Meta will cut roughly 8,000 jobs starting May 20, 2026, and Microsoft is offering voluntary buyouts to about 7% of its U.S. workforce in the first program of its kind in company history. Both giants framed the cuts as funding AI investment. The math, the timing, and one OpenAI chief’s blunt phrase suggest something else is driving the knife.
The two announcements landed inside the same 48 hours, on April 23, 2026, alongside a wave of fresh tech reductions that pushed 2026 layoffs past 95,000 by the time CNBC published its tally a day later. Yet the deepest reading of the announcements is not that AI replaced these workers. It is that AI is being used as the lever, and the cover, for a much older corporate move.
What Meta and Microsoft Actually Announced
Meta’s chief people officer, Janelle Gale, told staff in a memo that the company would cut about 10% of its workforce, eliminate 6,000 unfilled roles, and start exits on May 20, 2026.
“We’re doing this as part of our continued effort to run the company more efficiently and to allow us to offset the other investments we’re making,” Gale wrote. The internal memo was first reported by Axios and later confirmed by Meta to multiple outlets.
Microsoft’s program is structured differently. Bloomberg’s reporting on the internal communications shows the buyout targets U.S. employees at the senior director level and below whose age plus tenure equals 70 or more. Eligible workers and managers receive package details on May 7, 2026, and have 30 days to accept. Microsoft’s stock fell nearly 4% on the day of the announcement.
The two firms join Atlassian (1,600 cuts), Block (4,000), WiseTech Global (2,000), and Oracle (more than 25,000 across late 2025 and early 2026), each invoking AI without quite blaming it.
The Productivity Math Doesn’t Quite Add Up
Google chief Sundar Pichai has put the AI productivity dividend in numbers. He said internal AI tools have lifted Google engineering capacity by about 10%, measured in hours saved per week. Pichai also said in April 2026 that 75% of new code at Google is now AI-generated and reviewed by human engineers.
That 10% figure tracks neatly with the 7% to 10% workforce reductions at Meta, Microsoft, and Atlassian. The arithmetic looks tidy. It is also misleading.
Engineering is the easiest function to automate inside a software company. Tasks are well-specified, outputs are testable, and code compiles or it does not. Most corporate roles, including legal, marketing, finance, recruiting, and product management, do not work that way. They involve ambiguous briefs, competing stakeholders, and outputs that can only be judged subjectively.
Applying a coding-derived productivity number to those functions assumes a clean transfer that nobody has demonstrated. The companies are betting on that transfer anyway.
| Company | Cuts Announced | Share of Workforce | 2026 AI Capex or Spend |
|---|---|---|---|
| Meta | ~8,000 (plus 6,000 roles unfilled) | ~10% | $115B to $135B |
| Microsoft | Voluntary buyouts, ~7% of U.S. staff | ~7% | $80B+ on AI infrastructure |
| Block | ~4,000 | ~40% | Not disclosed |
| Oracle | ~25,000 | Estimated 15%+ | $25B+ on cloud and AI |
| Atlassian | 1,600 | ~10% | Not disclosed |
Sam Altman Named the Tactic in February
OpenAI chief executive Sam Altman gave the practice a label two months before the Meta and Microsoft announcements landed. Speaking at the India AI Impact Summit in February 2026, he called it AI washing.
“There’s some AI washing where people are blaming AI for layoffs that they would otherwise do, and then there’s some real displacement by AI of different kinds of jobs,” Altman said, in comments published by Fortune.
Independent data backs the skepticism. An analysis of layoffs from 2020 through early 2026 found only 16% were directly attributable to AI. The bigger drivers, by volume, are still pandemic-era over-hiring, post-merger consolidation, and ordinary cost cutting. About 55,000 layoffs in 2025 were tagged AI-related, or less than 1% of total job losses that year.
Meta’s Reality Labs Sleight of Hand
Meta’s announcement is harder to read at face value because of what happened in March. The company said on March 18, 2026, that it would shut down Horizon Worlds in VR by June 15, removing the app from the Quest store at the end of March. Reality Labs, the division that builds Meta’s metaverse and AR hardware, employed about 15,000 people as of January 2026 and had recently shed more than 1,000 of them.
Forty-eight hours later, Meta reversed the shutdown. TechCrunch reported that the CTO posted on Instagram Stories that Horizon Worlds in VR would remain available “for the foreseeable future” after backlash from existing users.
The reversal did not stop the staffing pullback. Reality Labs has now lost more than $83.6 billion cumulatively since 2020. The April cuts may quietly absorb employees whose metaverse work was already being wound down. Calling it an AI move sounds like a strategy. Calling it the long tail of a failed bet sounds like an admission.
