Updates
Corrections, methodology changes, and notable additions to the dataset. Newest first. Banners marked here surface on the homepage until their expiry date.
- UpdateMacro context — TrueUp's mid-June tally shows 150,000+ tech jobs cut year-to-date in 2026, with AI cited in 55% of events (152,415 workers across 135 companies). (TechCrunch, June 15)
- UpdateMacro context — Challenger, Gray & Christmas reported 97,006 US layoffs announced in May 2026, the highest May since 2020, with tech leading at 38,242. AI was cited in 38,579 cuts — the third straight month AI ranked as the #1 stated cause. (WSJ / Finimize, June 4)
- CorrectionDBS Bank's reported figure was revised from 4,000 to 2,500 after Reuters clarified the cut affected outsourcing contractors, not direct hires.
- UpdateOracle's 30,000-cut announcement added to the ledger. The company explicitly named AI-driven productivity gains as the reason in its Q4 earnings call. See Oracle report →
- UpdateMethodology change: the hero counter now shows US-only jobs, not global. Earlier reports' US allocations were back-filled where geography could be confirmed.
- CorrectionMicrosoft's March cut was initially logged as Explicit attribution. After a closer reading of the announcement language, the tier has been downgraded to Mixed: AI was mentioned alongside several other restructuring drivers.
- NoticeMultiple major announcements logged today (Amazon, Dow, Expedia, ASML). All four entries reviewed against AP and Reuters wire coverage before publication.
- UpdateGoogle's Q3 cut added. Attribution: Mixed; internal memo cited AI tooling rollouts as a contributor among several factors.
- UpdateIndustry taxonomy refined. 'Technology / E-commerce' merged into 'Technology' to reduce category sprawl; affected Amazon's three historical reports.
- NoticeSite launched. Initial dataset of 12 reports backfilled from public AP, Reuters, CNBC, and Bloomberg coverage from January through May 2025.
Spotted an error or have a tip? jobloss@secureainow.org. Corrections are public and dated.
Read the full methodology →