Companies Are Firing People for a Productivity Gain They Have Not Measured

Tech layoffs are running at 1,115 a day in 2026, and almost every announcement blames AI. A new Gartner study of 350 companies found the firms cutting the most got no better financial returns than the firms cutting the least. The layoffs are real. The justification often is not. Here is how to tell the difference inside your own company.

The number is hard to look away from. As of mid-June 2026, 247 layoff events have displaced nearly 184,000 workers this year, an average of 1,115 jobs lost every working day, nearly double the 564-a-day pace of 2025. And in announcement after announcement, executives name the same cause. Not restructuring. Not overhiring. AI.

What makes 2026 different from past downturns is that companies are saying the quiet part out loud. Where firms once hid cuts behind the language of efficiency, many now openly credit artificial intelligence. Meta cut roughly 8,000 roles in May while redirecting resources toward AI, even as it prepares to spend over $100 billion this year on data centers. Oracle executed the largest single cut of the year, around 30,000 jobs, shortly after strong earnings. This week alone, ServiceNow and Salesforce both announced cuts tied to expanding AI use, and GitLab restructured explicitly for what it called the agentic AI era.

I work with executives on AI implementation, so I want to separate two things that are getting collapsed into one. The technology is truly changing how work gets done. That part is true. But the claim that AI is driving these specific cuts, and that the cuts are paying off, is a claim the data does not support. And the gap between those two things is where a lot of careers are being spent.

The study the press releases left out

In May 2026, Gartner published findings from a survey of 350 executives at billion-dollar companies, all of them already deploying AI. The result was blunt. Eighty percent had cut headcount, some by as much as 20 percent. And the companies that cut the most showed nearly identical financial returns to the ones that cut the least. In several cases, the firms that cut less performed better.

Helen Poitevin, the Gartner analyst who led the work, said it plainly: chasing value only through headcount reduction is likely to lead most organizations down a path of limited returns.

That phrase, value only through headcount reduction, is the whole problem in five words. Cutting people is the easiest lever a leader can pull, because it is legible, immediate, and rewarded by the market on the day it is announced. Redesigning how work flows so the AI actually produces value is slow, messy, and pays off later if at all. Faced with that choice, a lot of executives reach for the lever they can pull this quarter and tell themselves the harder work will follow. The Gartner data is what happens when the harder work does not follow: the cost comes out, the value never comes in, and the company ends up smaller without ending up stronger.

This is not an isolated finding. In 2025, MIT's Project NANDA studied more than 300 disclosed enterprise AI initiatives and found that 95 percent were generating zero measurable return, despite tens of billions in spending. And in a randomized controlled trial the same year, the research group METR found that experienced developers using AI tools took 19 percent longer to finish tasks, even though they believed the tools had sped them up by 20 percent. The gap between how fast people felt and how fast they actually were is the same gap between the layoff narrative and the layoff results.

Put these three findings together and a mechanism appears. Companies are cutting workers based on the anticipated productivity of AI systems that have not delivered that productivity at scale. They fired first. The tools were supposed to catch up. Often they have not.

The term for this is AI washing

The polite phrase for the gap between the stated reason and the real one is AI washing, and it has moved from skeptic blogs into mainstream acknowledgment. In February 2026, OpenAI's own CEO acknowledged that companies are blaming AI for cuts they would have made anyway. A National Bureau of Economic Research working paper found that 90 percent of executives say AI has had zero employment impact at their own companies, even as their peers make AI the headline of every layoff.

The sharpest example is Block. In a 2025 memo, CEO Jack Dorsey was explicit that the cuts were not about replacing people with AI. By 2026, his shareholder letter attributed the elimination of roughly 4,000 positions to AI tools that had made the roles unnecessary. The business pressure had not changed. The frame had.

Why does the frame matter so much? Because Wall Street rewards it. When a profitable company cuts staff and cites efficiency, that reads as routine cost-cutting. When it cuts staff and cites AI, that reads as strategic vision, and the stock often moves accordingly. AI has become the most investor-friendly explanation available, which is exactly why it should be interrogated rather than accepted.

