AI Layoffs 2026: 185K Jobs Cut, but Gartner Says It's Not Working

TL;DR

Tech companies have cut 185,894 jobs in 2026 (averaging 1,115 per day), and most of them blame AI. But Gartner surveyed 350 executives and found no connection between workforce reductions and higher returns. Deutsche Bank analysts coined a term for it: AI redundancy washing. Forrester puts it more bluntly — 55% of employers who cut jobs citing AI already regret it, and half will quietly rehire within the year, often offshore at lower pay.

The 1,115-Per-Day Number

I’ve been tracking layoff trackers from Limassol since early 2025. Back then, companies said “post-pandemic restructuring.” In Q4 2025, it shifted to “efficiency gains.” By January 2026, every earnings call had the same three-word justification: “because of AI.”

The numbers from layoffs.fyi and TrueUp tell one consistent story. Through mid-June 2026:

• 267 layoff events across the tech industry
• 185,894 workers displaced
• 1,115 jobs lost per day on average, nearly double 2025’s pace

Block fired 40% of its workforce (over 4,000 people) with CEO Jack Dorsey citing “the growing capability of AI tools.” Atlassian cut 1,600 (10% of headcount). Meta has run multiple reduction rounds through the year, cutting approximately 10,000 across January, March, and May. Oracle opened 2026 with the single largest cut at 30,000.

Amazon alone accounted for 52% of tech layoffs in Q1 2026. At the same time, Amazon’s Seattle job postings collapsed 80% — from 22,700 in H1 2025 to 4,540 in Q1 2026, while Milan and Pisa quietly became its top 2-3 hiring cities globally. That pattern deserves a closer look.

What Is AI Redundancy Washing?

Deutsche Bank analysts coined the term in their January 2026 outlook: AI redundancy washing is the practice of attributing layoffs to artificial intelligence when the actual drivers are ordinary cost-cutting, post-pandemic over-hiring corrections, or pressure to fund infrastructure spending.

The analysts predicted it would be “a significant feature of 2026.” They were right. AI was cited as the stated reason in approximately 25% of all tech layoffs in March 2026, according to TechTimes reporting. But the same article noted that most of these companies are profitable — they aren’t cutting because they’re struggling. They’re cutting because $650 billion in combined AI infrastructure commitments from Amazon, Meta, Google, and Microsoft needs to come from somewhere, and payroll is the largest controllable line item.

Peter Cappelli, a management professor at Wharton, summarized the pattern: the headline is always “because of AI,” but read the actual announcements and executives are saying they expect AI to cover the work. They haven’t automated anything yet. They’re hoping.

That sentence is worth re-reading. Companies fire people first, then bet that AI will eventually fill the gap. The layoff announcement comes before the technology exists.

Gartner: No Correlation Between Layoffs and Returns

In May 2026, Gartner published findings from a survey of 350 global executives conducted in Q3 2025. Every company in the sample had at least $1 billion in annual revenue and had deployed at least one AI agent, intelligent automation, or autonomous technology.

The headline finding: 80% of companies reported workforce reductions, but there was no correlation between cutting headcount and achieving higher ROI. Workforce reduction rates were nearly equal between high-ROI and low-ROI companies.

Helen Poitevin, a Gartner analyst, put it directly: “Many CEOs turn to layoffs to demonstrate quick AI returns; however, this disposition is misplaced.”

The companies reporting the highest returns used AI for what Gartner called “people amplification.” They made existing workers more productive instead of replacing them.

This is consistent with Goldman Sachs research from senior U.S. economist Ronnie Walker, who found no “meaningful relationship between productivity and AI adoption” across the companies Goldman tracks.

The CFO Survey: 502,000 Jobs, Not 10 Million

A Duke University / NBER working paper published in March 2026 surveyed 750 U.S. CFOs through the Duke CFO Survey (a partnership with the Federal Reserve Banks of Atlanta and Richmond). The results were sobering in a different way than the headlines suggest.

44% of CFOs said they plan AI-related job cuts. That sounds alarming. But when they quantified the actual scope, the number was 0.4% of the total U.S. workforce — approximately 502,000 roles out of 125 million. About half would be white-collar positions.

John Graham, the Duke CFO Survey director, said: “It’s not the doomsday job scenario that you might sometimes see in the headlines.”

That 502,000 figure represents a 9x increase from 2025’s 55,000 AI-attributed layoffs. It’s real. But it’s not the millions that dominate LinkedIn discourse. And it tells us something: even the people making the cuts know the scale is limited.

Forrester’s Regret Index: 55% Already Wish They Hadn’t

Forrester’s Predictions 2026 report may be the most damning data point in this entire story. Their finding: 55% of employers who made AI-attributed layoffs already regret the decision.