Why the Word “AI” Triggers a Stock Jump
Block’s February 2026 announcement is the cleanest case study in financial reward. Chief executive Jack Dorsey said the company would cut roughly 4,000 jobs, nearly half its workforce, and tied the move to AI-driven efficiency. Block’s gross profit had grown 24% in the previous quarter.
The market reaction was immediate. Block shares rose roughly 20% on the announcement, ending the after-hours session at about $67. Bloomberg’s headline a few days later read, “Jack Dorsey’s 4,000 Job Cuts at Block Arouse Suspicions of AI-Washing.”
Dorsey’s own language put the bet plainly. “I think most companies are late,” he wrote on X. “Within the next year, I believe the majority of companies will reach the same conclusion and make similar structural changes.” Investors heard a forward-looking pitch and rewarded it. That financial signal is unlikely to be lost on every other tech CEO running the numbers in 2026.
The Real Strategy Is Forcing Adoption From Above
The most useful framing is not that AI replaces workers, and not that AI is invented as cover. It is that AI is being used as a forcing function.
Tech leaders believe AI will reshape how software, marketing, and operations get built. They do not yet know in detail how. So they create pressure. Cut headcount, hold expectations constant, and let the remaining teams figure out which workflows AI can absorb.
The workers most at risk are not necessarily those whose tasks can be replicated by AI. They are those who wait for pressure to arrive from outside rather than getting ahead of it now.
This reading explains why the cuts cluster in the 7% to 10% range, where Pichai’s productivity math holds. It explains why companies hire aggressively in AI roles even as they shed others. Meta’s Superintelligence Labs unveiling of Muse Spark earlier this year sat alongside reports that the company is paying nine-figure packages to a small group of AI researchers. The same employer signing pink slips is signing record offers.
What the Capex Numbers Reveal
Meta told investors on its January 2026 earnings call that 2026 capital expenditure would land between $115 billion and $135 billion, nearly double the $72.2 billion it spent in 2025. The guidance, reported by Data Center Dynamics, drove a 10% surge in Meta stock the day of the disclosure.
Microsoft has separately committed more than $80 billion to AI infrastructure in fiscal 2026. The combined three-year capex commitment from the two firms now exceeds half a trillion dollars. The 8,000 Meta layoffs save, on rough estimate, $1.5 billion to $2 billion in annual payroll. Set against $135 billion in capex, the savings are not the point.
The cuts are signaling, not financing. They tell Wall Street and the rank and file the same thing at once: AI is the priority, and headcount is no longer a measure of corporate health.
Watch the Hiring Lines, Not the Cuts
The honest test of whether 2026’s layoff wave is AI replacement, AI cover, or AI pressure will play out in 2027. If Meta, Microsoft, and their peers rehire at the same scale with different skill profiles, redesign workflows, and emerge measurably more productive, the case for AI as a real driver gets stronger. If they pocket the payroll savings without restructuring how work happens, the AI-washing critics will look right.
For now, the loudest signal is who these companies are paying, not who they are letting go. The same week Meta confirmed 8,000 cuts, recruiters were still floating eight- and nine-figure offers to senior AI researchers from rival labs. If you want to know where the giants are going, ignore the layoff list and study the offer sheet.
Frequently Asked Questions
How Many Tech Workers Have Been Laid Off in 2026?
More than 95,000 tech workers had been cut globally as of late April 2026, according to layoff trackers including Layoffs.fyi and TrueUp. The cumulative figure since 2020 is approaching 900,000. Oracle, Meta, and Microsoft account for the largest single-company shares of the 2026 total.
Why Are Meta and Microsoft Cutting Jobs While Spending More on AI?
Both companies say the cuts free up capital for AI investment. The financial gap between savings (low single-digit billions per year) and AI capex (more than $200 billion combined) is so wide that the layoffs are best read as a strategic signal rather than a funding mechanism. The cuts force remaining teams to adopt AI tools faster.
What Does “AI Washing” Mean in Layoffs?
AI washing describes the practice of attributing layoffs to artificial intelligence when the underlying reasons are post-pandemic over-hiring, weaker margins, or routine cost cutting. OpenAI chief executive Sam Altman used the term at the February 2026 India AI Impact Summit. Independent layoff databases find only about 16% of cuts since 2020 are directly tied to AI.
When Will Meta’s Layoffs Take Effect?
Meta’s exits begin on May 20, 2026, according to the internal memo from chief people officer Janelle Gale. The company is also dropping plans to fill 6,000 currently open roles. Affected employees in the United States receive severance and equity vesting per Meta’s standard package.
Is AI Actually Replacing White-Collar Jobs Yet?
In narrow domains, yes. Software engineering, customer support triage, and contract review show measurable AI displacement as of April 2026. Most professional roles involve ambiguous outputs that AI cannot yet verify autonomously. The broader “intelligence explosion” framing pushed by figures including AI entrepreneur Matt Shumer remains contested by economists and labor researchers.
Leave a Comment