Careeronaut cross-referenced five independent data sources to size the tech-sector contraction since 2022. The Bureau of Labor Statistics' Information sector — the federal government's official tech-economy metric — peaked at 3.115 million payroll jobs in November 2022. By April 2026 it had fallen to 2.773 million: an 11.0% decline, or 342,000 net jobs lost. Total US nonfarm employment over the same window hit an all-time record of 158.7 million in April 2026.
Layoffs.fyi's company-level tracker counts 645,094 individual tech layoffs across 2022 through 2025 — meaningfully larger than the BLS net figure, which captures the difference between gross layoffs and gross hires. Challenger, Gray & Christmas, which publishes the standard monthly US job-cut report, recorded 85,411 announced tech cuts in the first four months of 2026 alone — up 33% year-over-year, while total US announced cuts fell 50% year-over-year. Tech is the only major US industry where layoffs are growing in 2026.
The AI attribution has hardened: Challenger explicitly cites AI as the stated reason for 16% of all 2026 year-to-date job cuts, rising to 26% in April alone. Anthropic's Economic Index measures the share of an occupation's tasks performed by AI in real Claude usage; computer programmers top the index at 74.5% observed exposure — the highest of any occupation tracked. Indeed's software-developer job-postings index now sits 33% below its February 2020 baseline; demand for the most-AI-exposed white-collar job category has not recovered to pre-pandemic levels in five years.
Key numbers
Net tech jobs lost: BLS Information sector since Nov 2022 peak
−342,000
Down 11.0% from 3.115M to 2.773M — the federal government's official tech-sector employment metric
Total US nonfarm employment, April 2026
158.7M
All-time record — the rest of the economy is at peak even as tech contracts
Layoffs.fyi cumulative tracked tech layoffs, 2022–2025
645,094
Sum of 2022 (103,571), 2023 (264,320), 2024 (152,922) and 2025 (124,281) annual totals
Tech layoffs YTD 2026 vs YTD 2025 (Challenger)
+33%
85,411 tech cuts YTD — tech is the ONLY major US industry where 2026 layoffs are up year-over-year while overall US cuts are down 50%
Share of April 2026 US layoffs Challenger attributes to AI
26%
Up from 16% YTD. AI attribution is rising month-over-month in the Challenger reports
Computer programmer AI task exposure (Anthropic Economic Index)
74.5%
The highest observed exposure of any occupation tracked by Anthropic's Claude-usage data
Indeed Software Developer job-postings index vs Feb 2020 baseline
−33%
Software-developer demand still hasn't recovered to pre-pandemic levels 5+ years later
Total tech employment vs total nonfarm, indexed to Jan 2020
Information: ~95 · Total US: 105
Tech is below 2020 levels. The rest of the economy is well above. The 10-point gap is what 'bifurcation' looks like in one number.
The tech recession the rest of the economy doesn't see
Information sector vs total US nonfarm employment, indexed to January 2020 = 100. Source: BLS Current Employment Statistics series CES5000000001 and CES0000000001.
What this means
For tech workers: this is the official, government-data confirmation that 'the tech downturn' is not an artifact of social-media doomscrolling. The Information sector is 11% smaller than it was 41 months ago — a decline that, in total-economy terms, would be a full-blown recession.
For everyone else: this is the most concrete evidence yet that the US economy is bifurcating. Total payrolls are at an all-time high. Tech payrolls are below 2022 levels. Aggregate job-market headlines mask the divergence.
For journalists: the safest claims to make from this dataset are (1) the BLS net figure (-342,000 since the Nov 2022 peak) and (2) Challenger's announced-cuts numbers (85,411 tech, +33% YoY through April 2026). Layoffs.fyi and TrueUp are useful for company-level color but their cumulative totals are not directly comparable to BLS net (they capture gross layoffs, not net job loss). Be careful to keep gross vs. net straight.
For employers in tech: the talent-leverage equation has fully flipped. The 2021–22 'Great Resignation' narrative is dead in your industry — quit rates are at decade-lows (1.1% in March 2026 per BLS JOLTS) while AI displacement claims are rising. Engagement and re-skilling, not retention bonuses, are the larger problems.
“Every careers piece in 2026 cites tech layoff numbers from a different tracker and treats them as comparable. They aren't. Layoffs.fyi counts individual layoffs at private companies. The BLS measures sector-level payroll. Challenger counts what employers announce. Pulling all three together is the only way to size what's actually happening — and the picture they paint together is sharper than any single number on its own.”
Methodology
Careeronaut combined five independent, publicly accessible datasets, refreshed live from primary sources.
BLS Current Employment Statistics: monthly seasonally-adjusted nonfarm payroll counts for the Information sector (series CES5000000001) and total US nonfarm employment (series CES0000000001), pulled live via the BLS Public Data API. The Information sector covers publishing, telecommunications, data processing, broadcasting, software publishing, web search portals, and motion-picture industries. We intentionally use the BLS Information sector — not the broader 'tech' definition used by Layoffs.fyi — because it is the government's official metric and is auditable line-by-line on the BLS news release.
Layoffs.fyi: company-level layoff tracker maintained by Roger Lee since March 2020. We use the annual totals as reported on the public dashboard: 2022 (103,571), 2023 (264,320), 2024 (152,922), 2025 (124,281). Cumulative 2022–2025 total: 645,094.
Challenger, Gray & Christmas: monthly US job-cut report. We use the April 2026 release figures: 83,387 cuts in April, 300,749 year-to-date (down 50% year-over-year), with 85,411 cuts attributed to the technology sector year-to-date (up 33% year-over-year). AI is cited as the reason for 16% of 2026 year-to-date cuts and 26% in April alone.
Anthropic Economic Index (March 2026 release): per-occupation observed AI task exposure scores derived from anonymized Claude usage. Computer programmers top the list at 74.5%. Dataset published openly on Hugging Face.
Indeed Hiring Lab / FRED IHLIDXUSTPSOFTDEVE: Software Developer job-postings index, indexed to February 2020 = 100. Latest read: -33% versus baseline.
Caveats: (1) BLS Information sector is narrower than 'tech' as commonly used in business press; software developers at non-tech companies sit in Professional and Business Services. (2) BLS reports NET employment change. Layoffs.fyi and Challenger report GROSS layoffs. They are not directly comparable; the BLS net loss (-342,000) is smaller than the Layoffs.fyi cumulative (-645,094) because BLS counts new hires that offset layoffs. (3) Challenger tracks ANNOUNCED cuts, not executions; announcement-to-execution timing varies. (4) AI attribution in Challenger reports is based on the reason cited in the announcement, which can be a chosen narrative rather than the only cause. (5) Anthropic's exposure scores are 'observed in Claude usage' — they reflect what Claude is being used to do, not what AI generally is replacing. (6) Indeed postings ≠ hires. All five sources are openly linked in the campaign body so any number is verifiable in under five minutes by anyone with a browser.
For journalists
The Careeronaut team is happy to share the full methodology, raw data, or run a custom slice of the data for your story.
Email press@careeronaut.comAbout Careeronaut