The reasons track with what Gartner found:

• Only 2% of large layoffs were tied to actual AI implementation; the rest were speculative
• Companies laid off workers for AI capabilities that don’t yet exist, betting on future promises
• Forrester predicts that 50% of those cut workers will be quietly rehired, but offshore or at lower salaries

The Klarna case became the poster child. Klarna replaced 700 customer service employees with AI, trumpeted the savings publicly, then watched quality metrics decline and customer complaints spike. They rehired humans, but the reversal got a fraction of the press coverage.

Type A vs. Type B: A Data-Driven Split

JobSpikr’s 2026 ROI analysis introduced a useful framework for classifying AI layoffs:

The 6/94 split is striking. Genuine AI displacement is happening — data entry clerks (-34.1%), customer support (-28.8%), copywriters (-28.1%), and telemarketers (-29.3%) have seen real, sustained declines over multiple quarters. These roles were already shrinking before anyone called it “AI layoffs.”

But 94% of what’s being labeled “AI layoffs” in 2026 follows the Type B pattern: companies hired aggressively through H2 2025, then slashed headcount in Q1 2026 with AI as the stated justification. Amazon’s hiring surge followed by a Seattle posting collapse is the textbook example.

The Hiring Paradox

If AI is genuinely replacing workers at scale, you’d expect hiring for AI-adjacent roles to flatten. The machines should be handling it. Instead:

• AI Engineer postings: +654% (H1 2024 to H2 2025)
• AI Specialist demand: +436%
• AI Project Manager: +404%
• Prompt Engineering: +777% growth over 18 months
• AI Governance: +1,257% growth

Meanwhile, 78% of technology hiring managers plan to increase permanent headcount (up from 61% earlier in 2026), according to Robert Half’s 2026 talent report. Dice’s June 2026 data shows tech job postings up 23% year-over-year compared to May 2025, the strongest comparison of the year so far.

Companies are simultaneously cutting roles they call “AI-displaced” and frantically hiring people to build, maintain, monitor, and govern the AI systems they claim are doing the displacing. You can hold both facts at once, but the “AI efficiency” framing falls apart the moment you check the job boards.

If you want to spot the pattern yourself, here’s a quick Python snippet using the public layoffs.fyi CSV data (downloadable from their site) to flag companies where the hiring-vs-cutting signal conflicts:

import pandas as pd

layoffs = pd.read_csv("layoffs_2026.csv")
# Group by company, compare headcount changes vs AI-cited layoffs
company_cuts = (
    layoffs[layoffs["industry"] == "Tech"]
    .groupby("company")
    .agg(
        total_laid_off=("laid_off", "sum"),
        ai_cited=("ai_related", "sum"),
        events=("date", "count"),
    )
    .sort_values("total_laid_off", ascending=False)
)

# Flag Type B pattern: ai_cited > 0 but company also has open roles
# Cross-reference with job postings data (e.g. from LinkedIn API or Otta)
type_b = company_cuts[
    (company_cuts["ai_cited"] > 0)
    & (company_cuts["events"] >= 2)  # multiple rounds = restructuring
]
print(f"Type B candidates: {len(type_b)} companies")
print(type_b.head(10))

This won’t give you a definitive answer — you’d need to match it against each company’s open postings, but it flags the obvious cases where “AI-driven layoffs” and “now hiring AI engineers” appear on the same company page.

The MIT and McKinsey Reality Check

Two earlier studies provide context for why the ROI isn’t materializing:

An MIT study from 2025 found that 95% of enterprise AI pilot programs stalled, delivering little to no measurable impact on the P&L. Not “underperformed expectations.” Most of them never got past the pilot stage.

McKinsey’s State of AI 2025 report found that only 6% of organizations qualify as genuine “AI high performers” achieving 5%+ EBIT impact from their AI deployments. Deloitte’s parallel research pegged the typical payback period for AI investment at 2-4 years, far longer than the 7-12 months companies assumed.

The gap between what companies expect from AI and what they’ve measured so far is where AI redundancy washing lives.

What This Means If You Work in Tech

I wrote about the bifurcated job market in May and about whether AI is replacing juniors in June. This piece adds a third dimension: a lot of the displacement narrative is corporate theater.

That doesn’t mean the threat isn’t real. The 6% of layoffs that are evidence-led represent genuine structural change. Data entry, first-line support, templated copywriting, and telemarketing are declining for real. If your role is in that column, the “AI redundancy washing” framing won’t save your job.

But if you’re a software engineer, a data scientist, or anyone whose work involves judgment and ambiguity, the data says something different: companies that cut your kind of role saw no ROI improvement, and more than half regretted it.

