Cloudflare cut 1,100 jobs in May. Same week, it posted its best quarter ever — $639.8 million in revenue, up 34% year over year. Matthew Prince called it an AI story anyway.
That’s the pattern now. Meta’s net income hit $26.8 billion in a quarter it also cut 8,000 roles. Block eliminated 40% of its workforce while projecting close to $12 billion in gross profit. Nobody here was struggling. Everybody fired people anyway.
| Company | Cuts | Financial Context | Stated Reason |
|---|---|---|---|
| Cloudflare | 1,100 roles | $639.8M revenue, +34% YoY | AI shift; mostly hit internal audit and middle management |
| Meta | 8,000 roles | $26.8B quarterly net income | Reallocated 7,000 staff into core AI engineering |
| Block | 40% of workforce | $12B projected gross profit | Framed as a productivity resize |
The Reality of AI-Washing in 2026 Tech Layoffs
Challenger, Gray & Christmas has tracked AI as the top-cited layoff reason for four straight months in 2026. SHRM has a name for what’s underneath that number: AI-washing. Dressing up an ordinary headcount cut in language that sounds like strategy.
Wall Street rewards ambition. It punishes bad planning. So executives read from whichever script keeps the stock price green.
Two Trackers, One Set of Layoffs, a 48-Point Gap
Here’s where the story falls apart. One tracker counts loosely — SkillSyncer’s 2026 tracker found 56% of tracked layoff events named AI, automation, or machine learning somewhere in the announcement.
Another counts strictly. TechJack Solutions labeled just 7.8% of those same events “AI-Direct” — meaning the company explicitly named AI as the actual driver. The rest? Plain business-cycle correction, 84% of it.
Same layoffs. A 48-point swing depending on who’s counting. That gap is the story. Not the headline percentage.
Why This Is Happening Right Now
It’s not just a messaging choice. It’s a capital reallocation emergency. Meta alone has guided investors to $115 billion to $135 billion in 2026 capital expenditures — nearly double the $72.2 billion it spent in 2025 — most of it locked into a multiyear Nvidia deal covering Blackwell and Rubin GPUs. Borrowing hasn’t gotten any cheaper this year, so that money has to come from somewhere internal.
Payroll is the fastest lever. Georgetown’s Jason Schloetzer put it plainly: executives often cite AI when the real problem is cash flow, and they need capital freed up for infrastructure they’ve already committed to. Part of that squeeze traces back to the rising cost of running AI models at scale, a line item that rarely makes the press release.
Cognizant’s chief AI officer told Nikkei Asia something similar — AI sometimes becomes the scapegoat when a company simply overhired and wants to resize. Sam Altman used SHRM’s own term for it. Wharton’s Peter Cappelli went further: many companies citing AI hadn’t deployed it at scale yet. They were hoping.
It’s Also Just Copying Homework
Stanford’s Jeffrey Pfeffer has spent years studying what he calls social contagion in corporate layoffs — companies imitating each other regardless of whether the cuts actually help. His research found layoffs routinely fail to raise stock prices or cut real costs once severance and lost productivity are counted in. Companies do it anyway, largely because a board starts asking why every competitor is announcing cuts and theirs isn’t.
That framing fits 2026 uncomfortably well. A handful of companies say “AI efficiency” first. Everyone else reaches for the same phrase within a quarter, whether or not their situation resembles the original case at all.
The Departments Getting Cut First
Cloudflare’s own language gives it away. Prince wrote that most of the cuts hit “measurers” — middle management, finance, legal, internal auditing. Not engineers. Not the people building AI products. That’s the same measurer-class shrinkage showing up across the wider industry, not just at Cloudflare.
Meta followed a similar script. It cut roughly 8,000 roles while quietly moving 7,000 employees into AI-focused positions — a reshuffle tied directly to Meta’s ballooning AI spending and the pressure to show something for it.
Does Any of This Actually Pay Off?
A Gartner study surveying 350 executives at billion-dollar-plus companies found no clean link between cutting headcount for AI and getting a return on it. The companies seeing real gains weren’t the ones firing people and calling it efficiency. They were training existing staff to build their own automations.
Gartner expects more than 40% of agentic AI projects to be scrapped by the end of 2027. Costs spiral. The business case never solidifies. Dorsey framed Block’s cuts as a productivity story, but what actually changed at Block looks a lot closer to a straightforward resize.
Who Pays For This Later
Entry-level roles are disappearing fastest, and most coverage skips right past it. Stanford HAI’s AI Index found employment among developers aged 22 to 25 down nearly 20% since 2024, while headcount for developers over 30 continued to grow at the same companies.
That’s not AI replacing engineering. That’s a company deciding which rung of the ladder it can afford to remove — visible now in how entry-level hiring has changed across tech.
The math gets worse a few years out. Junior engineers are how a company grows the people who’ll manage its legacy systems once senior developers retire. Cut that bottom rung today, and there’s nobody trained to inherit the codebase in five years. AI agents don’t refactor systems they were never taught. They pattern-match around problems humans haven’t diagnosed yet.
The reset isn’t isolated to a company or two anymore. It’s showing up across the broader labor reset at Meta and Microsoft and spreading through the rest of the sector, quarter after quarter.
The next time a company announces layoffs and blames AI, the real question isn’t whether AI is involved. It’s whether they can name the specific task the software now does — or whether “AI” is just this year’s word for a decision that would’ve needed a different explanation five years ago.
Related: An AI Startup Posted a Job That Sounded Fake — Then 100,000 People Applied
