Data as of 28 September 2026. The full list is available as a free CSV download at the end of this article.
Through August 2026, US employers attributed 116,175 job cuts to artificial intelligence, about 22% of all announced cuts, according to Challenger, Gray & Christmas. That makes AI the leading stated reason for layoffs this year.
But "blamed AI" covers three very different things. Some companies said AI does the work now. Some said they are cutting people to pay for AI. And some big layoffs that get lumped in with AI were explained with something else entirely. (For the year-by-year picture, see our 2021–2026 timeline of AI tools and layoffs.) I went through the announcements from 2023 to 2026 and sorted them by what each company actually said.
At a glance
| What the company said | Companies |
|---|---|
| AI does this work now | Block, Atlassian, Salesforce, Chegg, Duolingo (contractors), IBM (hiring) |
| Cutting costs to fund AI | Meta, Oracle |
| Not AI, or explicitly denied | Amazon, Microsoft, Alphabet (2023), Autodesk |
Group 1: "AI does this work now"
These companies tied the cuts, or fewer hires, directly to AI doing the work.
| Date | Company | Cuts | What they said |
|---|---|---|---|
| Feb 2026 | Block | ~4,000 (~40%) | "Intelligence tools have changed what it means to build and run a company" |
| Mar 2026 | Atlassian | 1,600 (~10%) | Cuts announced citing the AI shift |
| Sep 2025 | Salesforce | ~4,000 support roles | "I've reduced it from 9,000 heads to about 5,000, because I need less heads" |
| Oct 2025 | Chegg | 388 (~45%) | "The new realities of AI" plus less traffic from search |
| Apr 2025 | Duolingo | Contractors only | Will "gradually stop using contractors to do work that AI can handle" |
| May 2023 | IBM | Hiring slowdown, not layoffs | CEO "could easily see 30%" of about 26,000 back-office roles replaced over five years |
Notice the pattern: most of these roles are support, content or back-office work, not engineering. Salesforce said it "no longer need[s] to actively backfill support engineer roles" because its AI agents handle more cases. Block is the exception that reached engineering: Fortune reported its leaders decided after testing models like Anthropic's Opus 4.6 and OpenAI's Codex 5.3 over the 2025 holidays.
Group 2: "We're cutting to pay for AI"
These companies are not saying AI replaced the people they let go. They are saying AI infrastructure is expensive and payroll is where the savings come from.
| Date | Company | Cuts | What they said |
|---|---|---|---|
| Apr 2026 | Meta | ~8,000 (10%) + 6,000 open roles cancelled | "to run the company more efficiently and to allow us to offset the other investments we're making" |
| Mar 2026 | Oracle | "Thousands" | Cutting jobs as it keeps ramping AI spending and faces investor pressure over debt for data centres |
Meta's case shows why this group is different: alongside the cuts, it said upward of 7,000 workers would move into new AI-focused teams. The jobs were not automated. The money moved.
A note on Oracle: you will see "30,000" in some coverage. That was an analyst estimate, not a company figure. CNBC reported the cuts as "thousands", and that is what we use.
Group 3: Not AI, or explicitly denied
| Date | Company | Cuts | What they said |
|---|---|---|---|
| Jan 2023 | Amazon | Just over 18,000 | "The uncertain economy and that we've hired rapidly over the last several years" |
| Jan 2023 | Microsoft | 10,000 | "Macroeconomic conditions and changing customer priorities" |
| Jan 2023 | Alphabet | ~12,000 | "We hired for a different economic reality than the one we face today" |
| Oct 2025 | Amazon | ~14,000 corporate | Jassy: "not really financially driven, and it's not even really AI-driven, not right now at least" |
| Jan 2026 | Autodesk | 1,000 (~7%) | Not "an effort to replace people with AI"; restructuring of sales |
| Jan 2026 | Amazon | ~16,000 corporate | "Reducing layers, increasing ownership, and removing bureaucracy" |
The 2023 cuts were the biggest of the whole cycle, and not one of these companies mentioned AI. Amazon's 2025–26 rounds are often reported as AI layoffs, but its own explanation was management layers and culture.
The ones that walked it back
- Klarna leaned hard on AI for customer service, then in May 2025 started hiring human support staff again after its CEO acknowledged the AI-only approach led to "lower quality".
- Duolingo's CEO later clarified the company had not laid off full-time employees because of AI, after a backlash to the "AI-first" memo.
- Block's Jack Dorsey posted the day after the cuts: "yes we over-hired during covid because i incorrectly built 2 separate company structures (square & cash app) rather than 1".
How to read the next "AI layoff" headline
- Which group is it? "AI does the work", "paying for AI", or AI mentioned only in passing?
- Which roles were cut? Support and content roles are a different story from engineering.
- Did the company over-hire in 2020–2022? If headcount grew fast then, the cut may be a correction.
- Is it a number from the company or an estimate? Oracle's "30,000" is the example to remember.
- Are they still hiring? Meta moved about 7,000 people into AI teams in the same announcement.
Remember that every AI figure here is self-reported. Challenger records the reason a company gives, not a verified cause.
Download the dataset
All 14 events in this article, with date, company, cuts, roles, how AI was framed, the stated reason and a source link, are in one file: ai-layoffs-2023-2026.csv. You are free to reuse it. A link back to this page is appreciated.
- Need it as JSON for an app or a notebook? Paste it into our CSV to JSON converter. It converts in your browser, nothing is uploaded.
- Want a quick chart, for example cuts by company? Paste the columns into the chart builder and export a PNG.
I will add new rows as companies announce cuts and update the date at the top when I do.