AI Became the Top Reason Companies Gave for Layoffs in 2026. The Bigger Risk is What It's Doing to Hiring.

By Courtney Turrin · July 28, 2026 · 13 min read read

Young professionals in a modern office setting — Photo by cottonbro studio, courtesy of Pexels.

AI is cited as the top reason for layoffs, but hiring has stalled more significantly.

In 2025, employers cited artificial intelligence in 54,836 announced layoffs, about 5 percent of all job cuts tracked that year by the outplacement firm Challenger, Gray & Christmas. By June 2026, that figure had climbed to 101,743 announcements, roughly 23 percent of all cuts, and AI had led every stated reason for job cuts for four straight months (Fig. 1).

[datawidget key=ai-cited-layoffs]

That number describes what companies said in press releases, shareholder letters, and public statements. It is not a measurement of what actually happened inside those companies. Challenger's report tracks announced reasons, self-selected by the employer, not verified causes confirmed by an outside party.

The labor data underneath the announcements tells a narrower, messier story than the headlines suggest. Independent economists who have gone looking for AI's fingerprint in employment numbers keep coming back with the same finding: something is happening, but it does not look like the sweeping automation story companies are telling. Meanwhile, the market appears to reward one version of that story more than the alternatives, which may explain why so many companies are reaching for the same explanation regardless of whether it's the true one.

Layoffs didn't spike. Hiring stalled.

The instinct is to assume a wave of AI layoffs means firing accelerated. It didn't. The national layoff and discharge rate sat at 1.1 percent in May 2026, well within its recent historical range. What changed is hiring. Employers announced 507,647 planned hires for all of 2025, down 34 percent from the year before and the lowest annual total since 2010 (Fig. 2).

[datawidget key=layoffs-vs-hires]

Quitting slowed at the same time, which is its own signal. In the information sector, the quit rate fell from 1.9 percent to 1.1 percent between May 2022 and May 2026, a decline Indeed's Hiring Lab reads as workers' confidence in finding something better having fallen sharply over four years. Economists have taken to calling this a low-fire, low-hire economy. It is a real phenomenon. It is not the same thing as machines replacing people.

Researchers keep looking for AI's fingerprint. Mostly, they can't find it.

There is one data point that does support the AI story on its face: information-sector layoffs, which cover most of the tech industry, rose to a rate roughly five times the national average over the past year, the sharpest increase of any sector tracked by the government's Job Openings and Labor Turnover Survey.

Yale's Budget Lab examined that anomaly directly, publishing both its analysis and the underlying data. They confirmed the rise is real, not statistical noise. But they also found that hires in the information sector rose alongside layoffs over the same period, and two independent measures, unemployment insurance claims and the sector's overall unemployment rate, do not show the kind of sustained damage a genuine AI-driven contraction would produce. Their reading: the numbers are "consistent with individual firms firing workers who are being hired by other firms, but not with a decline in the labor needs of the entire sector." Occupational churn in the sector, another marker economists watch for disruption, has stayed flat.

Other independent analyses have yielded the same insight. Researchers at the Federal Reserve Bank of New York concluded that AI is not the main driver of the current slowdown in hiring, even in occupations considered highly exposed to automation. The research firm Gartner found that companies citing AI as a reason for job cuts were not seeing better financial returns than companies that didn't, with one of its analysts describing the pattern as a "one-time exercise" rather than evidence of structural transformation. Oxford Economics went further, writing in a January 2026 briefing that firms don't appear to be replacing workers with AI on a significant scale.

Where AI's effect does show up: the bottom rung of the career ladder

None of this means AI has no measurable footprint on employment. AI's broader economic footprint, investment, product adoption, changes to how work gets done, is large and growing. What's specifically narrow is the causal link between AI and mass layoffs. The one place that link does show up clearly in the data is young workers in their first jobs.

Stanford economists Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen used payroll data from ADP, the country's largest payroll processor, to track employment by age and occupation since generative AI tools became widely available in late 2022. They found that early-career workers ages 22 to 25 in the most AI-exposed occupations saw 16 percent relative employment declines, controlling for firm-level shocks, while employment for more experienced workers in the same jobs stayed stable. The effect concentrates specifically in roles where AI automates a task rather than assists with it, and the pattern holds even after excluding technology companies and remote-friendly jobs from the analysis. For software developers age 22 to 25 specifically, headcount is down roughly 20 percent since late 2022.

