Meta’s AI infrastructure buildout is consuming nearly every dollar the company generates in free cash. In the second quarter of 2026, the social media giant reported free cash flow of just $784 million — a staggering 91% collapse from the same period a year earlier, when free cash flow exceeded $12 billion. The numbers reveal a company betting its entire financial cushion on a future that has not yet fully arrived.
The underlying business has not broken down. Net cash from operating activities actually rose 25% year-over-year, reaching $31.86 billion. The problem is what sits on the other side of the ledger: capital expenditures on servers, data centers, and networking infrastructure surged 83% to $31.08 billion in a single quarter, nearly canceling out every dollar of operating cash the business produced.
Meta expects to spend up to $145 billion on capital expenditures across the full year — a figure that dwarfs almost any corporate infrastructure program in recent history. CEO Mark Zuckerberg framed the timing bluntly on the earnings call: “There is a lead time where we’re investing in building out these data centers now,” he said, acknowledging that Meta is “not getting value out of them until they’re online.”
An AI Race With a Steep Entry Fee
The spending surge follows a period in which Meta was widely perceived as trailing OpenAI and Google in the competition to build frontier AI models. Zuckerberg responded in 2025 by reorganizing the company’s AI efforts under a new entity called Meta Superintelligence Labs. The lab has since released Muse Spark — a multimodal model powering Meta AI — followed by Muse Spark 1.1 and an image-generation model called Muse Image.
There are genuine early returns on that investment. Meta reported that AI-driven improvements to content recommendation and ad targeting helped push advertising revenue up 27%. Chief Financial Officer Susan Li insisted the company’s “strong operating cash flow” placed it in a “position of strength” to sustain the infrastructure buildout. That framing is technically accurate, though it glosses over just how little cash actually remains after the spending is done.
Meta’s longer-term ambitions extend well beyond better ad targeting. The company is developing larger foundation models, personal AI agents, business productivity tools, and what Zuckerberg has described as a potential enterprise computing business — investments whose payoff timelines remain genuinely uncertain.
Big Tech’s Collective Bet — and the Market’s Verdict
Meta is not alone in this posture. The largest technology companies are collectively pouring over $700 billion into AI infrastructure in 2026. Google’s free cash flow turned negative for the first time in decades earlier this month, a consequence of similar capital commitments. Microsoft, by contrast, held its spending plan steady and saw its shares jump as a result.
The market’s message to Meta was less forgiving. The company’s stock fell nearly 10% in after-hours trading following the earnings release. Investors appear to be rewarding discipline and penalizing open-ended commitment — at least until the returns become clearer. For a company that controls some of the world’s most-used communications platforms, the scale of this wager is worth watching closely.