AI Layoff Tracker

The short version. Verified job cuts are cuts with a filing or named source behind them, counted on the day they take effect; the main figure. AI-attributed is the subset where the employer named AI in words we can quote. Announced is a separate, labeled tier of announcement-stage plans, never mixed into the verified total. Nothing is estimated; every number links to a legal filing or named report, and country/US-state filters describe where the jobs were, not an employer’s headquarters.

How the AI tag works, in one line: only a primary or contributing cause counts, each with an exact supporting quote; investment in AI, future projections, or AI used to pick who goes do not qualify. A separate broader measure is labeled and never merged in.

Read the full methodology and sources → · extraction, dedup, coverage limits, why our totals differ, and API access, all in detail.

Official government filings and notices are collected directly (SEC EDGAR incl. Item 2.05 exit-cost filings, WARN notices from 48 US states and DC, Eurofound ERM for the EU), press-release wires and reviewed company IR feeds are monitored, and 705 reviewed news outlets across 203 countries surface coverage through GDELT’s 65-language index and Google News, allowlist-only, never crawled directly. Every published event links to its source. For the handful of US states that publish no usable WARN register (Arkansas, Wyoming, New Hampshire, Missouri, Hawaii, Oklahoma), we also show their official monthly BLS unemployment rate as a clearly separate context metric, sourced and dated, never mixed into the layoff counts. See the full source directory.

The region tabs are views over the worldwide data. The full country list for each tab loads here.

Every tracker measures a different thing, so the numbers should differ. We count verified events: cuts with a filing or named-outlet source behind them, each one clickable. The big announcement trackers count corporate intentions. Neither is wrong; they answer different questions. Our total sits below the headline announcement estimates, and the gap is fully explainable, here is exactly why, and why we treat it as a feature, not a shortfall.

1 · They book multi-year plans on day one; we count cuts as they happen. When a company announces “20,000 cuts over two years,” the announcement trackers record all 20,000 that day. We add each cut as its WARN notice or SEC filing actually appears. Over a year that is a large, permanent gap, their figure is a forecast, ours is an execution ledger.

2 · They include separations that name no event. Announcement totals fold in voluntary buyouts, deferred resignations, and attrition programs, including large federal-workforce reductions that file no WARN notice and name no company. In 2025 that was roughly 250,000–300,000 jobs of the announcement total alone. There is no document or named source to link, so we do not claim it.

3 · They count cuts no outlet ever named. Announcement surveys aggregate press mentions and estimates we cannot reproduce. We only publish what traces to a source, so an unsourced cut never enters our total.

4 · We date each cut by when it takes effect, not when it was filed. Most trackers count a layoff on the day its WARN notice is filed or the cut is announced. We count it on the day the jobs actually end, because that is what a worker lives through and what a labor-market reader wants to measure. The two bases answer different questions: filing date asks “when did we hear about it,” effective date asks “when did it happen.” The gap is small and can fall either way; in 2026 the effective-date basis runs slightly higher, because more notices were filed in 2025 for cuts that land in 2026 than were filed in 2026 for cuts landing in 2027. We store both dates, so any figure here can be recounted on either basis.

The bottom line, stated plainly. Our verified figure is a documented floor, smaller than the estimates, but every single number clicks through to a legal filing or a named report. We deliberately do not pad it to match a headline estimate, because a number a journalist can verify is worth more than a bigger one they cannot. And on the measure this tracker exists for, layoffs companies attribute to AI, our count actually exceeds the headline announcement trackers every year, because we surface AI attributions from primary sources they never itemize.

