Meta Layoffs Hit Safety Teams Ahead of a $62 Billion Trial

Meta Platforms reported $56.3 billion in first-quarter revenue on April 29, a 33 percent jump year over year with a 41 percent operating margin. Three weeks later, on May 20, layoff notices reached roughly 8,000 workers across the parent company of Facebook, Instagram, and WhatsApp, delivered in three waves starting at 4 a.m. local time in each region, beginning with Singapore and rolling through the United Kingdom and the Americas.

Which teams received those notices matters as much as the count. Meta’s integrity department, the group charged with removing hate speech and malicious content from platforms used by 3.56 billion people daily, was cut on day one, alongside cybersecurity staff and content designers. Two open legal proceedings against Meta, with a combined potential exposure running into the tens of billions of dollars, are directly tied to how well those departments function.

The Teams That Didn’t Make the Cut

The internal memo announcing the layoffs came from Janelle Gale, Meta’s chief people officer, who described the reductions as structural rather than performance-based. The company is reorganizing around AI-focused pods, small teams designed to “operate with a flatter structure” and “move faster,” and workers who do not fit the new model are out regardless of recent performance reviews. Reporting from Business Insider identified the first-wave cuts concentrated in three functions:

• Integrity workers covering platform abuse, civic and elections policy, and hate-speech enforcement
• Cybersecurity staff from within Meta’s broader infrastructure and security organization
• Content designers responsible for the interface language governing how users on Facebook, Instagram, and WhatsApp report content, manage accounts, and understand platform rules

Removing staff from those three areas carries specific second-order costs. The integrity team sets enforcement policies that govern what 3.56 billion daily active users can see and post, and its decisions appear directly in active legal and regulatory proceedings against the company. Cybersecurity staff protect infrastructure that governments, hospitals, and financial institutions across hundreds of countries depend on daily. Content designers write the reporting mechanisms and consent flows that regulators examine when assessing whether a platform has met its statutory obligations.

Meta had already moved away from third-party contractors handling content moderation in March, preceding the May internal cuts by roughly eight weeks. In 2023, when the company cut more than 21,000 roles under the “Year of Efficiency” banner, more than 100 positions in trust, integrity, and responsibility were eliminated based on Department of Labor filings reviewed at the time. The current round continues that same contraction at a company now carrying substantially larger open legal exposure than it did three years ago. Gale told staff that additional reductions in the second half of 2026 have not been ruled out.

Growing Legal Exposure Meets a Smaller Safety Function

Meta is defending two separate proceedings that are each directly tied to content moderation, and both carry firm deadlines in 2026.

New Mexico’s $62 Billion Case

The New Mexico Attorney General’s civil action against Meta, which has expanded to include content moderation claims, is set for trial on September 8, 2026. Meta’s Q1 2026 10-Q filed with the Securities and Exchange Commission discloses that the Attorney General has indicated an intent to seek up to $62.85 billion in penalties. The 10-Q does not assess the probability of that outcome, but the figure is more than ten times Meta’s first-quarter revenue.

Legal standards in platform-harm cases often turn on what precautions the defendant took. A documented sequence of reductions to safety staff, from the March contractor shift through the May internal integrity cuts, creates a record that plaintiffs can use to argue Meta knew about enforcement gaps and chose not to address them. A shareholder proxy proposal filed with the SEC in 2026 separately cited Meta’s content moderation practices as a material risk to shareholder value, noting the company had lost two landmark court cases tied to those practices in the preceding proxy period.

The EU’s Digital Services Act Front

The European Commission opened formal proceedings against Meta in April 2024 under the Digital Services Act (DSA, the European Union’s legal framework for platform content obligations). Preliminary findings issued in October 2025 reflected the Commission’s view that Meta had infringed DSA requirements on illegal-content reporting, appeals mechanisms, and researcher data access. Meta’s 10-Q confirms the company has an opportunity to respond before a final ruling, but a confirmed infringement finding could trigger fines of up to 6 percent of global annual revenue under DSA enforcement rules.

Meta’s full-year 2025 revenue crossed $200 billion for the first time. At 6 percent of that figure, the maximum DSA fine exposure on its own runs above $12 billion.

