60% of Companies Cut Workers for AI's Potential — Not Its Performance. Here's What the Data Shows.
In March 2026, 45,000 tech workers lost their jobs. Roughly 20.4% of those cuts were explicitly attributed to AI and automation by the companies that made them — a thirteenfold increase from 2023, when AI was cited in fewer than 2% of layoff announcements.
The numbers suggest AI is reshaping the workforce at an accelerating pace. But when researchers look past the announcements and into the actual data, a different picture emerges — one where the narrative is running well ahead of the technology.
The Harvard Business Review Survey
In January 2026, Harvard Business Review published findings from a survey of 1,006 global executives conducted in December 2025. The core finding: 60% of organizations had already reduced headcount in anticipation of AI's future impact. Another 29% had slowed hiring for the same reason.
The critical distinction was in the motivation. Only 2% of executives said large layoffs were tied to actual AI implementation. The remaining 58% were cutting jobs based on what AI might do — not what it was currently doing.
The survey, sponsored by Scaled Agile and conducted among executives familiar with their companies' AI initiatives, also found that 44% considered generative AI the most difficult AI type to evaluate economically. Despite this uncertainty, 90% reported their organizations received "moderate or a great deal of value from AI" — a finding the authors noted could reflect optimism bias as much as measured returns.
The authors' framing was precise: "Job losses and slowed hiring are real, even though companies are still waiting for generative AI to deliver on its promises." Individual productivity improvements — 10-15% in programming tasks, for example — had not yet translated into organizational-level efficiency gains large enough to justify the scale of workforce reductions.
The WARN Notice Gap
New York became the first U.S. state to require companies to disclose whether AI or automation contributed to layoffs, amending its Worker Adjustment and Retraining Notification (WARN) Act in March 2025. Since then, 162 companies have filed WARN notices covering approximately 28,300 workers.
Not one company, out of 162, disclosed that technology or automation contributed to its workforce reduction — even as many of those same companies publicly attributed cuts to AI in investor communications and press releases.
The disconnect is notable. Amazon filed WARN notices for 660 New York positions citing "economic" reasons — while CEO Andy Jassy had publicly stated that AI productivity would drive workforce changes. Goldman Sachs filed similar notices without checking the technology box. The law carries no meaningful penalties for noncompliance, which workforce analysts suggest is a significant factor in the zero-disclosure rate.
The result is a data point with two possible readings. Either AI genuinely isn't driving these layoffs and the public statements are overstated, or companies are choosing not to disclose AI's role when there's a formal checkbox — even as they emphasize it everywhere else.
Three Case Studies
Block: 4,000 Jobs, Stock Up 18%
In February 2026, Block CEO Jack Dorsey announced the elimination of roughly 4,000 jobs — nearly half the company's workforce — reducing headcount from over 10,000 to just under 6,000. Dorsey attributed the decision directly to AI capabilities: "A significantly smaller team, using the tools we're building, can do more and do it better."
Dorsey stated that the company was not in financial trouble. Block had just reported Q4 gross profit of $2.87 billion, up 24% year-over-year. The cuts, he said, were proactive: "I'd rather get there honestly and on our own terms than be forced into it reactively." He predicted that within a year, "the majority of companies will reach the same conclusion."
Block's stock surged nearly 18% on the announcement.
However, questions followed. Bloomberg reported that internal employees could not clearly explain how the AI transformation would work in practice. Josh Bersin, an HR industry analyst, noted that Block's gross margins trailed competitors like Visa, Mastercard, and Shopify — suggesting the financial motivation for cost-cutting existed independently of AI strategy. His assessment: "Don't take this announcement at face value. AI is not a 'job eliminating' strategy — it's an opportunity to re-engineer what you do."
Oxford Economics reached a similar conclusion more broadly: attributing cuts to AI "conveys a more positive message to investors" than citing weak demand or past overhiring.
Atlassian: The Five-Month Reversal
In October 2025, Atlassian CEO Mike Cannon-Brookes appeared on the 20VC podcast and stated that technology creation is "not output-bound" — that Atlassian would employ more engineers in five years, not fewer.
Five months later, in March 2026, Cannon-Brookes announced 1,600 layoffs — approximately 10% of the company's 16,000-person workforce — framed as adaptation to the AI era. More than 900 of the affected roles — over half of all cuts — came from the same software research and development departments where, months earlier, hiring growth had been predicted.
Atlassian stated it was self-funding further AI investment and strengthening its financial profile. The company's shares fell on the announcement. Whether the cuts reflected a genuine strategic shift, a response to investor pressure, or both, the timeline raised questions about how quickly "AI strategy" could change from hiring plan to layoff justification.
Klarna: The Retreat
Swedish fintech Klarna became one of the earliest and most vocal proponents of AI-driven workforce reduction, cutting approximately 40% of its workforce between December 2022 and 2024 through hiring freezes and attrition. CEO Sebastian Siemiatkowski declared that AI customer service agents were handling two-thirds of all support interactions within the first month of deployment.
