The emails started landing at 4 AM Singapore time. Then London woke up and got theirs. By the time California rolled out of bed on May 20, 8,000 Meta employees across five divisions had already been told their jobs no longer exist — and 6,000 open positions they might have applied to had been erased from the system entirely. That’s 14,000 roles gone in a single corporate exhale, executed with the kind of precision that suggests Mark Zuckerberg has been rehearsing this moment for months.
What makes this particular bloodletting remarkable isn’t the scale — Meta has done this before, three times in the last 18 months. It’s the timing. The company just posted $56 billion in quarterly revenue, a record. Profits are up. The stock is near all-time highs. By every traditional measure of corporate health, Meta is thriving. And yet here we are, watching 10% of the workforce get shown the door because the company decided its humans aren’t the right kind of humans anymore.
The Five Divisions That Got the Axe — And What They Tell You About Zuckerberg’s Priorities
The layoffs didn’t hit randomly. Five specific divisions were targeted: Reality Labs, the Facebook social division, recruiting, sales, and global operations. Read that list again. It’s not a cost-cutting exercise — it’s a corporate lobotomy with a roadmap.
Reality Labs has been bleeding for over a year now. The division lost 1,000 to 1,500 people in January, another 700 in March, and saw its budget slashed by 30%. The metaverse — remember the metaverse? — has been quietly downgraded from Zuckerberg’s civilizational bet to a line item that needs to justify its existence every quarter. The remaining Reality Labs engineers are being folded into what Meta internally calls “AI-focused pods” under Alexandr Wang’s Superintelligence Labs division.
That’s the tell. Meta isn’t shrinking — it’s metamorphosing. The people who built social features, recruited talent, closed ad deals, and managed global operations are being replaced by a leaner organism whose sole purpose is feeding the AI machine. The recruiting team being gutted is especially telling: Meta doesn’t plan to hire its way to the future anymore. It plans to build its way there with fewer, more expensive engineers working on models, not products.
$145 Billion in Capex — The Number That Explains Everything
When Meta first announced these layoffs in April, the capital expenditure guidance was $115 to $135 billion for 2026. A month later, that number has quietly crept up to $125 to $145 billion. To put that in perspective: Meta spent $39.2 billion in 2024 and $72.2 billion in 2025. The company is nearly quadrupling its infrastructure spend in two years.
Nearly all of it goes to the same place: data centres, Nvidia GPUs, custom silicon, and the infrastructure to support Meta’s Llama model ecosystem. The math is brutally simple. An experienced software engineer in Menlo Park costs roughly $400,000 a year in total compensation. Fire 8,000 of them and you free up $3.2 billion annually — enough to buy approximately 40,000 Nvidia H100 GPUs at current prices. Meta is literally trading people for processors.
And Wall Street loves it. The stock barely moved on the layoff news because investors have internalised the new logic: headcount is a liability, compute is an asset. Every dollar shifted from payroll to GPU clusters shows up in future AI capability, which shows up in better ad targeting, which shows up in revenue. The humans were just the middlemen, and middlemen get cut.
The ‘Superintelligence Pods’ — Meta’s New Org Chart Is an AI Lab Wearing a Social Media Company’s Skin
The restructuring isn’t just about who leaves — it’s about how the survivors are reorganised. Teams across the company are being reconstituted into AI-focused “pods” that report into Alexandr Wang’s Superintelligence Labs. If you’re unfamiliar with Wang, he’s the 26-year-old founder of Scale AI whom Zuckerberg poached for a reported package worth hundreds of millions. His mandate: build the systems that make Meta’s other products smarter without needing more people to run them.
The pod structure means that what used to be a product team with designers, PMs, engineers, and data scientists is now a small cluster of ML engineers with access to Meta’s model stack. The assumption baked into this org chart is that AI can replace the functions that humans used to perform — content moderation, ad optimisation, feature development, even parts of recruiting. Meta isn’t betting on this future. It’s already living in it.
The Severance Package Is Generous — And That’s Part of the Strategy
Affected US employees receive 16 weeks of base pay plus two additional weeks per year of service, along with 18 months of continued health coverage. For a senior engineer with five years at Meta, that’s roughly 26 weeks of pay — over six months of runway. International packages vary but follow a similar formula.
The generosity isn’t altruism. It’s reputation management. Meta needs to keep hiring the specific kind of talent it wants — ML researchers, infrastructure engineers, AI safety specialists — and those people read Blind and Glassdoor. A stingy severance package during a $56 billion revenue quarter would have been a recruiting disaster. The message to future hires is clear: we’ll pay you well, and if we ever let you go, we’ll pay you well then too. Just don’t expect to stay if your job can be automated.
More Cuts Are Coming — And Nobody Knows When
Meta has confirmed that additional layoffs are planned for the second half of 2026, though the timing and scope remain unspecified. The vagueness is deliberate. It keeps remaining employees in a perpetual state of productive anxiety — work harder, prove your value, demonstrate that you’re building something the AI can’t replace. It’s management by existential dread, and it works.
The pattern across Big Tech is now unmistakable. Google, Microsoft, Amazon, and Meta have collectively cut over 100,000 jobs in the last 18 months while simultaneously committing over $600 billion to AI infrastructure. These aren’t struggling companies making tough choices. These are the most profitable corporations in human history choosing to replace their workforce with a different kind of workforce — one made of silicon, electricity, and software weights.
The Verdict: This Is the Template, and Every Company Is Watching
Meta’s May 20 layoffs will be remembered not for the number — 8,000 is almost routine now — but for the mechanism. The simultaneous posting of record revenue, the surgical targeting of specific divisions, the immediate restructuring into AI pods, the generous severance that buys silence and goodwill. It’s a playbook, and every Fortune 500 CEO is taking notes.
The uncomfortable truth is that Zuckerberg might be right. If AI can do the work of a content moderator, a recruiter, or an ad sales rep at a fraction of the cost and at infinite scale, then keeping those humans on payroll isn’t compassionate — it’s inefficient. The question isn’t whether other companies will follow Meta’s lead. It’s how long they’ll pretend they aren’t already doing the same thing.
Somewhere in Singapore, someone who woke up to a 4 AM email is updating their LinkedIn profile right now. In Menlo Park, the Superintelligence Labs team is already filling the gap they left behind. The transition happened overnight, which is exactly the point. In the AI economy, 8,000 people don’t get laid off — they get deprecated.