.
- Require senior review for AI-generated code that reaches production.
- Track the percentage of AI-generated code separately from actual business or delivery outcomes.
- Benchmark salaries for AI-skilled engineers against comparable non-AI roles at the same seniority.
- Treat executive forecasts about future automation as scenarios rather than current measurements.
- Invest in architecture, review, security, and production-accountability skills that remain difficult to automate.
- Revisit workforce assumptions regularly because AI adoption, job postings, compensation, and productivity evidence are changing quickly.
Common pitfalls
- Treating every technology-sector layoff as an AI-caused job loss.
- Using announced layoffs as a direct count of jobs permanently replaced by automation.
- Looking only at total software hiring and missing the divergence between junior and senior roles.
- Assuming reduced new-grad hiring means software engineering as a profession is disappearing.
- Believing AI productivity claims without checking whether the study involved greenfield tasks or mature production codebases.
- Measuring only coding speed while ignoring review time, debugging, security defects, and code churn.
- Treating a larger percentage of AI-generated code as equivalent to higher engineering productivity.
- Assuming junior engineers gain the same long-term value from AI tools as experienced engineers with architecture and review responsibilities.
- Comparing AI salary premiums without controlling for seniority, geography, or specialization.
- Presenting forecasts about 90% or 95% AI-written code as measured industry-wide reality.
The short answer: AI is not eliminating software engineering. It is splitting the market in two. Entry-level tech jobs are shrinking fast. New-grad hiring at the largest tech companies is down more than 50% since 2019 (SignalFire). Demand for senior engineers holds or grows: 71% of the 2025-2026 rebound in software job postings came from senior roles (Indeed Hiring Lab). The U.S. Bureau of Labor Statistics still projects 15% growth for software developers from 2024 to 2034. This page collects every number you need, with sources and dates. You are free to cite any statistic with a link to this page.
Top 10 AI job displacement statistics for 2026
- AI was cited in 87,714 announced U.S. job cuts in 2026 through May – 22% of all layoffs, and already more than in all of 2025 (Challenger, Gray & Christmas).
- AI was the #1 stated reason for U.S. layoffs for five consecutive months, March-July 2026 (Challenger).
- Employment of software developers is projected to grow 15% from 2024 to 2034, much faster than average (U.S. BLS).
- New-grad hiring at the largest tech companies is down more than 50% since 2019; at early-stage startups it is down about 76% (SignalFire, 2026).
- Employment for workers aged 22-25 in AI-exposed jobs sits about 19% below trend, while experienced workers show no comparable decline (Stanford Digital Economy Lab, Aug 2026).
- Senior job postings rose 13.5% from January 2025 while entry-level postings fell 6.3% (Indeed Hiring Lab, July 2026).
- Experienced developers were 19% slower with AI tools in METR’s 2025 randomized trial – while believing they were 20% faster (METR).
- Microsoft and Google executives report 20-30% of their code is now AI-generated; independent telemetry measures 26.9% of merged production code (see Section 6).
- 45% of AI-generated code introduced an OWASP Top 10 security vulnerability (Veracode, 2025).
- Job postings that require AI skills pay 28% more – about $18,000 per year (Lightcast); PwC’s broader estimate is a 62% premium (PwC, 2026).
AI layoffs in 2026: how many jobs has AI replaced?
Direct answer: Companies attributed 54,836 U.S. job cuts to AI in 2025. In 2026, AI-attributed cuts reached 87,714 by the end of May – 22% of all layoffs – and AI led all layoff reasons for five straight months. But hiring also rose about 25% year over year. AI is shifting the labor market. It is not dismantling it.
Challenger, Gray & Christmas tracks the stated reasons for U.S. layoff announcements. AI became a tracked reason in May 2023. The numbers move fast, so every figure below carries its report date.
- Full year 2025: AI was cited in 54,836 announced job cuts – 5% of all cuts (Challenger, Jan 2026).
- Cumulative 2023-2025: 71,825 AI-attributed cuts (Challenger, Jan 2026).