By 2025, Klarna was publicly walking it back. The company acknowledged that cost-cutting had led to "lower quality" and began rehiring — approximately 20 customer service agents initially, with Siemiatkowski stating the company was "reinvesting in human support."
Klarna's experience illustrated a pattern: aggressive AI-attributed cuts followed by a discovery that the technology couldn't fully replace what was eliminated.
The Regret Data
Klarna is not alone. Forrester's Predictions 2026 report found that 55% of employers who made AI-attributed layoffs already regret the decision. More than a third of those companies have rehired more than half the roles they eliminated, many within six months — often at higher cost and with permanent damage to institutional knowledge and employer reputation.
The EY US AI Pulse Survey from December 2025 added another dimension: 96% of companies report AI productivity gains, but only 17% say those gains led to reduced headcount. The remaining 79% reinvested the gains into new capabilities, cybersecurity, and upskilling rather than headcount reduction.
When researchers from Resume.org surveyed 1,000 U.S. hiring managers in early 2026, 55% expected more layoffs — and 44% anticipated AI would be a top driver. The expectation of AI-driven layoffs appears to be self-reinforcing: companies cut because they expect others to cut, creating a cycle driven more by competitive anxiety than operational data.
The Counter-Argument
Not all AI layoffs fit the "anticipation over implementation" pattern. Some companies present measurable evidence for AI-driven workforce changes.
Fortune reported on a CEO who eliminated approximately 80% of his staff two years ago after employees resisted AI adoption. He says he'd do it again — and that the remaining 20% now produces more with AI tools than the full team did before.
Block's own Q4 numbers showed expanding customer base and improved profitability. Dorsey's argument that smaller teams with better tools can produce more output has theoretical support in productivity economics — even if the specifics at Block remain undetailed.
And the broader trend data is real. Fortune reported that 66% of CEOs plan to freeze or cut hiring through the rest of 2026, based on a survey of 350+ public-company CEOs. The freeze is producing structural changes: entry-level job listings have dropped 30% since 2022, and middle management postings have fallen 42%. These shifts may be partially AI-driven, partially post-pandemic restructuring, and partially competitive mimicry — the data doesn't cleanly isolate any single cause.
What the Research Says
Academic research has been more measured than corporate announcements.
Yale University's Budget Lab found no overall employment change across AI-exposed occupations — challenging the narrative of mass displacement.
Oxford Economics found that "productivity growth hasn't accelerated in a way consistent with widespread labor replacement."
Stanford research identified a 16% employment decline among early-career workers in AI-exposed fields since ChatGPT's launch — suggesting that entry-level positions may be disproportionately affected even as overall employment remains stable.
The Brookings Institution identified 6.1 million U.S. workers who face both high AI exposure and low adaptive capacity — primarily in clerical and administrative roles. Approximately 86% of these workers are women.
Broader forecasts vary widely. Forrester projects 10.4 million jobs eliminated by 2030. The World Economic Forum estimates 92 million jobs displaced but 170 million created — a net gain of 78 million. The range between these projections illustrates how uncertain the impact remains.
The Naming Problem
The term researchers and journalists have increasingly applied to this phenomenon is "AI-washing" — the practice of attributing financially motivated decisions to AI capabilities that don't yet exist or aren't fully implemented. It mirrors "greenwashing" in sustainability: using a forward-looking narrative to justify present-day actions that serve different purposes.
The mechanism is straightforward. Citing AI as a reason for layoffs signals innovation and forward-thinking to investors. Block's stock rose 18% on its announcement. Citing weak demand or past overhiring does the opposite. As one Oxford Economics researcher noted, AI provides "the least bad reason" for cutting headcount.
This doesn't mean AI is having no impact on employment. The Stanford data on early-career workers, the genuine productivity improvements in coding and customer service, and the 17% of companies that report AI-driven headcount reductions all indicate real shifts. But the gap between the scale of layoffs attributed to AI and the scale of AI actually deployed remains significant.
What Happens Next
The March 24 hearing in the Anthropic-Pentagon case (covered in our separate analysis) may establish precedent for how government contracts interact with AI safety standards. But the employment question operates on a different timeline.
84% of CEOs acknowledge that meaningful AI ROI is a multiyear project. Meanwhile, 53% of investors expect payback within six months. That expectation gap — between the time AI needs to mature and the time markets give companies to show results — creates pressure to demonstrate AI impact through visible actions like layoffs before the technology can demonstrate impact through measurable productivity.
If current layoff rates hold, RationalFX projects 264,730 tech job losses by the end of 2026. Whether those losses reflect genuine technological displacement or the continuation of a narrative cycle — cut, regret, rehire — will likely depend on whether AI's performance catches up to AI's reputation before the talent pipeline does permanent damage.
The data, for now, points in both directions. The layoffs are real. The AI justification is often not. And the 162 WARN notices sitting in a New York filing cabinet — each one with an unchecked box — remain the most concise summary of the gap between what companies say publicly and what they're willing to put in writing.
