Mark Zuckerberg once told his staff that 2026 would mark the year artificial intelligence began to dramatically reshape work at Meta. That prediction lands with force this week. On May 20 the company starts laying off roughly 8,000 people, about 10 percent of its workforce. It will also leave 6,000 open positions unfilled. The moves come as Meta redirects thousands more employees toward new AI projects and flattens management layers to speed decisions.
The numbers tell a stark story. Headcount stood at 77,986 at the end of March. Combine the cuts, the unfilled roles, and the reassignment of 7,000 workers, and roughly 20 percent of the company faces immediate disruption. Janelle Gale, Meta’s chief people officer, laid out the plan in an internal memo Monday. She described the shift as one that “will make us more productive and make the work more rewarding.” Some employees see it differently.
Inside the offices a sense of dread has taken hold. Longtime staffers question the direction of AI efforts led by new talent such as Alexandr Wang. They wonder whether their roles will survive the next round of reductions expected later this year. Employee ratings on the anonymous network Blind have fallen sharply. The overall score dropped 25 percent from its peak in 2024. Culture ratings plunged 39 percent. In most categories Meta now trails Amazon, Google and Netflix.
The pressure traces directly to spending. Meta expects to burn through $115 billion to $135 billion this year on infrastructure, much of it compute capacity for training ever-larger models. Zuckerberg himself linked the layoffs to those costs during an internal town hall. “That means that we do need to take down the size of the company somewhat,” he said, according to Reuters. He stopped short of claiming AI tools had replaced the departing workers. The math was simpler. Heavy capital outlays required lighter payrolls.
Yet the company keeps moving people toward AI. Gale’s memo announced transfers into four new organizations dedicated to building AI tools and applications. These groups adopt what the company calls “AI native design structures.” They operate with fewer managers per employee. Smaller pods and cohorts gain more ownership and move faster, Gale wrote. Two teams under CTO Andrew Bosworth focus on “AI for Work,” developing agents that can handle tasks once performed by humans. Another unit, Central Analytics, measures productivity gains from those agents. A fourth, Enterprise Solutions, will receive details soon.
The reassignments and cuts arrive together. Employees in the United States must work remotely on May 20. Layoff notices will land at 4 a.m. local time. Severance offers 16 weeks of pay plus two extra weeks for each year served. Gale acknowledged the difficulty. “We know days like this are extremely hard, and we appreciate you showing up for each other.” The words landed flat for many.
Frustration boiled over earlier this month. Meta rolled out the Model Capability Initiative, or MCI. The program records keystrokes, mouse movements and screenshots across hundreds of applications. The goal is to train AI agents that mimic human computer use. Some workers called the effort dystopian. Computers slowed under the extra load. Fears of data leaks spread.
More than 1,000 employees signed a petition. “Collecting and repurposing this kind of data raises serious concerns around privacy, consent, and trust in the workplace,” it reads. “It should not be the norm that companies of any size are permitted to exploit their employees by nonconsensually extracting their data for the purposes of AI training.” Protesters posted flyers in offices and flooded Meta’s internal Workplace platform with pictures of elephants, a pointed reminder of the layoff “elephant in the room.” Executives stayed largely silent for weeks after initial reports surfaced.
This tension reflects a broader shift. Meta has dialed back its once-expensive bet on the metaverse. It cut hundreds from Reality Labs earlier in the year. Now the focus sits squarely on competing with Google, OpenAI and Anthropic in foundational models and practical agents. Zuckerberg has pushed every employee to adopt AI coding assistants and to include AI usage in performance reviews. The message is clear. Adapt or risk obsolescence.
But adaptation carries costs that extend beyond dollars. Morale has suffered. Some veterans eye exits to pure-play AI startups where the mission feels less conflicted. Others remain, tasked with training the very systems that could shrink their teams. The company’s push to capture employee behavior data only heightens the unease. And the layoffs keep coming. Sources told Reuters to expect further reductions in the second half of 2026, possibly including a wave in August.
Industry watchers see Meta as an early test case. Similar moves have hit Cisco, Microsoft, Block and Coinbase. Challenger, Gray & Christmas data from March showed AI cited in thousands of job cuts across tech. The pattern suggests companies are trading headcount today for infrastructure tomorrow. Returns on those massive AI bets remain unproven. Productivity gains from agents exist mostly in pilots and internal demos.
Zuckerberg sounded pragmatic in his town hall. He separated the efficiency drive from any claim that AI had directly displaced workers. The infrastructure bill simply demanded offsets. That honesty, however blunt, failed to ease anxiety. One former employee told CNBC that an “emerging sense of dread” now stretches across departments. The Slashdot summary of that reporting captured the mood in stark terms.
Meta’s stock has held up through the announcements. Investors appear to accept the trade-off: higher capital expenditure today for leadership in AI tomorrow. Employees shoulder the human side of that equation. Some will land in the new AI pods with flatter reporting lines and bigger scope. Others will receive severance and begin again elsewhere. A few may stay and fight the surveillance tools they see as overreach.
The weeks ahead will reveal how cleanly the transition lands. New role details reach transferred employees Wednesday. Layoff emails arrive the same week. Managers lose layers. Teams reorganize around AI workflows. The company insists the result will prove faster and more rewarding. For now the dominant emotion is uncertainty. Dread. And a quiet calculation among thousands about what comes next when the machines start doing more of the work.
Recent coverage adds texture to the picture. The New York Times reported the 7,000 reassignments and the memo’s emphasis on flatter structures just hours before the layoffs begin. Reuters obtained the full internal document and detailed the employee revolt over tracking software. Both accounts confirm the scale and the friction. They show a company moving at speed to remake itself while its people absorb the shock.
Meta has navigated big shifts before. It survived the pivot from desktop to mobile. It absorbed criticism over content moderation and privacy scandals. The AI transition feels different. This time the technology threatens to reorganize the work of the knowledge workers who built the company. The coming months will test whether the bet pays off before the talent and trust erode further. Zuckerberg placed his wager years ago. This week the bill comes due.
WebProNews is an iEntry Publication