Meta axes hundreds to bankroll its ai obsession, reality labs bleeding out
Meta just trimmed another 1,000 humans from its payroll—roughly the capacity of a mid-size movie theater—so that Mark Zuckerberg can keep pumping 65 billion dollars this year alone into the black-box infrastructure feeding his ai dreams.
The bloodletting lands hardest on reality labs
Inside the VR division that burns 4 billion every quarter, recruiters, sales scouts and hardware engineers across Menlo Park, London and Singapore got the same 6 a.m. e-mail: your badge stops working in 30 days, unless you accept a relocation package to a city you never asked to live in. The company swears the number is “fewer than 1,000,” a rounding error against the 79,000 left standing, yet January already saw another thousand Reality Labs exits. Two purges, same division, same calendar year—do the math and the unit has shrunk by almost 20 percent in twelve months.
Wall Street barely flinched. Shares ticked up 1.4 percent on the news, the market’s way of applauding any sign that Zuckerberg is rationing the oxygen tanks inside his metaverse moon shot. Investors want margin, not metaverse goggles.
Ai talent poached from google arrives before the chairs cool
While HR calculates severance, engineering recruiters have been instructed to back-fill—just not with internal résumés. Three former Google VP-level researchers signed offers last week to run Meta’s new “agentic coding” group, tasked with building large language models that can debug Instagram’s backend while you sleep. The contradiction is brutal: fire the hardware glue-makers, hire the algorithm whisperers.
The official line: “Teams are restructured periodically to stay aligned with strategic objectives.” Translation: if you can’t train a neural net, you’re overhead.
Employees who survived the cull describe a new internal metric circulating on Workplace: ROI per GPU. Every project proposal must now forecast how many Nvidia H100 chips it will unlock. Headcount is secondary.
Meta’s 2024 capital expenditure guide already surpasses the gross domestic product of Slovenia. With this week’s layoffs, Zuckerberg signals he would rather cannibalize his own workforce than throttle the data-center build-out racing toward 600 billion in U.S. infrastructure by 2028. The machines are hiring; the humans are on notice.