Where this gets dangerous: the on-ramp is disappearing

There is one part of this story that is not washing, and leaders need to see it clearly because it is the part with the longest tail. The entry-level career on-ramp is collapsing. Stanford's Digital Economy Lab found that employment for software developers aged 22 to 25 fell nearly 20 percent from its 2022 peak, even as employment for developers over 26 grew.

The mechanism is structural and worth understanding. AI tools let senior engineers absorb the boilerplate coding, routine testing, and scripted debugging that junior roles were historically built around. The junior on-ramp is vanishing not because companies stopped needing engineers, but because the tasks that gave juniors their first five years of experience are now handled by tools in senior hands.

Here is the part almost nobody is pricing in. Destroying the entry-level pipeline does not just hurt this year's graduates. It shrinks the supply of experienced engineers available five to ten years from now, at the exact moment companies expect AI-augmented seniors to be their most valuable people. You are eating your seed corn and calling it a harvest.

The Replacement Exercise, run correctly

I teach a practice I call the Replacement Exercise: constantly replace yourself on the automatable parts of your job so you become irreplaceable on the parts that remain. The goal is to move from being a bee, doing volume work, to being a beekeeper, directing it.

What companies are doing right now is the corrupted version of that exercise. They are replacing the bees and assuming the honey appears on its own. But the Replacement Exercise only works when you redesign the human's role around higher-value work, not when you simply delete the human and hope the AI covers it. The Gartner data is what corrupted replacement looks like at scale: headcount down, returns flat, because nobody redesigned the work, they just removed the people.

Done right, this looks completely different. You identify the automatable volume, you move it to AI, and you redeploy the human toward judgment, relationships, and the decisions that AI cannot own. Headcount might still come down, but output and returns go up, because you changed the structure of the work rather than just its cost line.

What to do before you approve a single cut

If you are a leader staring at a proposal to reduce a team because of AI, ask one question before you sign. Have we measured the output gain on this exact workflow, or are we forecasting it? Forecasting is gambling with other people's careers and, as the data shows, often with your own returns.

Then ask a second question. If we remove these people, who absorbs the judgment work they were doing, and is the AI actually capable of that work today, or are we assuming it will be next quarter? Forrester reported in early 2026 that many companies announcing AI-related layoffs do not have mature, vetted AI applications ready to replace the roles. They cut the human and left the work uncovered.

The reputational cost nobody is modeling

There is a second cost to AI washing that does not show up in the Gartner returns data, and it compounds quietly. When a company cuts staff, blames AI, and then cannot deliver the promised efficiency, it does not just fail to gain. It loses trust, internally and externally, in ways that are hard to win back.

Inside the company, the people who remain are not fooled. They watched colleagues leave under an AI banner, and they can see whether the AI actually absorbed the work or whether it simply landed on them. When it lands on them, which the Forrester finding suggests is common, you have not made the organization more efficient. You have made it more overloaded and more cynical, and you have taught your best people that the company will dress up a cut as a vision whenever it is convenient. That lesson follows them to their next interview, and the strongest performers are exactly the ones with somewhere else to go.

Externally, the market is starting to notice the pattern too. When a National Bureau of Economic Research working paper finds that 90 percent of executives say AI has had zero employment impact at their own companies, even as those same executives' peers headline every layoff with AI, a credibility gap opens. The first wave of AI-branded cuts got rewarded by investors who read them as forward-thinking. The second and third waves will be read more skeptically, because the returns are not materializing on schedule, and Wall Street eventually prices in a story that keeps not coming true. The company that cried AI does not get the same benefit of the doubt the fourth time.

There is a clean test coming. Gartner predicts that by 2027, half of the organizations that planned to slash their service workforce through AI will abandon those plans, having missed their targets. When that data lands, it will tell us how many of today's cuts were strategy and how many were story.

Until then, the responsible position for any leader is simple. AI is real, the productivity is sometimes real, but the case for wholesale cuts has not been demonstrated, and three independent research bodies say so. So before your next workforce decision, ask yourself plainly: are you cutting because the AI is doing the work, or because the AI is a good thing to say while you cut? Your team can tell the difference. Eventually, so can your investors.

Sharon Gai is an AI transformation strategist, keynote speaker, and author of How to Do More with Less Using AI. She advises Fortune 500 companies on AI adoption and organizational redesign.