Skip to content

Meta Will Fire 8,000 People on May 20. The Same Year, It Will Spend $135 Billion on AI.

DS
LDS Team
Let's Data Science
12 min
The internal memo went out on April 19. Eight thousand roles are being cut, 10 percent of Meta's workforce, with more to come in the second half of 2026. Teams are being reorganized into AI pods under 28-year-old Chief AI Officer Alexandr Wang.

On Sunday, April 19, Meta employees received an internal memo that used phrases most of them had not seen before.

The memo promised a "step change in engineering productivity and product quality." It said the company was "fundamentally rewiring how we operate." It did not mention layoffs by name. By Monday morning, Reuters had the number: 8,000 people, or 10 percent of Meta's 78,865-person workforce. The cuts would take effect on May 20.

They are not the last round. A second wave is planned for the second half of 2026. Mark Zuckerberg has not said how large it will be.

The Trade Is Not Subtle

Meta's 2025 financials were the strongest in the company's history. Revenue hit $201 billion, up 22 percent year over year. Fourth-quarter net income and annual free cash flow both set records.

The company is cutting 8,000 jobs anyway.

The reason is on a different line of the balance sheet. Meta's 2026 capital expenditure guidance tops out at $135 billion, with a floor of 115 billion, nearly double the 72 billion dollars it spent in 2025.

Almost all of that new capex is going to AI infrastructure: data centers, compute capacity, and custom silicon. Meta has committed to a $27 billion joint venture with Blue Owl Capital to fund the Hyperion data center campus in Louisiana, and a separate $27 billion AI cloud deal with Nebius for dedicated compute capacity powered by Nvidia's Vera Rubin platform.

The math is a reallocation. Meta is shifting labor into compute. A senior engineer costs roughly $500,000 per year fully loaded. Eight thousand such cuts free up about four billion dollars annually, a fraction of one quarter of Meta's planned AI spend. The layoffs are not saving the company. They are making room.

Meta 2026 by the numbersValue
2025 revenue$201 billion
2025 revenue growth22% year over year
Q4 2025 net income$22.8 billion
2025 free cash flow$43.6 billion
2026 AI capex guidance$115 to $135 billion
Louisiana JV with Blue Owl Capital (Hyperion)$27 billion
Nebius AI cloud compute deal$27 billion
Layoffs announced April 198,000 employees
Percentage of workforce cut~10%
Cumulative cuts since 2022~25,000

The Divisions Getting Hit

The memo did not name specific teams. California WARN Act filings did.

On May 22, two days after the main layoff date, 124 positions will be eliminated at Meta's Burlingame office. On May 29, another 74 go at Sunnyvale. These filings are required by state law and name the earliest locations to feel the cut. The full list of affected divisions, confirmed by Reuters and The Information, reads like a map of the pre-AI Meta:

  • Reality Labs, the VR and AR division already cut in January (1,000 to 1,500 roles) and March (700 roles) of 2026
  • Facebook, the social network that made Meta and still accounts for the majority of its revenue
  • Recruiting, the team that spent three years staffing up for a headcount peak
  • Sales, particularly in the ad business
  • Global operations, the support function that runs policy, trust, and safety

Absent from the cut list: the teams being rebranded as AI. About 1,000 employees have already been moved into new roles with new titles, including "AI builder," "AI pod lead," and "AI org lead." The reorganization preceded the layoffs by several weeks. The pods are the product. The cuts pay for them.

The Man Running the Pods

Alexandr Wang is 28 years old. He is the youngest Chief AI Officer in the Fortune 50.

In June 2025, Zuckerberg paid $14.3 billion for a 49 percent stake in Scale AI, the data-labeling company Wang had built after dropping out of MIT. The price bought Meta half of Scale; it also bought Wang. He joined Meta as Chief AI Officer and immediately took over Meta Superintelligence Labs, a new division that absorbed Meta's AGI Foundations group and was given direct authority over the company's frontier-model work.

The first output of that division arrived on April 8, 2026: Muse Spark, Meta's first new foundation model since it stopped developing open-source Llama. LDS covered the shift then in Meta Paid $14.3 Billion for Alexandr Wang. Nine Months Later, He Killed Llama.

Wang's promotion was not uncontested. In March 2026, Andrew Bosworth, Meta's CTO, created a parallel Applied AI Engineering organization under VP Maher Saba, a Reality Labs veteran. The new unit reports to Bosworth, not Wang. Engineers across Meta are being reshuffled into Saba's org rather than Superintelligence Labs, a split that insiders describe as deliberate.

