Meta’s AI dream reels on more than lab-rat revolt

Its employee count is indeed shrinking as revenue per worker rises. Pushed further, the trend might slash a towering budget for total worker compensation, including healthcare and benefits, that ran to just over $30 billion in the first half of the year. Data centers, however, don’t come free. Capital expenditure has chewed up Meta’s formerly prodigious free cash flow, and investment is estimated to hit nearly $170 billion next year, according to LSEG data. At that pace, increased depreciation and amortization might well eat into any employment savings. Moreover, Meta’s total employee compensation is actually rising. Ignore severance costs, and it still spent nearly 30% more in the first half of 2026 than it did in 2025. AI researchers don’t come cheap, and Zuckerberg has notoriously bid up salaries as much as anyone. Technological revolutions don’t come cheap, whether in terms of money or affected workers. Silicon Valley is built on the idea that they are nonetheless inevitable. Little wonder, then, that tech giants like Meta Platforms are leading the charge on a spending spree to conjure ever-stronger artificial intelligence. Yet boss Mark Zuckerberg’s attempts to capitalize on the technology through mass firings and reorganizations have hit constant, painful snags, Reuters reported on Wednesday. It makes sense: revolutions also tend to be a problem for existing incumbents. Global AI investment will exceed $1 trillion this year, estimates Goldman Sachs. This is, in large part, a gamble that sophisticated models can automate more and more labor. Programming should be ground zero, since AI produces code far faster than humans. As a $1.4 trillion titan of the internet age and one of the leading spenders on data centers, Meta serves as a crucial proof of concept. According to Reuters reporting, the problems also go much, much further than money. Changes to the codebase are proliferating faster than features reaching users, and the security and reliability of automated work appear suspect. Plans to reorganize workers into “pods” and focus in on the very top tier of talent — perhaps justifying those huge salaries — look to have bred confusion. Resentment ran high over job reassignments and the possibility that humans were training their AI replacements. Ultimately, plans for firings and team restructurings were scaled way back. It’s rarely evident how best to use new technologies. Factories and offices took decades to optimally redesign work around electricity and computers, respectively. Drastically refashioning a giant firm like Meta presents inherent problems. Clayton Christensen’s classic book, *The Innovator’s Dilemma*, spells out the issue: existing products, employees and the managerial structures built around them are well adapted to the status quo, and trying to change them in the face of revolutionary industry shifts might just cause them to fall apart. Moving fast and breaking things works for startups with little to lose. Zuckerberg might not be able to afford that mindset.

Meta’s AI dream reels on more than lab-rat revolt

Costs, capex and the limits of head-count savings

Meta’s AI spending spree and the promise of automation

Reuters reporting highlights execution problems and employee backlash

Related News

Kura Sushi USA to host fireside chat at Barclays Global Consumer Conference

Evolution Metals plans adjacent Pohang land buy as part of magnet capacity expansion

Heico Corporation <HEI.N>: UBS Raises Target Price to $396 from $382

Terreno Realty sells Miami industrial property for $21.3 million

Global markets-World stocks steady as Nvidia boost tempers Jackson Hole caution