- 2026 monthly sequence: January 7,624 · February 4,680 · March 15,341 · April 21,490 · through May, 87,714 year-to-date – 22% of all 2026 layoffs, already above the full-year 2025 total (Challenger monthly reports).
- AI was the leading layoff reason for the fifth consecutive month in July 2026, with 10,970 cuts (Challenger, Aug 2026).
- The technology sector announced 154,445 cuts in 2025 (+15% vs. 2024) and 149,023 cuts through July 2026 (+67% year over year; 31% of all U.S. cuts).
- The independent tracker layoffs.fyi counts ~126,814 tech employees laid off across 278 companies in 2026 to date, led by Oracle (30,000), Amazon (16,100), Dell (11,000), and Meta (10,400) (layoffs.fyi, mid-August capture; not all cuts are AI-attributed).
Context that most pages skip. Engineers made up less than 30% of layoffs even at companies that cut deeply (SignalFire, 2026). Andy Challenger notes that hiring rose about 25% over the prior year, so AI is reshaping headcount rather than dismantling the labor market. And one widely quoted forecast – Anthropic CEO Dario Amodei’s May 2025 prediction that AI could remove about half of entry-level white-collar jobs within five years – remains a prediction, publicly disputed by other industry leaders (Axios).
Software developer hiring and job market trends
Direct answer: Is there a software engineer shortage? Yes and no. There is a surplus of entry-level candidates and a persistent shortage of senior engineers. Total tech postings remain about 30% below pre-pandemic levels, but software-development postings rose ~15% after agentic coding tools launched in early 2025 – and 71% of that rebound was senior roles.
The baseline: official projections
- The U.S. BLS projects 15% employment growth for software developers, QA analysts, and testers from 2024 to 2034 – “much faster than the average” (BLS Occupational Outlook Handbook).
- The BLS projects about 129,200 openings per year for the group over the decade, from a 2024 base of ~1.7 million software developers with median pay above $130,000.
The postings data
- U.S. tech job postings were 36% below early-2020 levels as of July 2025 (Indeed Hiring Lab).
- About half of the decline happened before ChatGPT launched – post-2022 over-hiring, higher rates, and budget cuts explain much of it (Indeed Hiring Lab).
- Software-development postings rose ~15% after Claude Code launched in late February 2025, while overall postings fell 7% in the same window (Indeed Hiring Lab, July 2026).
- 37% of new software postings mention AI in the job title (Indeed Hiring Lab, 2026).
Entry-level tech jobs and new-grad unemployment
- Unemployment for recent computer science graduates: 6.1% (2023 data) rising to 7.0% (2024 data). Computer engineering: 7.5% → 7.8% – the second-highest of 73 majors (Federal Reserve Bank of New York). Note: small samples make single-year figures noisy.
- About 43% of recent graduates are underemployed – working jobs that do not require a degree (NY Fed, end of 2025).
- In tech and engineering, fewer than 2% of postings were junior roles as of August 2025 (Indeed Hiring Lab).
The international picture
- South Korea: the Bank of Korea reports that 94% of net youth job losses (268,000 of 285,000, June 2022 – June 2026) occurred in high-AI-exposure sectors. IT services fell 31.4% and computer programming fell 16.6% for workers aged 15-29, while workers in their 50s gained 230,000 jobs (reported Aug 2026).
- United Kingdom: entry-level tech roles fell 46% in 2024, with projections of a 53% decline by the end of 2026 (Adzuna data, via Stack Overflow).
Entry-level tech jobs vs. senior roles: the great divergence
Direct answer: Three independent datasets – LinkedIn-derived hiring data, ADP payroll records, and job postings – all show the same pattern. AI pressure lands on entry-level tech jobs through reduced hiring, not layoffs. Senior engineers are protected and increasingly in demand.
- New grads were 7% of big-tech hires in 2024, down 25% from 2023 and more than 50% from 2019 (SignalFire, 2025).
- By 2026, new-grad hiring was down ~65% at tech majors and ~76% at early-stage startups versus 2019 (SignalFire, 2026).