The $135 Billion Bet on Machines Over Headcount

The logic behind the layoffs sits in the capital expenditure line. Meta’s Q1 2026 earnings release filed with the SEC confirms the company raised its full-year capital expenditure guidance to $125 billion to $145 billion, up from the prior range of $115 billion to $135 billion, citing higher component prices and additional data center costs. First-quarter capital expenditures alone reached $19.84 billion. The spending trajectory across three years:

• $39.2 billion in capital expenditures in 2024
• $72.2 billion in 2025, up 85 percent year over year
• $125 to $145 billion guided for 2026, nearly double the prior year

Analysts at Evercore estimate the cuts will generate roughly $3 billion in annualized payroll savings. Against the $135 billion midpoint of the capex range, that is about 2 percent of planned infrastructure spending. The math suggests the cuts are less about directly funding the AI build and more about demonstrating to investors that the people-cost line is being held while the machine-spending line doubles again.

Susan Li, Meta’s chief financial officer, told investors on April 29 that she could not predict the company’s optimal long-term workforce size given the pace at which AI capabilities are evolving.

The way to think about the investment is that we’re making a bet on the individual things that people care about, and that people are going to be more important in the future.

Mark Zuckerberg, Meta’s chief executive, said this at the April 29 earnings call. The approximately 7,000 workers being redirected rather than cut will move into newly created AI-focused organizational pods, with team names including Applied AI Engineering and Agent Transformation Accelerator. Each pod is designed to own its workstream end-to-end with fewer management layers than a traditional department structure. For a breakdown of how Meta’s new AI pod structure was assembled, the May 20 report on Meta’s AI pod restructuring covers the mechanics.

Inside the Workforce Left Standing

The approximately 7,000 workers being redirected into AI pods are not watching this from a position of safety. The Wall Street Journal reported that more than 1,500 Meta employees signed a petition demanding the company end data collection for AI training, responding to the Model Capability Initiative, an internal program that records keystrokes, mouse movements, and screen activity to develop AI agents capable of replicating those tasks. One unnamed policy employee told Wired that morale was low in part because U.S. staff felt they were “being used to train the AI models that will replace them.”

That climate had been building since before May 20. Employees built at least three internal countdown websites tracking the days to the layoffs, one labeled “Big Beautiful Layoff.” Satisfaction ratings on Blind, an anonymous professional network widely used in the technology sector, fell 25 percent from a 2024 peak, with culture scores declining 39 percent over the same period. Separately, UK-based employees began organizing a unionization drive through United Tech and Allied Workers in the weeks preceding the cut.

At the same time, Meta Superintelligence Labs, the company’s dedicated AI research division, is reported to be offering compensation packages of up to $100 million for senior AI researchers. For workers let go in the May wave, the U.S. severance formula covers 16 weeks of base pay plus two additional weeks per year of service, along with 18 months of health coverage, triple what Meta offered in 2023 rounds. Immigration support and career-placement services are also included.

Meta’s Numbers Inside a Broader Tech Contraction

Meta’s May round is not isolated. Cisco announced roughly 4,000 cuts the same week. Oracle eliminated an estimated 20,000 to 30,000 roles in March. The pattern at every major company is identical: record revenue, elevated AI infrastructure spending, and workforce reductions framed as offsetting one with the other.

A Goldman Sachs survey found that AI-driven layoffs are running at more than 16,000 payroll reductions per month across the economy in 2026. Across the technology sector, more than 73,000 jobs were cut at 95 companies in the first four months of the year, with projections that the full-year count will exceed the 124,000 roles eliminated across all of 2025.

Since November 2022, Meta has eliminated more than 33,000 positions, cutting 11,000 in that first round, 10,000 in March 2023, and roughly 3,600 more in January 2025. The company ended March 2026 with 77,986 employees on the books, a figure that has been declining since.

In earlier rounds, the stated justification was correcting pandemic-era over-hiring. That framing is not available in 2026, when Meta is cutting against a backdrop of record quarterly profit and a near-doubling of its annual AI infrastructure budget.

The September 8 trial date in New Mexico arrives roughly 16 weeks from now. If Meta enters that courtroom with a demonstrably smaller integrity function than it had when the case was filed, and the state’s legal theory holds that inadequate moderation represents a documented pattern rather than isolated failure, the May restructuring memo will appear as an exhibit. If the case settles before trial, or the legal theory fails to survive judicial review, the cost-cutting thesis holds and the AI buildout continues on its current trajectory. Either way, the number that will define Meta’s autumn will not be found on the revenue line.