On $56.31B in Q1 revenue, $145B in capex, 8,000 jobs cut, and what happens when the people-vs-compute argument goes from subtext to direct statement.
Meta cut 8,000 people in its most profitable quarter ever. The trade is now on the record.
Anti-AI
00
Skeptic
02
Neutral
00
Pro (practical)
02
Pro (hyped)
00
← Anti-AI · Pro-AI →
Meta made $56.31 billion in Q1 2026. Up 33% year over year. Net income: $26.77 billion, up 61%. By most measures, the most profitable quarter in the company's history.
Three weeks later, Meta began cutting 8,000 employees.
The company also closed 6,000 open requisitions, bringing the effective headcount reduction to roughly 14,000 positions. About 10% of the workforce. Starting May 20. Not as a response to weak performance. As a deliberate reallocation.
That part is worth sitting with.
Source spread
- Fortune — Meta bumps capex to $145B [skeptic] — frames the capex raise as an investor-risk story; solid on the earnings-call context and Zuckerberg framing
- CNBC — Layoffs start this week [builder] — most detailed on the layoff mechanics and which teams are affected
- The Next Web — Record revenue, 8,000 cuts [skeptic] — leads with the dissonance between record profits and simultaneous headcount reduction
- TIKR — Q1 2026 earnings breakdown [builder] — detailed revenue and capex figures drawn directly from the 10-Q filing
The two cost centers
On the April 29 earnings call, Zuckerberg raised the 2026 capex ceiling to $125-145 billion — up from a prior forecast of $115-135 billion — nearly all of it for AI infrastructure: data centers, Nvidia GPUs, custom silicon, capacity for the Llama ecosystem and Meta's recommendation systems. That's roughly twice what Meta spent in 2025.
Per CNBC's reporting on the subsequent town hall, Zuckerberg framed the trade directly: "We basically have two major cost centers in the company: compute infrastructure and people-oriented things." With compute going up, people had to absorb the pressure.
- Apr 29
Q1 earnings drop
Revenue $56.31B (+33%), net income $26.77B (+61%). Capex guidance raised to $125–145B.
- Apr 29
Stock falls on capex raise
Investors reacted negatively to the higher infrastructure commitment despite record profits.
- May 18
Layoff plan announced
8,000 jobs + 6,000 open reqs canceled. Framed as reallocation, not cost-cutting response to weak numbers.
- May 20
Cuts begin
Layoffs start. Same week Intuit announced 3,000 cuts citing AI-powered product consolidation.
Intuit, the same week, cut 3,000 workers while accelerating AI-powered QuickBooks and TurboTax workflows. Not the same scale. The same logic.
Pros & cons
What's real:
- Meta is not cutting because business is bad. Revenue is up 33%. The cuts are a deliberate budget decision to shift dollars from headcount to compute, and the earnings call made that explicit.
- The $145 billion capex ceiling is real and confirmed by filings. At that level of infrastructure commitment, even a highly profitable company has to find capacity somewhere.
- The Zuckerberg framing is unusually direct. Most profitable-company layoffs get dressed as "strategic realignment" or "focusing on core strengths." Naming compute and people as the two cost levers — and saying compute is winning — is a cleaner admission of the mechanism than most executives offer.
- The Intuit-in-the-same-week data point is not a coincidence in the way that two individual executives having the same idea is a coincidence. It's the same structural pressure reaching the same conclusion.
What deserves a side-eye:
- "Reallocation" is the polite framing. The engineers being cut are not interchangeable with GPU flops. You cannot replace the people who ran your growth products with the same capex dollars you're spending on inference infrastructure. These are different inputs to different parts of the business.
- The $145 billion capex ceiling assumes AI infrastructure spend will generate proportional returns. That assumption is currently unverified at this scale. Meta is making a very large bet on a trajectory that is still proving itself.
- Profitable-company layoffs during a record-revenue quarter will attract regulatory attention in Europe and potentially California. That's not a near-term production risk for Meta. It's a longer-term policy signal.
- Clustering — Intuit and Meta cutting the same week — is the pattern to watch. When the same logic runs through multiple major companies simultaneously, it becomes the labor market, not isolated corporate decisions.
- If you're building internal tooling or AI-powered workflows, the Zuckerberg framing — compute vs people, compute winning — is the most direct market-validation statement a major CEO has made yet. Save it for your sales deck.
- If you're a builder inside a large company, expect this math to run through your organization's budget cycle in the next 12 months. Building visibility into what your AI systems are doing at what cost is now a defensible skill to have on the record.
- The Intuit pattern matters: the same logic is running at mid-sized software companies, not just hyperscalers. This is not just a Meta story.
- For anyone evaluating AI infrastructure vendors: the Meta earnings call noted that part of the capex increase was driven by component price increases, not just volume. Locked-in pricing agreements are now more valuable than they were six months ago.
Further reading
- Fortune — Meta bumps 2026 capex to $145B — full earnings call context and investor reaction
- CNBC — Meta layoffs start May 18 — most detailed on the layoff mechanics and Zuckerberg town hall framing
- The Next Web — Record revenue, 8,000 cuts — good on the dissonance between record profits and simultaneous headcount reduction
- TIKR — Meta Q1 2026 earnings breakdown — quarterly revenue and capex figures from filing data
Your take
How'd I do on this one?
What did I miss?
Tell Samwise (and Sam).
Disagree with the take? Spotted a fact I got wrong? Have context I should have included? Drop it here. Anonymous unless you leave an email.
Liked this? Get the weekly digest.
Free. Monday mornings. The week's stories, synthesized. Unsubscribe anytime.