AI Layoffs in 2026: What the Numbers Actually Mean for Job Seekers

The figures in this article come from the Layoffs.fyi AI layoffs tracker, maintained by Roger Lee. It counts only layoffs that the company itself or press coverage attributed to AI — whether to fund AI investment, to replace people with AI, or to respond to AI-driven disruption of the business. Data below reflects the tracker as of September 10, 2026.

The Trend Line

Three years of data show an acceleration that is hard to dismiss as noise.

Two things stand out. The headcount roughly tripled year over year while the number of separate events only doubled — meaning the average AI-attributed layoff is getting larger, not just more frequent. And 2026 is not a finished year. The most recent entry in the tracker is dated September 3, so this total covers roughly eight months.

The 2024 number deserves a caveat. Only four events were recorded that year, and a single one — SAP restructuring 8,000 roles — accounts for most of the total. In 2024 almost nobody described a layoff as an AI layoff. Part of what the curve measures is a genuine shift in employment, and part of it is a shift in how companies are willing to explain themselves publicly.

Who Made the Largest Cuts in 2026

The percentage column is where the human story lives. Oracle cutting 21,000 people is the biggest absolute number in the tracker, but at 13% of staff it is a restructuring. Block cutting 40% of its workforce and WiseTech cutting 30% are different events entirely — those are companies rebuilding themselves around a smaller permanent headcount.

Three Different Things Called “AI Layoffs”

Reading through all 111 events in the 2026 list, they sort cleanly into three categories. The distinction matters because each one implies a different story about you, and a different way to handle the question in an interview.

1. Funding the bet

The company is not replacing your job with AI. It is cutting costs somewhere to pay for an AI investment somewhere else. Atlassian cut 1,600 roles explicitly to self-fund further AI and enterprise sales investment. Intuit cut 3,000 while pursuing partnerships to rebuild TurboTax, QuickBooks and Mailchimp around AI. Meta slashed the Reality Labs budget by 30% and moved the money to superintelligence research. Autodesk shed 1,000 mostly customer-facing sales roles to redirect spend toward cloud and AI.

If this was your layoff, your role was not obsolete. It was a budget line. That is a much easier thing to explain, and it is usually verifiable from public reporting.

2. Direct substitution

This is the category people picture, and it is real but narrower than the headlines suggest. The clearest cases in the data involve work that is high-volume, rule-governed and reviewable:

• Content moderation. TikTok's automated systems reached the point of handling 85% of all content removals for guideline violations, and two separate cuts followed — 300 roles in July and 250 more with the closure of its Nashville office in August.
• Documentation. Snowflake eliminated its entire technical writing and documentation team, 70 people, as AI tools took over content generation.
• Production creative. SSense cut its whole full-time photography, retouching and hair-and-makeup operation, 215 roles, moving to AI-generated fashion imagery.
• Tier-one support. MercadoLibre cut 116 roles; former employees described having spent much of their work training the chatbots that ultimately replaced them.
• Manual back-office workflows. Opendoor closed its entire India operation, 250 roles that existed to run manual processes.

Engineering is not exempt, but the pattern there is compression rather than elimination. Freshworks cut 500 roles, with CEO Dennis Woodside telling staff that “over half of our code is written by AI.” The team got smaller; it did not disappear.

3. The business model broke

The rarest category and the most brutal. AI did not automate these companies' work — it destroyed demand for their product. Tailwind Labs is the starkest example in the whole dataset: developers could suddenly generate Tailwind CSS on demand, paid template revenue fell 80%, documentation traffic dropped 40%, and the company laid off three people because it could not otherwise make payroll. ZoomInfo cut 600 as AI repriced the entire B2B sales-intelligence market and eroded demand for its core data products.

If your employer was in this category, the interview framing is straightforward and blameless: the market moved underneath the company. Everyone in your industry knows it happened.

The Roles Moved — They Did Not Simply Vanish

The single most useful pattern in this data is easy to miss if you only read the totals. Several companies cut and hired in the same breath.