For developers in Europe and Cyprus specifically: the geographic arbitrage pattern is worth watching. Amazon’s shift from Seattle to Milan mirrors a broader trend. Forrester predicts that rehired workers come back offshore at lower salaries. Eastern European and Southern European tech hubs (Warsaw, Bucharest, Lisbon, and yes, Limassol) are net beneficiaries of this reshuffling. The layoffs might be badly reasoned, but the geographic redistribution of work they trigger is real.

Three things worth doing now:

1. Learn to quantify your AI fluency. When the next round of cuts comes, managers will protect the people who can demonstrate productivity gains from AI tools, the ones who go beyond Copilot autocomplete. If you can show that you shipped a feature 3x faster using Claude Code or reduced your review cycle with AI coding agents, that’s a data point in your favor.

2. Build skills that sit next to AI, not under it. AI Governance postings grew 1,257% in 18 months. Security, compliance, and architecture roles are growing precisely because AI creates new risk surfaces. I covered the LiteLLM vulnerability story last week — every AI deployment creates new attack vectors that need human oversight.

3. Watch the rehire signal. Forrester’s 50% quiet-rehire prediction means job openings will appear in the same companies that just cut — often within 6-12 months, often in different geographies, often at different titles. If you were cut by a Type B company, don’t write them off as a future employer.

FAQ

Are tech layoffs really because of AI?

Partially. JobSpikr’s analysis found that roughly 6% of AI-attributed layoffs in 2025-2026 follow an evidence-led pattern where roles genuinely declined over multiple quarters before cuts. The remaining 94% are speculative — companies cut first and hope AI fills the gap later. Deutsche Bank calls this pattern “AI redundancy washing.”

What is AI redundancy washing?

A term coined by Deutsche Bank analysts in January 2026 for the practice of attributing layoffs to AI when the real drivers are cost-cutting, post-pandemic hiring corrections, or pressure to fund infrastructure spending. Wharton professor Peter Cappelli described the same phenomenon: companies fire people first, then expect AI to cover the work.

How many AI layoffs have there been in 2026?

Through mid-June 2026, 185,894 tech workers have been laid off across 267 events, averaging 1,115 per day. AI is explicitly cited in approximately 25% of those cuts. The Duke CFO Survey projects about 502,000 AI-related job losses across the full U.S. economy in 2026, a 9x increase from 2025, but still only 0.4% of the total workforce.

Do AI layoffs improve company performance?

No, according to the best available evidence. Gartner surveyed 350 executives and found no correlation between workforce reductions and higher ROI. Companies that cut the most saw returns similar to those that cut the least. Forrester found 55% of employers who made AI-attributed cuts already regret the decision.

Which companies are doing the most AI layoffs?

The largest cuts in 2026 include Oracle (30,000), Block (4,000+, or 40% of workforce), Meta (ongoing weekly cycles), Atlassian (1,600), and Amazon (52% of Q1 tech layoffs). Amazon’s pattern is notable: Seattle postings dropped 80% while Italian offices became top hiring locations, suggesting geographic cost arbitrage rather than genuine AI displacement.

Sources

• Gartner: AI Layoffs May Create Budget Room but Do Not Deliver Returns — the 350-executive study finding no ROI correlation
• Fortune: CFOs Admit AI Layoffs Will Be 9x Higher — Duke/NBER CFO survey of 750 firms
• TechTimes: Tech Layoffs Hit 1,115 a Day — June 2026 layoff numbers and analysis
• CNBC: Deutsche Bank Says the Honeymoon Is Over for AI — origin of the “AI redundancy washing” term
• HR Executive: The AI Layoff Trap — Why Half Will Be Quietly Rehired — Forrester’s 55% regret finding
• JobSpikr: AI Layoffs 2026 — The ROI Reality Check — Type A vs Type B framework and role decline data
• TechTimes: Tech Layoffs Reach 142,000 in 2026 — $700B AI infrastructure connection
• Blockchain Council: Are Tech Layoffs Really Due to AI? — Wharton analysis and company case studies
• Dice: Tech Job Postings Rebound 23% YoY — hiring rebound data

Bottom Line

AI redundancy washing is the 2026 version of “pivoting to mobile” or “blockchain strategy”: a convenient label that helps executives justify decisions they would have made anyway. The Gartner data is direct enough — cutting headcount doesn’t improve AI ROI. And Forrester found that more than half the companies that tried it wish they hadn’t.

Real AI displacement exists. It’s concentrated in routine, templated roles that were already declining. If your work involves repetitive pattern-matching with clear inputs and outputs, the automation risk is genuine and growing.

But for the 94% of “AI layoffs” that are Type B — speculative cuts made under investor pressure with no deployed replacement — the best response is paying attention to who’s quietly rehiring six months later, and making sure your skills are on the right side of that 6/94 split.