META Raises Capex to $145B, Up 30 Percent YoY, While Cutting 10 Percent of Staff

Meta just executed a workforce reset that masks a deeper strategic pivot: the company is laying off 10% of staff while simultaneously tripling its AI infrastructure bet. Mark Zuckerberg announced 8,000 layoffs, the cancellation of 6,000 open roles, and the redeployment of 7,000 employees into AI-focused positions. At the same time, Meta guided capital expenditure to $145B, a staggering 30%+ increase year-over-year. This is not austerity; it is a ruthless reallocation of resources from legacy operations into artificial intelligence.

The subtext is that Meta is building its own AI chips, its own data center infrastructure, and its own large language models. For years, Meta paid Nvidia hundreds of millions for GPU capacity. Now, the company is betting it can design chips in-house, train models at lower unit costs, and own the entire stack. The layoffs are a byproduct: accountants, recruiters, and middle managers in non-core business units are being culled to fund the AI arms race. This is Zuckerberg's gamble that vertical integration will give Meta an unassailable edge in the AI race against Microsoft, Google, and Amazon.

The stock market is struggling to price this thesis. Meta trades at a substantial discount to Nvidia despite delivering record revenue growth and now committing $145B to capex. Retail investors see the 10% layoff and sell. Institutional investors see the $145B capex and buy, betting that the company will emerge as a compute giant with moatA sustainable competitive advantage that protects long-term returns on capital.. The question is whether Meta can actually execute: can it design competitive chips, build reliable data centers, and train foundational models faster and cheaper than Nvidia can supply GPUs? If yes, Meta is worth significantly more. If no, the company has starved its core advertising business of talent and wasted capital.

For the AI ecosystem, Meta's capex is a signal that hyperscalers are no longer content to be Nvidia customers. They are becoming Nvidia competitors. This is bearish for Nvidia's pricing power over the long term, even if short-term capex volumes stay strong. It is also a sign that the AI capex cycle may be hitting an inflection point where growth rates start to slow as companies shift from buying to building.