The ROI Fallacy: The Layoffs Paid for the Machines. The Machines Haven't Paid Anyone Back.

Here is the question the 2026 layoff wave never answered: if the people were cut to fund the machines, where is the return on the machines?

We now have a real answer, and it comes from inside the enterprise-research establishment, not from critics. In May 2026, Gartner published the result of surveying organizations that restructured around AI. About 80 percent reported workforce reductions as part of their autonomous-technology push. Then the analysts did the simple, brutal thing: they compared the companies that cut deeply with the companies that didn't, and looked at who actually got a return on their AI investment.

There was no relationship. Workforce reduction rates were nearly identical between the organizations reporting strong AI returns and those reporting modest or negative ones. Cutting people did not produce the payoff. In Gartner's own framing, the layoffs "may create budget room, but do not deliver returns." What did correlate with returns was nearly the opposite move: reinvesting in the workforce — organizations that used AI to amplify people, and spent on the skills and operating models to do it, were the ones that saw results.

Sit with the size of that. AI was the leading stated reason for layoffs in March and April of this year. The running count of 2026 layoffs where the employer named AI is nearly fifty thousand people — approaching the total for all of 2025 in half the time. Behind each of those numbers was a board deck that treated headcount reduction as the proof of AI transformation: we cut, therefore we transformed. The best available enterprise data now says the second half of that sentence never happened.

Did AI do this, or did we?

This is where the desk's standing question earns its keep, because the mechanism matters. An AI model did not identify fifteen percent of the workforce as redundant. Executives made a bet — in many cases before their AI deployments were mature enough to measure — that the machines would cover the gap. The layoff came first; the capability was assumed. That sequencing is a human decision with a human incentive behind it: a workforce reduction reads instantly on an earnings call, while an AI system's actual contribution takes years to prove. The cut is legible to Wall Street. The return is not. So the cut became the announcement, and the announcement became the strategy.

There is a second-order cost that the survey data only gestures at. When an organization cuts the people who understood why things work — why a supply chain failed in a particular way, why a client protocol exists — it does not just lose labor. It loses the context its AI systems needed to be useful. The models summarize; the veterans knew what the summary was missing. And sometimes the record is even blunter than the survey. When Meta cut roughly 8,000 people — about ten percent of staff — the public framing was an "AI efficiency push." At the internal town hall, by the reporting of it, Mark Zuckerberg told his own employees the layoffs were about capex, not AI productivity: headcount cut to help fund $135 billion of compute and hundred-million-dollar researchers. The 10-K's capex and cash-flow lines say the same thing without the quote. Our Meta dossier holds the full paper trail. One of the largest AI-attributed layoffs of 2026, characterized by its own chief executive as a financing decision wearing the technology's name.

The honest ledger, then. What is documented: the cuts happened at scale, AI was the stated reason, and the best available research finds no return attached to the cutting. What is not documented: that AI capability was ready to absorb the work — the null ROI result suggests, but does not prove, that in most firms it was not. We are not claiming the technology is useless; Gartner's same data shows returns flowing to companies that deployed it with their people rather than instead of them.

Which leaves the question where it belongs — not with the machine, but with the people who signed the layoff memos. If the data says cutting staff doesn't produce AI returns, and the cutting continues anyway, then the layoffs were never really about AI. They were about the budget room, and AI was the story that made the budget room look like the future.

We will keep the ledger as the filings come in. The claim taxonomy in our Layoffs by AI archive tracks exactly who said the machine did it — the company, the media, or no one. The gap between that claim and the measured return is now one of the clearest numbers on this desk.