When Reuters reported last week that Meta is planning layoffs affecting up to 20% of its roughly 79,000 employees, the stock went up nearly 3%. That reaction tells you almost everything about how investors currently view the relationship between AI spending and human headcount at large technology companies. The layoffs are not a sign of distress — they are a sign that the strategy is working as intended. Meta has been explicit for more than a year that AI would allow it to run with fewer people while achieving more. The market is pricing that narrative as credible.
The numbers sit in an uncomfortable juxtaposition. Meta's AI-related capital expenditure for 2026 is projected at $115 to $135 billion — nearly twice what the company spent on all capital investment in 2025, and a figure that puts it among the largest infrastructure investment programmes of any company in any industry in history. Simultaneously, the company is cutting potentially 15,000 to 16,000 jobs, with the Reality Labs VR division already seeing more than 1,000 cuts and the AI unit itself having shed 600 positions in October after Alexandr Wang, the former Scale AI CEO now running Meta's superintelligence effort, consolidated leadership. Financial analysts estimate the restructuring could save $7 to $8 billion annually — a material offset, though not close to covering the incremental AI infrastructure spend.
The logic connecting the two moves is straightforward, even if its human consequences are not. Meta's bet is that AI can automate enough of what its current workforce does that a smaller headcount produces the same or greater output — freeing capital that would have gone to salaries to go to GPUs instead. The efficiency gains from AI are being recycled into more AI investment, not distributed as profit or returned to employees in the form of maintained employment. This is not particular to Meta; it is the explicit strategy that every large technology company is pursuing to varying degrees, and Meta is simply executing it at visible scale and with unusual candour.
What makes Meta's version of this story distinctive is the product context. Meta's AI infrastructure does not primarily serve external customers in the way Amazon Web Services or Microsoft Azure do. It serves the 3.5 billion people who use Facebook, Instagram, WhatsApp, and Threads — making their advertising more targeted, their feeds more engaging, and their experiences more personalised. The efficiency gains from AI at Meta are not primarily about reducing the cost of enterprise software; they are about making the world's largest social platforms more effective at the business of attention capture. The scale of the AI investment reflects the scale of that opportunity: if you can marginally improve engagement or ad conversion across 3.5 billion users, the economics are extraordinary.
The Zuckerberg superintelligence vision adds a longer-horizon dimension that the layoffs coverage tends to crowd out. Meta is not just automating its current operations; it is trying to build what Zuckerberg has called "personal superintelligence" — AI systems sophisticated enough to serve as meaningful intellectual partners for individual users. The company has offered pay packages worth hundreds of millions of dollars to recruit the researchers it thinks can get there. Wang and former GitHub CEO Nat Friedman are running the effort. This is a frontier research programme funded by a company that generates over $160 billion a year in revenue from advertising, which gives it a different kind of staying power than a startup dependent on venture capital rounds.
The signal that the layoffs news sent, more than anything else, is that the market has internalised a particular story about the near future: AI spending is investment, headcount reduction is efficiency, and the combination produces higher earnings. That story may prove correct. It may also prove to be an accounting fiction that collapses when the AI productivity gains are slower or smaller than projected, leaving companies with both the infrastructure debt and the reduced workforce capacity to cope with it. The 3% stock bump on layoffs news suggests investors currently believe the first scenario. History would counsel some caution about narratives of this kind, though it would also note that the platforms Meta runs are genuinely enormous and the AI tools being deployed into them are genuinely capable.