- At tech majors, overall hiring fell 25% versus 2019, but engineering hiring fell only 11%. Engineers now make up 55% of tech-major hiring, up from 46% in 2019 (SignalFire, 2026).
- Stanford’s analysis of ADP payroll data finds employment for workers aged 22-25 in AI-exposed occupations is about 19% below trend (August 2026 update; the gap was 13% in the original August 2025 paper and has widened in each revision). Experienced workers in the same occupations show no comparable decline (Stanford Digital Economy Lab).
- The same research states plainly: there is no widespread, economy-wide job displacement from AI. The mechanism is reduced hiring, not increased layoffs.
- Senior postings rose 13.5% from January 2025; mid-level fell 6.7%; entry-level fell 6.3%. Software development has the highest senior share of any field: 69.3% of postings (Indeed Hiring Lab, July 2026).
- 71% of the increase in software-development postings from May 2025 to May 2026 came from senior roles (Indeed Hiring Lab).
- Senior engineers capture about 5x the AI productivity gains of junior engineers (Opsera 2026 benchmark, 250,000+ developers).
Why the split? AI automates codified, routine tasks first. Those tasks defined most junior workloads. Senior engineers supply what AI cannot: architecture judgment, code review, and accountability for production systems. This is why companies that scale engineering with AI now hire senior Python developers instead of large junior teams.
AI coding statistics: what productivity studies actually show
Direct answer: Adoption is near-universal – 90% of developers use AI. Measured productivity is contested. Controlled studies show gains of 26-56% on routine and greenfield work, near-zero to negative results on complex, mature codebases, and a large gap between how fast developers feel and how fast they are.
Adoption and reliance
- 90% of software professionals use AI, up 14 points in one year; the median developer spends two hours a day with AI tools (Google DORA 2025, ~5,000 respondents).
- 84% of developers use or plan to use AI tools, but favorability fell from over 70% to 60%, and trust in AI accuracy fell to 29% (Stack Overflow 2025, 49,000+ respondents).
- 66% of developers spend more time fixing “almost-right” AI code; experienced developers are the most skeptical group (Stack Overflow, 2025).
The measured effects
- GitHub’s randomized trial: developers with Copilot finished a greenfield task 55.8% faster (95% CI: 21-89%; n=95). Less-experienced developers gained the most (Peng et al., 2023).
- Field experiments at Microsoft, Accenture, and a Fortune 100 firm (4,000+ developers, run by Microsoft, MIT, Princeton, and Wharton researchers): Copilot raised output by 26% on average (Cui et al., 2024).
- McKinsey research finds AI saves roughly 35-45% of time on routine coding tasks but under 10% on high-complexity work.
- METR’s randomized trial: experienced open-source developers were 19% slower with AI tools on mature codebases they knew well – while forecasting a 24% speedup and self-reporting a 20% speedup afterward. That is a 39-point perception gap (METR, July 2025; 16 developers, 246 real tasks).
- METR’s February 2026 follow-up points positive but is statistically inconclusive: returning developers showed an estimated 18% speedup with a confidence interval of -38% to +9%, and newly recruited developers showed ~4% (CI -15% to +9%). METR calls this “very weak evidence” due to selection effects and is redesigning the study (METR, Feb 2026).
- Opsera’s benchmark of 250,000+ developers: AI cuts time-to-pull-request by up to 58%, but AI-generated PRs wait 4.6x longer in code review (Opsera, 2026).
- An enterprise survey found 43% of AI-generated code changes need debugging in production, and developers now spend ~38% of the work week on verification and debugging (Global Surveyz via VentureBeat, 2026).
- DORA’s framing is the most useful summary: AI is a “mirror and multiplier” – it amplifies the strengths and weaknesses a team already has (Google DORA 2025).
For tool-level adoption and trust data – which assistants developers use and how much they trust them – see our companion page: AI coding assistant statistics.
