Meta layoffs starting this week stress harsh AI reality inside Zuckerberg’s company

When Meta CEO Mark Zuckerberg told employees about his plan in late 2022 to lay off 11,000 employees, in cuts that would later expand to 21,000, he was contrite in admitting that he overhired during the Covid pandemic.

"I got this wrong, and I take responsibility for that," Zuckerberg said in a message to staffers in November of that year as the company's stock was in freefall. In early 2023, Zuckerberg said the cuts were necessary as part of Meta's "year of efficiency."

More than three years later, with the latest round of mass layoffs set to begin this week, the tone at the top has changed dramatically. Starting Wednesday, Meta is reducing its workforce by about 10%, or about 8,000 jobs. The company also scrapped plans to fill 6,000 open roles, according to a memo about the layoffs in April.

The current downsizing follows cuts of about 1,000 staffers in January in the company's Reality Labs unit, and reductions in March impacting hundreds more workers, along with the decision to shift away from third-party vendors and contractors tasked with content-moderation tasks.

Meanwhile, Meta is ramping up its investments in artificial intelligence, lifting its 2026 guidance for capital expenditures last month by as much as $10 billion, reaching as high as $145 billion.

In announcing the coming job cuts, a week before disclosing the capex increase, Meta told employees that the reductions are "all part of our continued effort to run the company more efficiently and to allow us to offset the other investments we're making."

There was no apology from Zuckerberg. Meta declined to comment for this story.

Internally, there's an emerging sense of dread across wide swaths of the company, according to current and former Meta employees who asked not to be named in order to speak freely on the matter. That's in part because more cuts are expected this year, including a potential round of layoffs in August, followed by another round later in the year, some of the sources said.

Finance chief Susan Li said during the first-quarter earnings call that executives "don't really know what the optimal size of the company will be in the future." Regarding AI investments, Li said, "our experience so far has been that we have continued to underestimate our compute needs even as we have been ramping capacity significantly as the advances in AI have continued and our teams continue to identify compelling new projects and initiatives."

Across the tech industry, workers are watching as stock prices balloon and AI startups soar to monster valuations while employers are simultaneously cutting headcount due to the rapidly emerging power of AI. So far in 2026, there have been almost 110,000 layoffs at 137 tech companies, according to Layoffs.fyi, after roughly 125,000 cuts all last year.

At the current pace, cuts could approach the peak in 2023, when there were over 260,000 layoffs, as many software and digital media companies rightsized following the Covid hiring boom.

'Replaced by machines'

Umesh Ramakrishnan, chief strategy officer at executive search firm Kingsley Gate, said the current trend of AI taking jobs is hard for workers, but welcomed by investors.

"It's easy to tell somebody, 'Hey, listen, I made a mistake by hiring more people than I should have,'" Ramakrishnan said. "Now the world understands that jobs are being replaced by machines, and if you're not doing that, shareholders are getting upset."

Cisco is the latest tech giant to make such an announcement, telling investors alongside quarterly earnings last week that it was eliminating fewer than 4,000 jobs.

"The companies that will win in the AI era will be those with focus, urgency, and the discipline to continuously shift investment toward the areas where demand and long-term value creation are strongest," Cisco CEO Chuck Robbins wrote in a blog post, titled "Our path forward."

Cisco shares popped more than 13% on Thursday, their best day since 2011, after the company reported better-than-expected results and lifted its AI infrastructure guidance.

Wall Street still isn't sold on Meta's story, but that's largely because the company's AI strategy has been scattered and remains largely in flux. The stock is down about 7% so far this year and almost 5% over the past 12 months, underperforming all of its megacap peers other than Microsoft.

Whatever anxiety investors are experiencing, the feelings inside the company are more intense, with some long-time staffers questioning Meta's AI pursuits under AI chief Alexandr Wang, while also weighing if now is the time to leave for opportunities at other companies in the AI race, according to current and former employees.

Data aggregated by Blind, an anonymous professional network that requires users to verify their employment with a work email address, reveals some of the internal malaise.

Meta's overall rating by employees on Blind has declined 25% from a peak in the second quarter of 2024 to the current period, with a 39% drop in its culture rating. In every category other than compensation. Meta has seen a ratings decline and dramatically underperforms rivals Amazon, Google and Neflix, the Blind data reveals.

The company's full-court press with AI included the recent debut of an employee tracking tool intended to collect data from staffers' actions, such as mouse movements and keystrokes on their work computers. The Model Capability Initiative (MCI), as it's called, is part of Meta's efforts to train AI models to power digital agents that can perform various coding and white-collar tasks.

Employees have characterized the data tracking tool as "dystopian," according to messages viewed by CNBC, with some workers expressing fear that personal information could be leaked. Some Meta workers have noted that their workplace computers appear slower since the company initiated the project, adding to their frustration, sources said.

Meta workers responded by creating an online petition that urges Zuckerberg and leadership to shutter the project.

"Collecting and repurposing this kind of data raises serious concerns around privacy, consent, and trust in the workplace," the petition says. "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."

Leo Boussioux, an assistant professor of information systems at the University of Washington's Foster School of Business, described Meta as one of many companies currently overhauling its workforce and operations to accommodate "the fact that AI is changing the way we work."

Boussioux said one goal could be to increase fear or pressure, using AI-related threats and layoffs as a "form of weapon to enable a culture change." But, he said, it could also reflect "poor management that does not know how to enable this in a more comfortable way for the employees."

—CNBC's Stephen Desaulniers and Lora Kolodny contributed to this report.

WATCH: Meta's overall numbers were impressive, says Jim Cramer.