The authors have also tested, and largely ruled out, the most obvious alternative explanation. In a February 2026 follow-up, they report that while rising interest rates affected employment broadly, the evidence does not support rates as an explanation for the specific, disproportionate decline in entry-level hiring within AI-exposed occupations. They are more cautious about timing: with the fullest set of controls applied, the decline in AI-exposed occupations only becomes statistically significant starting in 2024, and they note the earliest drop in software-developer hiring, right after ChatGPT's late-2022 release, likely reflects something other than AI given how quickly it appeared.

On the evidence so far, that is not a plausible explanation for a company halving its entire workforce.

The market rewards the story, which is its own kind of evidence

If the data linking AI to large-scale layoffs is this thin, why has the explanation become so common anyway? Part of the answer sits in how financial markets respond to the word.

A meta-analysis of 34,594 layoff announcements across 78 published studies found that the average stock market reaction to a layoff announcement is significantly negative. But the average conceals a sharp split. Investors do not punish layoffs that are framed as proactive, forward-looking management decisions, such as cost discipline ahead of a downturn. They do punish layoffs framed as reactive, a response to falling demand or a business in trouble. The size of the penalty also scales with the size of the layoff.

That split gives companies an incentive to choose their framing carefully, and "AI transformation" is one of the few available framings that reads as proactive rather than defensive. Brookings researcher Molly Kinder has made the same point directly: telling investors a layoff is about AI conveys a more positive message than telling them it's about weak demand or a hiring mistake made two years earlier.

These studies can't fully separate framing from firm health: a company already doing well may find it easier to sell 'proactive' and be believed. Block's case, examined below, is useful precisely because the public numbers let both explanations be checked against the same facts.

One company, two explanations, eleven months apart

Block, the financial technology company behind Cash App and Square, offers the clearest single illustration of how much the explanation can shift without the underlying facts changing.

In March 2025, when Block cut 931 employees, CEO Jack Dorsey emailed staff a plain accounting of the reasons: 391 people let go for strategic realignment, 460 for performance, 80 managers cut to flatten the org chart. He was direct about what wasn't driving the decision. "None of the above points are trying to hit a specific financial target, replacing folks with AI, or changing our headcount cap," he wrote.

Eleven months later, cutting nearly half the company, roughly 4,000 people, from more than 10,000 employees to under 6,000, the explanation had changed entirely. "Intelligence tools have changed what it means to build and run a company," Dorsey wrote in his shareholder letter. "A significantly smaller team, using the tools we're building, can do more and do it better." He was explicit that the business itself was healthy: "We're not making this decision because we're in trouble. Our business is strong." Block's stock rose as much as 24 percent on the announcement.

The public numbers complicate the framing. Block's headcount had roughly tripled from 3,835 employees at the end of 2019 to over 10,000 before the February 2026 cuts, a pandemic-era hiring surge now being unwound and described as a technology strategy rather than a correction. Mizuho Americas analyst Dan Dolev told the Wall Street Journal that, in his view, the vast majority of the cuts were probably not due to AI. Block is not an isolated case: Salesforce's CEO has said he "needs less heads" after cutting 4,000 customer-support roles and attributing the reduction to AI.

None of this means AI played no role at all in either round of Block's cuts. It means the company had two different stories available to tell about very similar underlying decisions, and it chose the one markets responded to better.

This kind of comparison is possible for Block mainly because its explanations happen to be on the public record in unusual detail. For most companies, a shareholder letter or a press release is the last word anyone outside the company sees. Severance packages are typically conditioned on the departing employee signing a release of legal claims, waiving the right to sue over the reasons behind their termination. That's a standard feature of how layoffs work, not evidence of wrongdoing. But it means the employer's own explanation rarely gets tested in a courtroom or through discovery, which is a structural reason the public conversation about why any layoff happened runs almost entirely on companies' say-so.

The same tools are fueling a surge in people working for themselves

AI is not only something companies cite when cutting workers. The same tools are changing the other side of the labor market, by lowering the cost of starting a business alone.