Where we lead, and where we don’t, stated honestly. Because our figure is built from receipts, it is not always smaller. Measured like-for-like against the public trackers by category:

• Against WARN-only aggregators (the legally-filed US floor), we come out higher: we import the same WARN notices and add SEC filings and named-news reports on top, so our US verified total clears the WARN floor rather than stopping at it.
• Against tech-event trackers, our worldwide technology job-cut total is at or above the largest of them by volume; we carry fewer tiny private-startup events, but more total tech job losses, because we catch the big filed cuts they sometimes miss.
• On AI attribution, our broad AI-linked measure is at or above the scale of the announcement surveys’ AI figure, and every entry carries the employer’s own words. Our strict measure is deliberately smaller than a survey, because a survey infers a reason and we require a quote.
• Against announcement surveys’ all-industry total, we run lower, on purpose: the gap is receiptless cuts (federal-workforce reductions, buyouts, attrition, and small closings that file nothing). We do not claim what we cannot source.
• On coverage no one else offers, we are the only one of these measuring the global, all-industry, source-linked universe. Everyone else is US-only, tech-only, or a survey.

Where each kind of tracker fits:

• Announcement surveys, monthly totals of announced US cuts from press reports and company statements, including estimates and multi-year plans. Typically published as press releases; no per-event public database.
• Editorial newsroom trackers, selected major announcements with newsroom verification. No downloadable dataset; selective by design.
• Sector trackers, technology-focused trackers built from announcements and crowdsourced reports. Their scope matches our Technology filter, not our all-industry total.
• Official statistics, US BLS JOLTS, UK ONS, Eurostat count all separations economy-wide (millions/month) with no company detail. A different universe entirely.
• This tracker, verified events in the headline, announcement-stage figures in a separate labeled tier, corrections logged openly, data and code public. When our number differs, the difference is definitional, and both definitions are stated here so either can be cited correctly.

The full directory of every pipeline, the SEC, all state WARN registries with live links, Eurofound ERM, and the news index, lives on the Data Sources page. The disclosures below cover what is not yet included.

Why the country count grows over time. The number of countries is not a setting we can raise; it reflects where large, press-covered layoffs have actually happened in our window. GDELT already searches every country on earth in 65+ languages, so a country appears the moment a credible outlet there covers a qualifying layoff. As events occur and as we add more trusted local outlets, the count rises on its own. This is honest by design: we show the countries where verifiable events exist, not a padded list.

Known gaps, stated plainly. We do not yet operate direct connectors for Canada SEDAR+, UK RNS, ASX, TDnet/EDINET, NSE/BSE, HKEXnews, SGXNet, SENS, DART or TASE; they are maintained as official-source research candidates and will be named as live only after a stable public interface, tests and source-health monitoring exist. A few countries also publish official per-company redundancy records we do not ingest yet, including Belgium’s FPS Employment collective-dismissal reports, Italy’s weekly CIGS decree lists, and Sweden’s varsel statistics. Most countries, including Germany and Mexico, treat employer identity in redundancy filings as confidential, so press coverage through GDELT in local languages is the primary source there. Events too small for any press coverage, any WARN threshold, or the ERM threshold of 100 jobs will not appear in any tracker, including this one.

Why this log is a feature, not a warning. Many early entries are the pipeline catching and fixing its own extraction bugs during pre-launch QA, which is the point: a system that finds and discloses its own errors is more trustworthy than one that hides them.

Errors are corrected openly, not silently. Every correction to published figures is dated and described here, newest first, and corrected rows are also flagged edited: true in the API. The list scrolls, because it grows a little every day as the data self-corrects.

For reproducible monitoring, the machine-readable quality status endpoint reports dataset revision, recent corrections, collector health, retained-source integrity and the status of each coverage workstream. Pending work is shown as pending, not silently treated as coverage.