How much code is written by AI? Vibe coding statistics
Direct answer: Executives report 20-30% of code at Microsoft and over 30% of new code at Google is AI-generated. Independent telemetry measures 26.9% of merged production code. “Vibe coding” – prompting AI to generate most of an application – already describes a quarter of Y Combinator’s Winter 2025 startups, whose codebases were ~95% AI-generated.
- Satya Nadella: 20-30% of Microsoft’s code is written by AI, with the strongest results in Python (April 2025). Self-reported; no published methodology.
- Sundar Pichai: more than 30% of Google’s new code is AI-generated, up from ~25% six months earlier (2025). Self-reported.
- Microsoft CTO Kevin Scott predicts 95% of code will be AI-generated by 2030. A prediction, not a measurement.
- 25% of Y Combinator’s Winter 2025 startups had codebases that were ~95% AI-generated (TechCrunch, March 2025).
- Telemetry across 4.2 million developers: AI-generated code rose from 22% to 26.9% of merged production code between November 2025 and February 2026 – while measured productivity gains stayed near +10% (LogRocket analysis, March 2026). More code is not the same as more output.
Code quality is the counterweight
- Copy-pasted code rose from 8.3% (2020) to 12.3% (2024) of changed lines. Refactored (“moved”) code fell from ~24% to 9.5%. 2024 was the first year copy-paste exceeded refactoring (GitClear, 211M lines analyzed).
- By mid-2026 the gap widened: copy-paste reached 15.7% and moved code fell to 3.8%. Code churn runs at roughly double the pre-AI baseline (GitClear, 2026).
- 45% of AI-generated code introduced an OWASP Top 10 vulnerability. Java failed security checks 72% of the time; cross-site scripting defenses failed 86% of the time. Security pass rates stay flat near 55% even as syntax correctness climbs to ~95% (Veracode, 100+ LLMs tested).
AI engineer salary statistics and skill premiums
Direct answer: AI skills raise developer pay. The size of the premium depends on the method: 12% at mid-level in Europe, 28% across U.S. postings, and up to 62% in PwC’s raw cross-role average. Average AI/ML engineer total compensation reached $242,507 in 2026.
- Postings that require AI skills pay 28% more (~$18,000/year); postings requiring two or more AI skills carry a 43% premium (Lightcast, 1.3B postings analyzed).
- PwC’s 2026 Global AI Jobs Barometer reports a 62% wage premium, up from 56% the prior year. PwC notes this is a raw average and “may not imply a causal relationship” (PwC).
- European benchmark: a ~12% AI premium at mid-level across 1,600+ companies (Ravio); tiered U.S. estimates run 6.2% (entry) to 18.7% (staff).
- Average AI/ML engineer total compensation: $242,507 (Levels.fyi, 2026); base salary range $134,000-$193,250 (Robert Half).
- 51% of AI-skill postings are now outside IT – AI skills spread beyond tech itself (Lightcast).
- Baseline: software developer median pay exceeds $130,000 (BLS, 2024); early-career CS and computer engineering majors earn the highest starting wages of all majors at $87,000-$90,000 (NY Fed).
The premium data explains a hiring pattern we see directly: companies pay for engineers who can build with AI, not for engineers displaced by it. If that is your gap, you can hire senior AI/ML engineers vetted for production work.
The future of software engineering: forecasts to 2030
Direct answer: Major forecasts agree on transformation, not elimination. The WEF projects a net gain of 78 million jobs globally by 2030 and ranks software developers among the top five growing roles. BCG expects 50-55% of U.S. jobs to be reshaped within three years, with only 10-15% potentially eliminated over five or more.
- WEF Future of Jobs 2025: 170 million jobs created, 92 million displaced – net +78 million by 2030. Software and application developers rank in the top five fastest-growing roles; AI/ML specialist roles grow 82% (WEF).
- 86% of employers expect AI to transform their business by 2030; 39% of workers’ core skills will change (WEF).
- BCG (2026): 50-55% of U.S. jobs will be reshaped by AI within two to three years; 10-15% may be eliminated over five-plus years (BCG).
