Meta forecasts weak revenue as AI spending soars to $145 billion

Meta's earnings reveal a company caught between present constraints and future ambitions. The timing mismatch is structural: Zuckerberg realized last summer that Meta risked losing ground in the AI race and accelerated spending on both talent and computing capacity. Because building data centers takes years, Meta scrambled to rent compute from third-party cloud operators—a necessary short-term fix that nonetheless locks the company into double-paying for infrastructure (renting now, building and operating its own later) with no immediate monetization path. This is the core reason investors are alarmed: Meta is not investing in proven growth engines but rather betting that AI superiority will preserve its ad business and unlock new revenue streams that do not yet exist at scale.

Zuckerberg's defense—that Meta must own its AI stack end-to-end, from chips to models, just as it controls its social platform infrastructure—is historically sound. Facebook's early advantage came partly from engineering control and operational efficiency. However, the comparison also carries a warning: fifteen years ago, Microsoft made a similar bet on vertical integration and control at the moment the industry was shifting to mobile and the cloud. Meta's claim that it can build personal agents and enterprise tools rests on the same full-stack logic, but trajectory and relevance in a fast-moving AI field matter more than sheer volume or internal optimization. The company's ad business, while still outpacing competitors in growth, is no longer delivering the momentum boost (simultaneous impressions and price-per-ad growth) that would make investors tolerate capex spending on faith.