• 2026-07-28: 401 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-28: 5 entries enriched. Automated role-category extraction from stored source evidence
• 2026-07-28: 3 entries removed. Undated duplicate cleanup: news/SEC rows with no date that duplicate a dated event of the same company and headcount (they had bypassed the date-gated dedup guard).
• 2026-07-28: 73 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-27: 3 entries corrected (fields: country, industry). Automated classification audit: label mismatch flagged and independently confirmed by two LLM passes (DeepSeek)
• 2026-07-27: 1 entry merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-27: 400 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-27: 6 entries enriched. Automated role-category extraction from stored source evidence
• 2026-07-27: 1 entry reclassified. Automated source-evidence reassessment
• 2026-07-27: 54 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-26: 2 entries removed. 2025 historical-year backfill honesty correction: phantom/double-count rows (Meta 8k = lawsuit allegation in a Fortune macro roundup; US-federal 5k = Benzinga commentary on Nov BLS drop; DOGE 10k = cross-agency aggregate double-counting VA/NOAA/IRS). -23,000 jobs. Per TECHLOG backfill #1.
• 2026-07-26: 6 entries removed. Undated duplicate cleanup: news/SEC rows with no date that duplicate a dated event of the same company and headcount (they had bypassed the date-gated dedup guard).
• 2026-07-26: 23 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-26: 407 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-26: 12 entries enriched. Automated role-category extraction from stored source evidence
• 2026-07-26: 134 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-25: 6 entries corrected (fields: industry, country). Automated classification audit: label mismatch flagged and independently confirmed by two LLM passes (DeepSeek)
• 2026-07-25: 5 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-25: 9 entries enriched. Automated role-category extraction from stored source evidence
• 2026-07-25: 390 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-25: 1 entry removed. ghost row: wiped meta (empty company + job_count 0) inflating entry/distinct counts
• 2026-07-24: 4 entries removed. FL WARN W-1511 test/sandbox notice: one WARN# spans AT&T+BOEING+BOEING test, fabricated counts (78788 fake)
• 2026-07-24: 2 entries removed. Undated duplicate cleanup: news/SEC rows with no date that duplicate a dated event of the same company and headcount (they had bypassed the date-gated dedup guard).
• 2026-07-24: 9 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-24: 400 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-24: 6 entries enriched. Automated role-category extraction from stored source evidence
• 2026-07-24: 18226 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-23: 41 entries corrected (fields: country). Legacy repair: country filled from deterministic evidence (a US state code on the row, or a US state government WARN source). Rows without such evidence were left blank rather than guessed.
• 2026-07-23: 54 entries corrected (fields: company). Legacy repair: the notice's site address was stored inside the employer name, which split the company from its own other rows. Name normalized; no count, date or source changed.
• 2026-07-23: 1 entry corrected (fields: industry). Automated classification audit: label mismatch flagged and independently confirmed by two LLM passes (DeepSeek)
• 2026-07-23: 17 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-23: 4 entries removed. Undated duplicate cleanup: news/SEC rows with no date that duplicate a dated event of the same company and headcount (they had bypassed the date-gated dedup guard).
• 2026-07-23: 23 entries removed. Audit 2026-07-23: WARN notices the state later RESCINDED or CANCELLED. A withdrawn notice is a layoff that did not happen, so counting it added 5,050 jobs nobody lost. The importer now drops these at the boundary; these 23 stored rows are removed and their hashes suppressed.
• 2026-07-23: 1 entry removed. Audit #1 (2026-07-23): the cited Times of India source states Starbucks has not confirmed how many jobs will be lost; the ~1,000 figure in the stored excerpt is an earlier Seattle/Kent retail event already covered by WA WARN rows. Count unsupported by its own source.
• 2026-07-23: 400 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-23: 21549 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-22: 22 entries removed. Undated duplicate cleanup: news/SEC rows with no date that duplicate a dated event of the same company and headcount (they had bypassed the date-gated dedup guard).
