The Numbers Behind Meta’s Layoffs Nobody’s Talking About
Everyone is reading Meta’s layoffs as AI replacing people. As a hiring manager who has been through layoff rounds myself, I think the real story is being missed. If you work in tech and want to understand what Meta’s layoffs actually signal, this is for you. The headline number is real, but the reason behind it is not simply that AI does the jobs now. It is that AI infrastructure is staggeringly expensive, and someone has to pay for it. The video above walks through the article, and below is the written version of the numbers behind Meta’s layoffs and the cost story nobody is talking about.
The real numbers
Meta cut roughly 8,000 roles, reassigned another 7,000 to AI, and canceled 6,000 open positions, for a total impact near 14,000.
The announcement was about 8,000 employees, roughly 10 percent of the workforce, reduced to shift resources toward AI. Another 7,000 were reassigned to AI-focused teams, and 6,000 previously open roles were canceled. Add it up and the total impact lands near 14,000, though about 6,000 of those were unfilled positions. If you work at a company like this, you already know how fast those numbers turn from a headline into your own team.
To Meta’s credit, the severance is solid: 16 weeks of base pay plus two weeks for every year of service. I get into this in the video. I recently was laid off by another company with no severance at all, so I do not take that lightly. The cuts hit engineering, product, cybersecurity, and content design.
How these layoffs happen
Impacted employees usually lose system access and get the news by email, often on the same day.
Having been through rounds of layoffs, I know the pattern. Teams that are impacted typically get an email and lose access to systems, sometimes before they have fully processed the news. Offices emptied as employees were told to work from home.
There is a harder detail in the reporting. Employees at Meta reportedly pushed a petition to stop workplace tracking data, and the tracking continued. As I understand it, that activity data on how people work was used to help train AI to perform some of these roles. That is a quiet part of the story that deserves more attention than it gets.
The cost nobody is talking about
The deeper driver is not AI doing the work, it is the enormous cost of building and running AI infrastructure.
Here is where I part ways with the simple narrative. I found that some roles were surely eliminated because of AI directly. But the larger force is the cost of AI itself: the compute, the energy, the power, and the infrastructure needed to run these systems at scale. All of it is extraordinarily expensive to operate.
To keep investing in that infrastructure, companies have to cut costs somewhere, and the fastest lever is people. What I learned watching this play out is that the conversation should be about cost and return, not just replacement. Nobody can reliably calculate the return on investment for these AI tools yet, which makes the spending even harder to justify.
What it means going forward
Expect more companies to cut staff to fund AI, even without proof the AI pays for itself.
Meta has been through multiple rounds, and it is not alone. Cisco and others have announced cuts, and the running total for the year has climbed into the six figures. Zuckerberg has said he does not expect further layoffs this year, but as we all know, a statement one week can be followed by an announcement the next.
The pattern will continue. The pattern is likely to continue in a way that touches almost every large technology company, because the incentive structure that produced these cuts has not changed, and you should expect the same logic to reach your corner of the industry eventually. To fund AI, companies cut costs, and the primary way to cut cost is through people. Until someone can measure the actual return on all this AI investment, I expect the layoff-to-fund-infrastructure cycle to keep repeating across the industry.
The takeaway
Meta’s layoffs are real, roughly 8,000 cut with a total impact near 14,000, and the severance is genuinely generous. But reading them as pure AI replacement misses the point. The deeper driver is the massive cost of AI infrastructure, the compute and energy that companies are funding by cutting headcount. Until anyone can prove the return on that investment, expect more companies to keep trimming staff to pay for the buildout, and expect the official story to stay vaguer than the numbers.
Watch the full discussion in my video on Meta’s layoffs, and I walk through the article detail in the video. I also lay out the AI infrastructure cost argument in the video. Here is my question for the comments: do you buy the AI-replacement story, or is this about infrastructure cost? Subscribe for more direct conversations on tech and AI.