- Anthropic’s Economic Index: coding remains the dominant AI use case – computer and mathematical tasks make up about one-third of Claude.ai conversations and roughly half of enterprise API traffic. Agentic coding runs 79% automation vs. 21% augmentation, a signal of where workflows move next (Anthropic).
- Predictions vs. reality check: in early 2025 Dario Amodei predicted AI would write ~90% of code within six months. Inside frontier labs that came close; across the industry the measured range stayed near 25-40% (Forbes, Feb 2026). Treat round-number predictions as scenarios, not data.
Methodology and sources
This page prioritizes primary sources: government statistics (BLS, NY Fed), peer-tracked studies (METR, arXiv), large-sample industry research (DORA, Stack Overflow, SignalFire, Stanford Digital Economy Lab), and named vendor telemetry (GitClear, Veracode, Opsera). Every statistic carries its publication date, because several figures are moving totals. Executive claims about the share of AI-written code are labeled self-reported. Where credible sources conflict, we show both figures rather than pick one.
We also exclude numbers that fail verification. Two examples you will see elsewhere: the claim that entry-level developer hiring fell “73%” and the claim that “14% of workers have already been displaced by AI.” Neither traces to a primary source with a published methodology, so neither appears in our dataset. If you find an error on this page, tell us and we will correct it and log the change.
How to cite this page
You may reuse any statistic or chart on this page with attribution. Suggested citation:
Uvik Software. “AI Job Displacement Statistics 2026: Layoffs, Hiring & Productivity Data for Software Engineering.” Uvik.net, August 2026, https://uvik.net/blog/ai-job-displacement-statistics/.
Uvik Software provides senior, AI-augmented Python and AI/ML engineers to teams in the US and Europe. If the data above matches what you see in your own hiring, talk to us about senior Python developers or AI/ML engineers.
FAQ: AI and software engineering jobs
Will AI replace software engineers?
Not at the occupation level. The BLS projects 15% growth for software developers from 2024 to 2034, and the WEF ranks the role among its top five growing jobs through 2030. The measurable pressure sits at entry level, not across the profession. We analyze the full argument in <a href="https://uvik.net/blog/will-artificial-intelligence-replace-programmers/">Will AI replace programmers?</a>
How many jobs has AI replaced?
Challenger, Gray & Christmas attributed 54,836 U.S. job cuts to AI in 2025 and 87,714 in 2026 through May - 22% of all layoffs. AI led all layoff reasons for five straight months (March-July 2026). The same firm reports hiring rose about 25% year over year.
Is there a software engineer shortage or a surplus?
Both. Entry level shows a surplus: recent CS-grad unemployment reached 6.1-7.0%, and big-tech new-grad hiring fell more than 50% versus 2019. Senior level shows a shortage: 71% of the 2025-2026 rebound in software postings came from senior roles.
Does AI make developers more productive?
The evidence is split. A GitHub trial found a 55.8% speedup on a greenfield task, and a 4,000-developer study found 26% average gains. METR’s 2025 trial found experienced developers were 19% slower on mature codebases, and its 2026 follow-up was statistically inconclusive. Gains concentrate on routine work.
How much code is written by AI?
Executives report 20-30% at Microsoft and over 30% of new code at Google (self-reported). Independent telemetry across 4.2 million developers measures 26.9% of merged production code as of February 2026.
Is AI-generated code safe?
Often not without senior review. Veracode found 45% of AI-generated code introduced an OWASP Top 10 vulnerability (72% for Java). GitClear found copy-pasted code now exceeds refactored code, and churn runs at double the pre-AI baseline.
Do AI skills increase developer salaries?
Yes. Lightcast measures a 28% premium (~$18,000/year) in postings that require AI skills. PwC reports 62% as a raw cross-role average. European mid-level data shows a more conservative 12%.
Are junior or senior developers more affected by AI?
Junior developers, decisively. Stanford’s payroll research puts employment for ages 22-25 in AI-exposed jobs about 19% below trend, while experienced workers show no comparable decline. Senior engineers also capture about 5x the AI productivity gains of juniors (Opsera).