• 2026-07-22: 4 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-22: 309 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-22: 400 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-22: 1 entry enriched. Automated role-category extraction from stored source evidence
• 2026-07-21: 3 entries corrected (fields: industry, country). Automated classification audit: label mismatch flagged and independently confirmed by two LLM passes (DeepSeek)
• 2026-07-21: 11 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-21: 202 entries enriched. Curated employer-domicile registry backfill (deterministic public HQ facts for the largest multi-country events)
• 2026-07-21: 809 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-21: 1 entry enriched. Automated role-category extraction from stored source evidence
• 2026-07-21: 1 entry enriched. Automated source-evidence context enrichment
• 2026-07-21: 77 entries reclassified. Automated source-evidence reassessment
• 2026-07-21: 15727 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-20: 29 entries enriched. Automated industry classification: company identity + retained excerpt, double-confirmed by two model passes, fixed vocabulary, blank fields only
• 2026-07-20: 1 entry corrected (fields: industry). Automated classification audit: label mismatch flagged and independently confirmed by two LLM passes (DeepSeek)
• 2026-07-20: 3 entries enriched. Automated source-evidence context enrichment
• 2026-07-20: 4 entries reclassified. Automated source-evidence reassessment
• 2026-07-20: 37 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-20: 2 entries enriched. Automated role-category extraction from stored source evidence
• 2026-07-20: 400 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-19: 2 entries enriched. Automated role-category extraction from stored source evidence
• 2026-07-19: 1 entry corrected (fields: company). Company-name alignment: id 70595 (2026-07-13 Volkswagen 50,000) renames to Volkswagen Group so both statements group as one company.
• 2026-07-19: 400 entries corrected (fields: reason_tags). Reason-tag backfill: fixed-vocabulary tags classified only from the row's stored excerpt (ERM rows: Eurofound's own recorded restructuring type)
• 2026-07-19: 59 entries corrected (fields: reason_tags). Reason-tag backfill for 2026 seeded announcement events: ai_automation for employer-stated AI, possible_ai for broad press links, restructuring baseline. Aligns the Reasons filter and chart with the seeded rows.
• 2026-07-19: 1 entry corrected (fields: job_count, source_url, excerpt). 2025 audit batch 2: Nissan row 70233 stored the cumulative 20,000 figure; CBT News states 11,000 NEW cuts announced 2025-05-13 on top of the separately-recorded 9,000 from Jan (row 60932), storing both 20,000 and 9,000 double-counts
• 2026-07-19: 2 entries merged. 2025 audit batch 2: Recruit Holdings/Indeed/Glassdoor July 11 announcement of 1,300 cuts is triple-counted (rows 70339 Jul-10, 70341 Jul-11, 26457 Jul-11 all 1,300); keeper 26457 carries the AI attribution
• 2026-07-19: 4 entries corrected (fields: job_count, source_url, excerpt). 2025 historical-year audit: conservative-count corrections to the employer's own announced figures (Amazon press release states ~14,000 corporate roles vs the Reuters up-to-30,000 estimate; GM announced 1,750 indefinite layoffs vs the 1,200 Factory Zero slice; Target confirmed ~1,000 layoff notices with 800 vacant roles inside the 1,800 headline; Blue Origin ~1,400 = 10% of ~14,000 per Reuters/Via
• 2026-07-19: 20 entries merged. 2025 historical-year audit: confirmed duplicate rows of the same announcement (dossier: UPS Oct-28 48K disclosure duplicated by a Challenger-roundup row and superseding the Apr-29 20K stage; IRS Apr-5 same-day pair; HHS Mar-27 RIF quadruplicated plus FDA/agency sub-slices of the announced 10K breakdown; State Jul-11 pair (keep conservative 1,300); Education Jul SCOTUS row re-reporting the Mar RIF;
• 2026-07-19: 7 entries corrected (fields: industry). Industry precision pass on batch-3 seeds: e-commerce, travel and media companies had defaulted to Technology; corrected through the fixed vocabulary.
• 2026-07-19: 251 entries enriched. Curated employer-domicile registry backfill (deterministic public HQ facts for the largest multi-country events)
• 2026-07-19: 3 entries corrected (fields: industry, country). Automated classification audit: label mismatch flagged and independently confirmed by two LLM passes (DeepSeek)
• 2026-07-19: 15 entries merged. Daily cross-source dedup: same layoff event reported by multiple sources, confirmed by DeepSeek