Wang is also responsible for the most high-profile departure from the old regime. Yann LeCun, the Turing Award winner who led Meta FAIR for a decade, left in late 2025 after publicly criticizing Wang as "young and inexperienced." Five months later, LeCun raised $1.03 billion for his own startup AMI, which LDS covered in Yann LeCun's AMI Labs Closed a $1 Billion Seed Round to Build World Models.

How We Got Here

JUN 2025
Meta buys 49 percent of Scale AI for $14.3 billion
Alexandr Wang joins as Chief AI Officer. Meta Superintelligence Labs is created.
LATE 2025
Yann LeCun departs
Meta's chief AI scientist leaves after calling Wang "young and inexperienced."
JAN 2026
Reality Labs cuts 1,000 to 1,500 roles
The first layoff round of the year hits the VR/AR division.
MAR 2026
Bosworth creates rival AI unit under Maher Saba
Applied AI Engineering splits Wang's authority. 700 more employees are cut.
APR 8, 2026
Muse Spark launches
The first model from Meta Superintelligence Labs. Llama is declared closed.
APR 19, 2026
Internal memo announces 8,000 layoffs for May 20
Second-half 2026 cuts also announced but not sized. Reality Labs, Facebook, recruiting, sales, and global operations named.
MAY 20, 2026
Layoff execution day
Burlingame (124) and Sunnyvale (74) WARN filings kick in on May 22 and May 29.
H2 2026
Second wave of cuts planned
Size not disclosed. Industry analysts expect up to another 8,000.

What Practitioners Should Actually Notice

For engineers at Meta, the detail that matters is not the layoff number. It is the pod structure.

Traditional product-engineering hierarchies at large tech companies have been organized around surfaces: Facebook News Feed, Instagram Reels, WhatsApp Business, Reality Labs VR. The pod model flips this. An AI pod is defined by a capability, not a surface. One pod might own agentic reasoning across every Meta surface. Another might own personalization. A third might own the evaluation harness. Engineers are expected to ship into whichever surface needs the capability that week.

This is the structure that Anthropic and OpenAI have run since founding. Both are far smaller than Meta. The question Meta is betting on is whether a 70,000-person company can run like a 1,000-person lab without losing the coordination benefits of its scale. The companies that have tried this before, including Google's multiple Bard reorgs in 2023 and 2024, did not succeed cleanly.

For engineers at other companies, the layoffs are a market signal. Meta and Oracle, both profitable, both cut the same way in the same quarter. LDS covered the Oracle version of this story in Oracle Cut 30,000 Jobs With a 6 AM Email. It Posted $6 Billion in Profit Last Quarter. The thesis is consistent: the 2026 AI capex wave is not being funded by new revenue. It is being funded by headcount.

The Other Side

Meta and its defenders argue this is not a crisis. It is a reallocation inside a profitable company.

The company points to a non-layoff fact: it ended 2025 with $43.6 billion in annual free cash flow. It can fund its AI spend without firing anyone. It is choosing to fire anyway because the roles being cut, in its view, are not the roles that will build what comes next. Alexandr Wang, in a public memo quoted by Business Insider in March, wrote that Meta's AI future requires "a much smaller and more talent-dense team" than the company currently has.

Critics see something less strategic. Sam Lessin, a former Facebook executive and partner at Slow Ventures, posted on X that "Meta is firing people because OpenAI has 3,000 employees and a $500 billion valuation. The pod thing is cover for headcount panic." Employees on Team Blind, the anonymous tech forum, described the reorganization in terms less flattering than "rewiring."

One data point to watch is the second wave. If the H2 2026 cuts come at a similar 8,000 level, Meta will have eliminated close to 20 percent of its workforce in a single year. At that point, even Wang's talent-density argument runs into the coordination problem.

The Bottom Line

Meta's 2025 was its best year ever on every metric that matters to Wall Street. Revenue was up 22 percent. Net income was a record. Free cash flow was a record. The stock closed March 2026 at an all-time high.

And 8,000 people will lose their jobs on May 20 because none of that was about them.

The Signal for Practitioners

The AI capex build-out is now funded by workforce cuts at profitable companies. Oracle, Meta, and Microsoft have all announced reductions while simultaneously posting record earnings and committing hundreds of billions to AI infrastructure. Data scientists and ML engineers who believe their job is safe because their company is profitable should re-read the line. The layoffs are not about profit. They are about which kinds of engineers the next decade needs.

Mark Zuckerberg has not spoken publicly about the cuts. Wang's internal memo ended with a line that reads better aloud than on paper: "Superintelligence is coming. We will build it or we will not."

Sources

Practice with real Ad Tech data

90 SQL & Python problems · 15 industry datasets

250 free problems · No credit card

See all Ad Tech problems
Free Career Roadmaps8 PATHS

Step-by-step roadmaps from zero to job-ready — curated courses, salary data, and the exact learning order that gets you hired.

Explore all career paths