The Layoffs Nobody Can Move Sideways From

The short version: The May 2026 tech layoffs are not a downturn. Profitable, growing companies like Meta, Intuit, Cisco, Cloudflare, and Coinbase cut more than 92,000 jobs in five months to restructure around AI. Because every firm gets the same AI at the same time, the lateral job moves that softened past layoffs are disappearing. The companies that win will redeploy the people they keep onto judgment work, not simply shrink.

Key takeaways

• Over 92,000 tech jobs vanished in five months, led by profitable, growing companies, not struggling ones.

• When every competitor gets the same AI at once, eliminated roles disappear industry-wide, closing the usual sideways exit.

• Cutting volume work is the easy half; redeploying kept staff onto judgment work is the half that actually compounds.

• Run the replacement exercise: move yourself from bee to beekeeper before the company decides for you.

---

There is a number from this past month that I cannot stop thinking about. More than 92,000 tech jobs have vanished in five months, and the companies doing the cutting are not the companies you would expect. They are not bleeding cash. They are not unwinding pandemic over-hiring. They are profitable, growing, and restructuring on purpose.

Meta began company-wide layoffs on May 20th, cutting roughly 8,000 people, about ten percent of its workforce, with more cuts planned for the back half of the year. In the same breath, the company is spending somewhere between 115 and 135 billion dollars on AI infrastructure in 2026. Read that again. They are removing humans and adding compute, and they are doing both as fast as they can.

It is not just Meta. Intuit is letting go of about 3,000 employees, roughly seventeen percent of staff, to redirect resources into baking AI into its products. Cisco announced 4,000 job cuts while openly saying automation is replacing human work in certain functions. Cloudflare is cutting more than 1,100 roles, about twenty percent of its people, after its internal AI usage jumped more than 600 percent in three months. Coinbase trimmed roughly fourteen percent of its workforce, with CEO Brian Armstrong saying plainly that the company needs to become AI-native.

This is not a recession. It is a redesign.

Every layoff cycle I have lived through told the same story. A company over-hired, the economy turned, and the cuts corrected the mistake. Then the economy recovered and the jobs came back somewhere, even if not at the same firm. Workers moved sideways. A laid-off engineer at one company became a hired engineer at a competitor. The pain was real but the labor market absorbed it.

What is happening now breaks that pattern, and it breaks it in a specific way that I think most leaders have not fully absorbed. As one analysis of the May wave put it, when the same AI capabilities are available to every company in an industry simultaneously, the restructuring happens simultaneously. The role categories that get eliminated get eliminated industry-wide, at the same time, which blocks the lateral moves that absorbed displacement in every previous cycle.

That is the quiet horror in the data. The exit ramp is closing. If your job is automatable, and the company down the street has the same models you do, then the company down the street is not hiring for your role either. They just cut it too.

It helps to name the parallels, because this is not three or four companies acting strangely. It is a coordinated reordering of entire industries. Meta, Intuit, Cisco, Cloudflare, and Coinbase are not in the same business, yet they made the same call in the same month for the same reason. When you see that many unrelated firms reach an identical conclusion at once, you are not looking at company-specific decisions. You are looking at an industry-wide repricing of human labor against machine labor, and it is happening faster than any retraining program, any university curriculum, or any government policy can respond to. That speed mismatch is the part that should worry leaders and workers alike, because the institutions we usually rely on to cushion these transitions move in years while this is moving in quarters.

The capability arrived before the imagination did

Here is where I want to be careful, because the doom narrative is easy and mostly useless. The honest reading of this moment is not that AI is destroying work. It is that AI is destroying a particular kind of work faster than most organizations can imagine new work to put people on.

I lived a version of this years ago at Alibaba during Double 11, the largest shopping festival on earth. We needed hundreds of thousands of marketing banners in a matter of weeks, each one tailored to a different product, a different audience, a different moment in the funnel. There was no human team on the planet that could produce that volume at that quality in that window. So we used generative systems to do it, and they produced banners at a scale and speed that made the old way look quaint. That was the moment I understood AI was not another efficiency tool. It was a different category of leverage.

But here is what I also learned. The humans did not disappear. Their work moved up the stack. They stopped making banners and started directing the systems that made banners, judging which ones worked, deciding which campaigns to run, owning the outcome. The volume work went to the machines. The judgment work stayed with the people, and there was more of it than before, not less.

