Meta's Cash Flow Cratered 91% Even as Revenue Hit Record Highs

Meta just posted $60.80 billion in quarterly revenue, up roughly 28% year-on-year and yet another record for the company. Shares still dropped about 10% anyway. The disconnect points to a number buried further down the earnings report: free cash flow of just $784 million by the end of Q2 2026, a 91% collapse from the $8.55 billion Meta had on hand at the same point last year.

That's a lot of money to lose track of, and the explanation isn't a mystery. Meta has been pouring cash into two expensive bets: its long-running Reality Labs division, which builds the VR and AR hardware behind products like the Quest headset, and a much larger buildout of AI infrastructure. The latter is the bigger story right now, and it's not slowing down. Meta expects to spend somewhere between $130 billion and $145 billion this year alone on AI infrastructure.

For a company already spending at that scale, the shrinking cash cushion is the kind of number that makes investors nervous, and it's why CEO Mark Zuckerberg and CFO Susan Li spent the earnings call making the case that all that money is doing real work.

The pitch to investors

Zuckerberg's argument leaned heavily on advertising, still the engine of Meta's business. "Our investments in AI are accelerating every major part of our core business," he told investors. He pointed specifically to how large language models are now being used to improve ad targeting, saying Meta has "expanded the context that we can take into account around a person's organic and ads activity to determine an ad's relevance, driving significant increases in relevance and conversions on both Facebook and Instagram."

The numbers back up at least part of that claim. Advertising alone brought in $59 billion of Meta's quarterly revenue. Zuckerberg went further, claiming Meta's ad business is growing faster, in dollar terms, than any other publicly reported ad business in the industry. "So these AI investments are paying off," he said.

Beyond advertising, Zuckerberg floated a second, less proven revenue path: selling AI capacity and tools directly to businesses. That could mean APIs, AI agents built for enterprise customers, or even selling raw compute power outright. Li framed this as a response to an industry-wide shortage. "The industry has under-built historically for the wave of AI adoption, making existing capacity, including our own, extremely valuable," she said.

Betting bigger, not smaller

That framing matters, because it's effectively Meta's justification for continuing to spend at a massive scale rather than pulling back. If capacity is genuinely scarce across the industry, then owning more of it, even at enormous upfront cost, looks less like overreach and more like securing a scarce resource before competitors do. Zuckerberg made the same point plainly: "As AI usage in our products and businesses continues to ramp, we continue to invest aggressively in infrastructure to meet the demand."

The trouble is that this is largely a story about the future. The revenue Meta is reporting today comes overwhelmingly from an ad business that existed well before this round of AI spending, even if AI tools are now sharpening it. The enterprise compute-selling business Zuckerberg described is still mostly aspirational, a "large opportunity" rather than a proven revenue stream with numbers attached to it in this report.

That gap between the pitch and the current cash flow is exactly what spooked investors enough to knock 10% off the share price despite a record quarter. Meta is asking shareholders to trust that $130-145 billion in infrastructure spending this year will eventually show up as returns, at a moment when the company's actual cash reserves just shrank by 91%.

What happens next

Meta isn't alone in making this bet. The broader AI industry has been operating on similar logic for a while now: build capacity first, worry about proving the return later, because whoever under-invests risks getting locked out of the next wave of demand. Meta's own executives are explicitly framing their spending in exactly those terms.

Whether that logic holds depends on things that haven't happened yet. Enterprise customers actually buying compute and agents at scale. Ad revenue growth continuing to outpace the industry as AI tools get more sophisticated. Investors deciding a shrinking cash cushion is an acceptable price for a bigger bet on the future.

For now, Meta's answer to any doubts is more spending, not less. The next few quarters will show whether that confidence is warranted or whether the gap between what Meta is promising and what its cash flow shows keeps widening.