Paytm cut 400 roles in June while simultaneously announcing plans to hire roughly 4,000 people in AI, product and technology over the following nine months. Pentera cut marketing and headquarters roles while continuing to fill engineering positions. GitLab cut 350 people and exited 22 countries, reducing its geographic footprint by 37% — a relocation of work as much as a reduction of it.

There is also a clear structural target running through the largest cuts: middle management. Amazon framed its 16,000-person reduction around reducing layers and removing bureaucracy. Coinbase capped management at five layers, eliminated pure manager roles, and moved toward what it called “one person teams.” Patreon flattened management layers as part of its restructuring. If your job was primarily coordination — running status meetings, routing information between teams, aggregating reports — that function is being compressed harder than individual contributor work.

What This Changes in the Interview Room

Three practical shifts follow from all of this.

“How do you use AI in your work?” is now a real question, not small talk. When a company has just restructured around AI, the hiring manager is screening for whether you will accelerate that or resist it. Have a specific, honest answer ready: a tool you actually use, a task it changed, and what you still do yourself. Vague enthusiasm reads as poorly as blanket skepticism. Our guide to using AI tools to prepare for interviews is a reasonable place to build that answer from.

Output per person is the metric being rewarded. DeepL's CEO described a structural shift in how many people it takes to do work well. Multiverse told the press its revenue per employee was up 37%. Whatever your function, the version of your accomplishment that lands in 2026 is the one with a ratio in it — work absorbed, throughput per head, scope covered without adding people.

Explaining a layoff is now a structural story, not a personal one. This is genuinely easier than it was three years ago, because the interviewer has almost certainly seen the same headlines.

How to Talk About Being Laid Off

The goal is three sentences: what happened, that it was structural, and what you did next. No bitterness, no over-explaining, no volunteering more than was asked.

“My team was cut in March when the company restructured to fund its AI platform investment — about 10% of staff went in the same round. I'd been leading the billing integrations there for three years. Since then I've been going deep on how our kind of workflow gets rebuilt with LLM tooling, which is part of why this role interested me.”

That works because it names the cause, gives the scale so it reads as structural rather than performance-related, and pivots to something forward-looking within a few seconds. Three specific mistakes to avoid:

• Don't editorialize about the decision. “Honestly it was a terrible call and the product's fallen apart since” may even be true. It still costs you the room.
• Don't inflate your role in it. If you were one of 8,000, say so. It is stronger, not weaker.
• Don't leave the gap unexplained. Interviewers read silence as something worse than the truth. Name what you have been doing, even if it is study and contract work.

If your layoff was in the substitution category — your specific function was automated — the move is to name it before they wonder. Acknowledge the function changed, then talk about where your judgment still applies. A moderator who understands policy edge cases is valuable to the team that now supervises an automated system. A technical writer who knows what documentation readers actually get stuck on is valuable to the team that reviews generated docs.

For the mechanics of building these answers properly, our STAR method guide covers the structure, and how to answer “tell me about yourself” covers where in your narrative the layoff belongs. The full 2026 interview preparation guide pulls the whole sequence together.

Questions Worth Asking Them

The same data that makes this market uncomfortable gives you unusually good material for the end of the interview. Asking about AI strategy signals that you have read the landscape:

• “Which parts of this team's work have changed most as AI tooling came in over the last year?”
• “Is this role backfilling something, or is it new scope?”
• “How does the team decide what stays human review?”

The answers tell you whether you are joining a team that is investing or one that is mid-contraction. That is worth knowing before you accept. Our list of questions to ask your interviewer has more, and if you get to an offer, salary negotiation after an offer matters more than usual in a market where employers assume candidates are anxious.

The Honest Read

92,713 is a large number and it is still growing. It is also worth keeping in proportion: it counts layoffs across the entire global tech sector over eight months, and the tracker explicitly includes cuts where AI was the stated reason for reallocating budget rather than the thing doing the work. Companies also have an incentive to attribute layoffs to AI, because “we are becoming an AI-native company” reads better to investors than “we over-hired.” Some share of this curve is narrative.

What is not narrative: the roles being cut cluster in identifiable places, the surviving teams are smaller, and the people getting hired are the ones who can show what they produce rather than how many people they coordinate. Interviews in this market reward specificity and candor about the AI question more than they ever have. That part is entirely within your control.