The Nasdaq Economic Institute, in its first published research, analyzed Census Bureau business-formation data and found that applications from one-person businesses have risen more than 20 percent since early 2025, even as applications from companies likely to eventually hire employees have stayed roughly flat. Nearly half of that growth in solo formations came from industries with high AI adoption, including technology, finance, and professional services. Nasdaq's chief economist, Phil Mackintosh, called the pattern "a meaningful signal for where AI's economic impact may show up first."

It is not clear from the data whether this shift is being driven mainly by necessity or by opportunity, and it is likely both. Some share of people going solo are almost certainly doing so because the corporate hiring market described earlier in this piece has made reentry slow and uncertain, and self-employment is the option left standing. Others appear to be choosing it on its own merits. MBO Partners, which has surveyed the independent workforce annually for 15 years, found in its 2025 report that the number of independent workers earning over $100,000 a year rose 19 percent to 5.6 million, nearly double the 2020 figure, and that 59 percent of independent workers say they earn more than they did in a traditional job.

That last figure deserves scrutiny rather than a straight read. Self-reported earnings comparisons from people who chose to leave traditional employment are not a random sample, and the same MBO survey found that only 41 percent of solopreneurs rely on their business as their primary source of income, meaning many respondents are describing supplemental income, not a full replacement for a salary.

"Become your own boss" undersells what that actually requires. Gallup's American Job Quality Study found that 53 percent of self-employed workers have no retirement plan at all, compared with 19 percent of traditional employees, and that self-employed workers are nearly three times as likely to have no health insurance. The same study found self-employed workers log longer average hours than employees. And going independent is not a guaranteed outcome even on its own terms: Bureau of Labor Statistics data on business survival show that nearly half of new businesses close within five years, a rate that has held steady for decades and applies before AI tools entered the picture.

Put together, this is not simply a story about corporate layoffs pushing displaced workers into a worse alternative, nor is it a story about AI unlocking effortless entrepreneurial upside. It's a labor market where the traditional path of one employer, one paycheck, and one benefits package is being replaced, for a meaningful share of workers, with something that offers more autonomy and, for a minority, more income, in exchange for less security. Whether that trade is a net improvement depends entirely on which side of the distribution a given worker lands on, and the data available right now can't say in advance which side that will be.

Whatever the real reason, getting rehired has gotten much harder

Regardless of how any individual layoff should be explained, the people on the losing end of it are facing a labor market that punishes displacement more severely than it did two years ago.

The share of unemployed workers who have been out of work 27 weeks or longer, the standard definition of long-term unemployment, reached 27.3 percent in June 2026, up from 21.5 percent in July 2024 (Fig. 3).[datawidget key=long-term-unemployment] The information sector, which includes most of the roles at the center of the AI-layoff debate, carries the longest mean unemployment duration of any industry tracked in the government's occupational data.

The group most affected by AI keeps growing every year

Set aside for a moment whether any individual company's stated reason for its layoffs holds up. The more consequential pattern in this data is what it suggests about where the labor market is headed.

Right now, AI's confirmed effect on employment is concentrated in one place: workers in their early twenties entering AI-exposed technical occupations for the first time. That is a leading indicator, not a contained one. Every year, another cohort of new graduates enters those same occupations, and if the entry-level effect documented by Brynjolfsson and his colleagues persists, it compounds with each hiring cycle rather than resolving on its own. At the same time, the broader hiring market these workers are entering has already slowed by every measure in this piece, and the people it does eject are taking longer than they have in years to find the next job.

The self-employment track absorbs some of that pressure, and for a portion of workers it is a genuine improvement over what they had. But it is not a like-for-like replacement for a corporate salary and benefits, and the data available does not support treating it as a safety net that will catch everyone the traditional labor market is failing to reabsorb.

None of that requires believing any particular company lied about its motives for a specific round of cuts. It requires taking seriously what happens when the group facing the clearest AI-driven employment effect keeps growing every year, the broader hiring market keeps cooling, and the alternative path available to displaced workers comes with a coin flip's odds of surviving five years. That is the story sitting underneath the announcements, whether or not any single one of them is being told straight.