That is the difference between the companies that will come out of this wave stronger and the ones that will just be smaller. Cutting headcount because AI can do the volume work is the easy half. The hard half is redeploying the humans you keep onto the judgment work that AI cannot do. Most of the companies in the May layoff numbers have done the first half. I have seen very few do the second.

The Hive Structure, and why most org charts are wrong for this

I describe the right structure for an AI-native company as a hive, and it has only two roles. There are bees and there is the beekeeper. The bees are the AI systems, executing volume work at machine speed, tireless and fast and narrow. The beekeeper is the human, directing the bees, judging their output, and making sure the whole thing serves a purpose the bees cannot understand on their own.

What you do not do is add tiers. You do not build a layer of middle managers whose job is to manage the people who manage the bees. The whole point of the hive is that one beekeeper can direct an enormous amount of bee labor. That is the leverage. Every extra human tier you insert between the beekeeper and the outcome is friction that eats the leverage you just bought.

When I look at the May layoffs through this lens, I see companies that understand they have too many bees doing work the machines now do better. What I do not see is companies that have figured out how many beekeepers they need, or who those beekeepers should be, or how to turn a former bee into a beekeeper. That last move, helping a person climb from doing the work to directing the work, is the single most valuable thing a leader can do right now, and almost nobody is investing in it.

The productivity numbers hide the strain

There is a second dataset from this same window that complicates the easy story in both directions. The workforce studies coming out this spring show that AI is, in fact, making people more productive. A Gallup analysis found that 65 percent of employees at AI-adopting organizations say it has improved their productivity. A Federal Reserve and NBER study of corporate executives found positive labor productivity gains that vary by sector and are expected to strengthen through 2026, with the largest effects in high-skill services and finance.

So the tools are working. People are getting more done. And yet the same research shows the catch. Analysis of workplace data found that AI is significantly increasing the intensity and scope of employees' work and raising the threat of burnout. Worker confidence in AI actually fell, and a large share of employees fear automation could replace their jobs within two years.

Put those facts next to the layoffs and a clearer picture forms. AI lets a smaller group of people produce what a larger group used to. Companies are responding by keeping the smaller group, cutting the rest, and loading more onto the survivors. That is a coherent strategy for one or two quarters of margin. It is a terrible strategy for retaining the very people you decided were worth keeping, because you are burning them out at the exact moment they have the most leverage to leave.

The companies that get this right will use the productivity gains to do more, not just to do the same with fewer. There is a version of this where AI absorbs the drudgery and humans get pointed at higher-value problems that were always sitting unaddressed because nobody had the capacity. And there is a version where AI absorbs the drudgery, the company pockets the savings, and the remaining humans inherit a heavier load and a thinner reason to stay. The technology does not pick between those outcomes. Leadership does.

What this means if you still have your job

If you are reading this with a job intact, the data is not telling you to relax. It is telling you to do an exercise I call replacement. Constantly replace yourself on the tasks a machine can do, so that you become irreplaceable on the tasks that remain.

Sit down and list every task you did last week. Be ruthless. Next to each one, write whether a capable AI system, the kind that already exists today, could do eighty percent of it. The tasks that get a yes are not your job security. They are your exposure. The tasks that get a no, the ones requiring judgment, relationships, taste, accountability, and the ability to decide what should be done rather than how to do it, those are where you want to spend more of your time, starting now.

The workers who survive this are not the ones who avoid AI. They are the ones who climbed from bee to beekeeper inside their own role before the company made the decision for them.

What this means if you run the company

If you are making these cuts, or you are about to, I want to offer one reframe that has nothing to do with sentiment and everything to do with results. Headcount reduction is a cost story. It shows up once, it makes a quarter look better, and then you are running a smaller version of the same company. Capability redeployment is a growth story. It is harder, it is slower, and it is the only version of this that produces a company that does more next year than it did this year.

The firms in the May numbers that treat AI as a way to shrink will shrink. The firms that treat it as a way to point their best people at problems they never had the capacity to attack will compound. Same technology, same layoffs, completely different outcome, and the difference is entirely in whether leadership did the hard half.

So here is the question I will leave you with. When your company finishes cutting the roles AI made redundant, what exactly are the people you kept going to do that they could not do before? If you cannot answer that in a sentence, you have run a cost-cutting program and called it a transformation. Which one is it?

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.

---