Nearly 80,000 tech workers have lost their jobs in the first quarter of 2026. Almost half of those cuts, according to the companies doing the cutting, are because of AI.

The companies announcing these layoffs are simultaneously spending close to $700 billion on AI infrastructure this year. Amazon alone plans to spend $200 billion while cutting 16,000 people. Meta is pouring up to $135 billion into AI while trimming headcount. The message to investors is clear: we’re so committed to AI that we’re reshaping the entire workforce around it.

MIT’s “GenAI Divide” report, published in August 2025 after surveying 153 business leaders and analysing 300 public AI deployments, found that 95% of companies investing in AI are seeing zero measurable return. Separately, a National Bureau of Economic Research survey of C-suite executives across the US, UK, Germany and Australia found that nearly 90% said AI had no impact on workplace employment over the past three years.

So either these companies have cracked something that almost none of their peers have, or something else is going on.

The convenient narrative

Even Sam Altman thinks it’s the latter. At the India AI Impact Summit in February, he told CNBC-TV18: "There’s some AI washing where people are blaming AI for layoffs that they would otherwise do." He later sharpened the point: "Almost every company that does layoffs is blaming AI, whether or not it really is about AI."

When the CEO of OpenAI, the person with more to gain from the “AI is transforming everything” narrative than almost anyone alive, is calling this out, you should probably pay attention.

Bloomberg has run multiple investigations into what it calls the “AI-washing of job cuts.” Their March reporting described it bluntly: companies are “exploiting fear to dress up old-fashioned cost-cutting as technological futurism.” The layoff tracker TrueUp counted over 59,000 tech jobs gone by early March, with the number accelerating through April. But when Bloomberg dug into Block’s 4,000-person cut — which Jack Dorsey attributed to AI capabilities — they found it looked a lot more like reversing pandemic-era overhiring.

None of this is a conspiracy. It’s just how incentives work. A CEO who announces layoffs because “we over-hired in 2021 and margins need fixing” gets a sympathetic nod from analysts and a modest stock bump. A CEO who announces the same layoffs because “AI is transforming how we work” gets a 22% stock surge, a Wired profile, and an invitation to speak at Davos.

What the numbers actually show

The mismatch between narrative and reality is stark once you lay it out.

Of 108,435 job cuts recorded in January 2026, AI was explicitly cited as a reason in roughly 7,600 cases — about 7% of the total. The rest were attributed to the usual reasons: restructuring, cost reduction, strategic realignment. The kind of corporate housekeeping that happens every year, cycle after cycle, with whatever the fashionable explanation happens to be.

In 2001, it was “the internet didn’t deliver.” In 2008, it was “global financial conditions.” In 2023, it was “post-pandemic normalisation.” In 2026, it’s AI. The layoffs happen regardless. The story changes.

I don’t say this to dismiss AI’s impact on work — it’s real, and I’ve written extensively about it. I watched Block demonstrate what genuine AI-driven restructuring looks like when Dorsey showed specific tools, specific output gains, and specific capability thresholds that triggered the decision. That article started with me planning to write about AI-washing, but Block’s numbers were too clean and too specific to ignore. They might actually be the real thing.

The problem is that for every Block, there are a dozen companies where the AI story doesn’t survive contact with the evidence. Where there’s no named tool, no measurable output gain, no specific capability that crossed a threshold. Just a press release, a narrative, and a stock price that cooperates.

The $700 billion question

This is where it gets genuinely strange. These same companies are spending extraordinary amounts on AI. Amazon, Microsoft, Alphabet and Meta are collectively approaching $700 billion in AI infrastructure spending this year. Data centres, chips, servers, compute — the physical machinery of the AI era.

Spending money on AI infrastructure and actually using AI to transform your workforce are different things entirely, though. You can pour billions into GPU clusters and still have no idea how to redesign a single business process around AI capabilities. The infrastructure is the easy part, relatively speaking. It’s expensive and it’s visible, which makes it perfect for earnings calls. The hard part is figuring out which roles AI actually changes, how to redeploy people, how to redesign workflows. That requires the kind of patient, unglamorous operational work that doesn’t photograph well.

And this is the bit that makes me uneasy. The AI infrastructure spending is real. The job cuts are real. But the connection between the two is, in most cases, paper-thin. Companies are spending on AI over here and cutting people over there, and the narrative stitching them together is doing most of the heavy lifting.

Who this actually hurts

The cynical reading is that none of this matters much — companies have always found reasons to cut staff, the reasons are always partially true and partially convenient, and the market allocates accordingly. And there’s something to that.

AI-washing does real damage, though, in two specific ways.

It corrodes trust inside the organisations doing it. I wrote recently about the trust trap that forms when companies frame AI deployment as headcount reduction. When remaining employees hear that AI justified their colleagues’ redundancies, they don’t embrace the AI tools enthusiastically. They resist them, quietly and persistently. A workforce convinced the technology exists to replace them will find a thousand ways to ensure it doesn’t work well enough to prove them right. The companies announcing AI-driven layoffs today are poisoning their own AI adoption tomorrow.

The bigger problem, though: it makes it harder to have an honest conversation about the transition that’s actually happening. AI is genuinely changing some kinds of work. Zencoder’s engineering team went from 36 people to 30 and nearly doubled their throughput — that’s a real structural change backed by real numbers, not a press release. And there are broader shifts building underneath: the graduation cliff (fewer graduate hires, no entry-level pipeline), fewer middle managers, training programmes that don’t work at scale. These deserve serious responses.

When every garden-variety restructuring gets labelled “AI transformation,” though, the signal gets lost in the noise. People stop believing any of it. The sceptics write off every warning as corporate theatre. The people who should be adapting assume it’s all hype. And the genuine cases, the companies that really are building differently, get lumped in with the performers.

The tell

There’s a simple test. When a company announces AI-driven layoffs, ask three questions. Can they name the specific AI tool or capability that changed the economics? Can they show measurable output or productivity data from before and after? Are they still hiring AI talent while cutting elsewhere?

Block could answer all three. Dorsey named Goose, their internal AI coding assistant. He cited a 40% increase in engineering output. And they were hiring senior AI engineers even as they cut 4,000 roles. You can disagree with the decision, but the story hangs together.

Most companies can answer none of them. The layoff announcement says “AI” the way a previous generation’s announcement said “digital transformation” — as a direction of travel, not a specific cause. It’s strategy language doing the work of operational evidence.

I’m not sure what percentage of AI-cited layoffs in 2026 are genuine and what percentage are theatre. Nobody is, including the CEOs making the announcements. Altman’s honest enough to admit he doesn’t know either, which is more than most people involved in this will concede.

What I do know is that the companies treating AI as a justification rather than a transformation are going to find themselves in an awkward position in about eighteen months. The ones that used AI as cover for cost-cutting will have smaller teams but the same processes, the same tools, the same capabilities — just fewer people doing the work. The ones that actually rebuilt around AI capabilities will look fundamentally different.

The market will notice the difference eventually. It always does. The question is how much damage gets done to how many people’s livelihoods in the meantime, while the theatre plays on.