AI layoffs tracker

Why a confidence rating, not just a list

Almost every “AI layoffs” list treats a headline as data. If a company cut jobs and a journalist mentioned AI in the same article, it goes on the list — regardless of whether the company said AI caused it, said something more hedged about AI investment, or explicitly denied AI was the reason. That collapses three very different situations into one number, and it is the main reason AI layoff statistics are so easy to overstate.

This tracker splits every entry into three tiers instead:

• Company-attributed. A named executive, on the record — an earnings call, SEC filing, employee memo or interview — said AI is a reason for the specific cut. Salesforce's Marc Benioff saying “I need less heads” with AI absorbing support volume is the clearest example on this page.
• Mixed signal. The company cites AI as an investment priority or a reason to reallocate budget, without directly saying AI performs the eliminated work — or names AI alongside other stated causes. Workday and Cisco fall here: both cited AI as where resources are moving, not as what replaced the cut roles.
• Press-framed only. Only journalists or analysts linked the cuts to AI; the company gave a different reason or none. Amazon's October 2025 round is the sharpest case here — CEO Andy Jassy said on the record it was “not really AI-driven,” even as outlets ran headlines linking Amazon and Target job cuts to AI reshaping retail.

As of August 1, 2026, 11 of 25 entries are company-attributed, 6 are mixed signal, and 8 are press-framed only — and 1 of the company-attributed cases (Klarna) were significantly walked back within a year of the original claim.

The aggregate trend: Challenger, Gray & Christmas data

The single most cited source for AI-layoff statistics is Challenger, Gray & Christmas, an outplacement firm that has coded “Artificial Intelligence” as a distinct cited reason for job cuts since 2023. Their number counts company press releases and public statements that mention AI as a reason — it is a count of announcements, not an independent measurement of what actually caused each cut, and a single announcement can cite more than one reason. With that caveat, the trend is unambiguous:

2023 and 2024 figures are derived by subtracting Challenger's published cumulative “since 2023” totals from each other, since Challenger does not publish a standalone 2023 annual figure. All other numbers are stated directly in Challenger's monthly and year-end reports.

Two things stand out. First, AI's share of total cuts has grown far faster than total cuts themselves — from under 2% of all 2024 cuts to roughly 23% of all 2026 cuts through June, even though overall layoff volume did not rise nearly as sharply. Second, month to month it is volatile: May 2026's 38,579 AI-cited cuts was the highest single month on record, and June fell 64% from it. Treat single-month spikes as noisy, and the year-over-year direction as the signal.

What the macro numbers do — and don't — show

The U.S. Bureau of Labor Statistics runs the Job Openings and Labor Turnover Survey (JOLTS), the government's official layoffs-and-discharges data. It recorded 1.7 million layoffs and discharges across all industries in May 2026, with the layoff rate in Information — the sector that contains most AI-exposed tech jobs — rising to 2.4% year over year, up from 1.3%. JOLTS does not ask employers why they laid workers off, and it has no “AI” category at all. It confirms that layoffs are elevated in AI-exposed sectors; it cannot say how much of that is AI.

That gap — a government data series with no cause field, next to a private firm's count of self-reported cited reasons — is the actual state of AI-layoff measurement in 2026. Nobody publishes a rigorous, causally identified count of jobs AI has eliminated, and treat any headline claiming otherwise with real skepticism. For the occupation-level question of which jobs are most exposed to AI capability, rather than which jobs have already been cut, see our AI job risk tool, which works from four separate exposure studies instead of announced layoffs.

AI attribution is a narrative choice, not always a fact

Citing AI in a layoff announcement is, in part, a communications decision — and it cuts both ways. Some executives reach for AI because it reads better to investors than “we over-hired in 2021” or “our costs were too high,” a pattern CNBC has termed “AI-washing.” Microsoft cut roughly 15,000 roles across 2025 while posting record profit, with AI language present in company communications but management-layer reduction cited alongside it — a pattern that makes a clean single cause hard to defend either way.

The reverse also happens. Amazon's Andy Jassy explicitly rejected the AI framing for the October 2025 round of cuts, calling it a cultural reset against organizational “layers” — four months after his own memo said AI would reduce Amazon's total corporate workforce over time. Both statements came from the same executive about the same company; they describe different things (a specific event vs. a multi-year trend) and neither one is necessarily false.

IBM's 2023 case shows how a real statement can mutate into a false one on the way through the press. CEO Arvind Krishna told Bloomberg he would pause hiring on up to 7,800 roles AI might replace over five years — a hiring-pause projection. It is now widely recirculated online as “IBM laid off 8,000 workers to replace them with AI,” a factually different claim that this tracker deliberately corrects rather than repeats.

Methodology, inclusion criteria and limitations

What qualifies for an entry

• A specific, dated announcement from a named company, sourced to a primary statement (memo, SEC filing, earnings call) or a named, reputable outlet.
• AI must appear somewhere in the public discussion of the cut — either from the company or from credible press coverage — even when we rate the entry press-framed-only. The rating describes the strength of the AI link, not whether AI was mentioned at all.
• No estimated headcounts. If a source gives a percentage but not a number, the headcount field is left blank rather than back-calculated from a separately reported total headcount, since those totals shift quarter to quarter.

What is deliberately excluded

• Rumors, anonymous-source-only reports, and unconfirmed internal leaks.
• Layoffs with no public AI connection at all, even if AI is plausibly a contributing factor — plausibility without a stated or reported link is not enough for inclusion.

Sources

Company-level entries are drawn from primary company communications where available (SEC filings, employee memos, earnings-call transcripts) and otherwise from named outlets including CNBC, Bloomberg, Axios, Reuters, the Washington Post, TechCrunch and trade press such as CFO Dive and Modern Retail — never from aggregator sites republishing a claim without their own reporting. Aggregate trend data comes from Challenger, Gray & Christmas monthly and year-end job-cut reports, captured 2026-07-06, and from the BLS JOLTS program.

Update cadence

Monthly. Aggregate figures are refreshed within the first week after Challenger publishes its prior-month report; company-level entries are added on a rolling basis as they are announced and independently sourced. The table is a typed array in data.ts in this page's source folder — new rows are appended there with a source link, never estimated.

Known limitations

• This is a curated sample, not a census. It covers large, well-reported companies. Layoffs at smaller employers are systematically underrepresented, as they are in Challenger's data too — Challenger itself has noted AI-driven cuts are likely undercounted because many companies avoid naming AI publicly for legal or reputational reasons.
• Attribution is not causation. A company-attributed rating means an executive said AI was a reason — it does not independently verify that claim. Executives have incentives to overstate AI's role (a growth story for investors) and to understate it (avoiding blame for over-hiring or a weak product cycle).
• US and large-company skew. Most sourcing here is US-headquartered, English-language coverage. Non-US and privately held companies are underrepresented.
• Net effect is out of scope. This tracker counts cuts, not hiring. Several companies here — IBM, Accenture, CrowdStrike — hired heavily in AI-adjacent roles the same year they cut elsewhere. A cuts-only table cannot tell you whether AI is net-negative for a company's headcount, only where cuts were announced and how they were explained.

Corrections and additions are welcome — email the address in the footer with a source link and the specific entry it affects.

Why we built this

Planetary Labour builds an autonomous go-to-market engine — software that runs marketing work itself. We are, plainly, on the automation side of this question, which is exactly why we would rather publish the distinction between a real AI-attributed layoff and a press-framed one than a number that flatters either argument. The pattern in the data above is more interesting than a single scary total: companies are using AI language in very different ways, and only a fraction of the layoffs given AI headlines were actually attributed to AI by the company doing the cutting.

Frequently asked questions