• 2026-07-19: 2 entries enriched. Automated source-evidence context enrichment
• 2026-07-19: 92 entries reclassified. Automated source-evidence reassessment
• 2026-07-18: 1 entry corrected (fields: job_count, excerpt). R8 Oracle reconciliation: converge on the company-grounded figure. Oracle's FY2026 10-K discloses headcount fell from ~162,000 (May 31, 2025) to ~141,000 (May 31, 2026), a net ~21,000 reduction the filing links to 'the adoption and deployment of AI technologies across our operations'; the prior 30,000 was press-reported gross plan size originating from a TD Cowen analyst estimate. Gross plan repor
• 2026-07-18: 1 entry merged. R8 Oracle reconciliation: the Feb 20,000 row is analyst-forecast coverage (TD Cowen 20K-30K, 'may lay off up to 30,000') of the same Fiscal 2026 Oracle Restructuring Plan executed Mar 31, 2026 and held as the Apr 6 30,000 row; a single plan was double-counted as 50,000. Both Times of India reports are retained on the canonical event.
• 2026-07-18: 1 entry corrected (fields: job_count, country, excerpt). Dow Jan-2026 correction (plan R2 #3, amended): count restored to 3,700 – ERM factsheet 204559 books the net-new portion of the 4,500-job global plan announced Jan 29, 2026 (CBS); the remaining 800 were announced July 2025 and separately retained (row 61702 / ERM 204001), so 4,500 double-counts. The CBS source-attach step had overwritten the unpinned row to 4,500 and blanked country; both restored
• 2026-07-18: 4 entries enriched. Automated source-evidence context enrichment
• 2026-07-18: 1 entry corrected (fields: job_count, layoff_date, announced, ai_explicit, ai_causation, review_status, excerpt). R2 #1 Intuit correction: extractor parsed '17% of workforce' as 17 jobs. Corrected to the article-stated ~3,000 ('Intuit is letting 17% of its staff go, or about 3,000 people' – TechCrunch 2026-05-20; 'approximately 3,000 employees' – CFO Dive). Original CNBC URLs are 403-blocked; count evidenced via accessible allowlisted alternates. AI fields set to the company's explicit on-record denial: CEO s
• 2026-07-18: 1 entry corrected (fields: job_count, source_url, excerpt). Meta 2026-03-25 Reality Labs/recruiting/sales round undercounted: extractor stored 100 from a 'hundreds' headline (economictimes). Corrected to 700 per The Information/Bloomberg-sourced reporting ('about 700'); primary CNBC 2026-03-25 article is 403-blocked, accessible alternate quoted. Challenger gap plan R2 #2. Original source report retained on event 43623.
• 2026-07-18: 1 entry merged. R2 #4 Lucid duplicate: 'Lucid' (livemint) and 'Lucid Motors' (TechCrunch) 1,500-job rows for the 2026-06-22 announcement are the same event (18% cost-reduction round, same day); company_key normalization keeps 'motors' so dedupe never clustered them. Merged into the TechCrunch-sourced canonical row retaining both source reports.
• 2026-07-18: 29 entries reclassified. Automated source-evidence reassessment
• 2026-07-18: 2 entries corrected (fields: industry, country). Automated classification audit: label mismatch flagged and independently confirmed by two LLM passes (DeepSeek)
• 2026-07-15: Florida test rows removed, 87,600 jobs. Florida’s official WARN export contains internal test entries, which are fictitious notices sharing one WARN number and using non-existent zip codes. Eight such rows were removed, the largest a fake 78,788-worker “AT&T” notice that briefly ranked as our biggest entry. Our importer now skips test-named rows, and each removed row is permanently blocked from re-import.
• 2026-07-15: Country assigned to 88 news and SEC entries. These rows had no country recorded, which hid them from the regional views and country charts, though they were always in the worldwide totals. Each was resolved from its own source article. The largest were Oracle (30,000, spanning the US, India, Canada, Mexico and Uruguay, so “Multiple countries”) and BBC (2,000, United Kingdom).
• 2026-07-15: Ideal US Talent Systems RI corrected from 9,891 to 2. The Rhode Island notice states the company-wide figure with only 2 RI employees affected, and the per-state filings for DC, GA, IL and VA are already separate entries. Counting the company-wide total under RI double-counted the event.
• 2026-07-15: Ten non-events removed. These were SEC-filing extraction mistakes: severance dollar figures and workforce-reduction percentages misread as headcounts, WARN Act boilerplate clauses from acquisition agreements, and three duplicate rows of one Meta story carrying wrong dates.

Spotted something off? Every entry links to its primary source so you can check us. Send corrections through the contact page and they get priority.