MENLO PARK — On Tuesday evening, the number that rattled Meta Platforms investors was not on the revenue line. Revenue of $60.8 billion for the second quarter grew 28 percent from a year earlier, a rate most companies would treat as the headline. Meta framed it that way. Wall Street’s attention landed somewhere else.
Free cash flow for the quarter came in at $784 million, down from $8.5 billion in the second quarter of 2025. That is not a rounding error or a seasonal quirk. It is what happens when a company generating record advertising revenue redirects most of the proceeds into infrastructure at a scale that dwarfs its own recent history. Capital expenditure in the quarter reached approximately $31 billion, part of a full-year commitment the company now guides at $130 billion to $145 billion. For a business that spent roughly $37 billion on capital expenditure across the entirety of 2024, that is a 3.5-fold increase in two years.
According to the company’s second-quarter earnings release, diluted earnings per share came in at $6.18, down 13 percent from $7.14 a year ago and short of analyst estimates by roughly a dollar. Operating income fell 8 percent to $18.8 billion despite the revenue surge. The shares fell roughly 10 percent in after-hours trading. That reaction is not a verdict on Meta’s advertising business, which by any conventional measure is accelerating. It is a signal about how investors are pricing the risk that $130 billion in annual infrastructure spending does not produce returns on a timeline they can model.
Ad impressions grew 14 percent year over year in the second quarter; average price per ad climbed 12 percent. Advertising revenue of $59.4 billion was up 27 percent. The 3.6 billion people who use at least one Meta platform daily represent a network that has no structural rival in social media. The core business is not what investors are pricing against.
Mark Zuckerberg has been precise about where the money is going, if not about when it returns. “AI is accelerating our core business today, powering our next generation of products, and opening the door to entirely new enterprise opportunities,” he said in the earnings statement. That formulation maps onto three distinct bets. The first—AI improving ad-serving through Advantage+, Meta’s automated campaign system—is already generating measurable pricing improvement. The second—Llama 4, Meta’s open-source model, embedded across Facebook, Instagram, and WhatsApp—is observable if not yet directly monetized in any auditable way. The third, the enterprise opportunity Zuckerberg gestured toward, was the one investors pushed on in the call and received the least on.
Zuckerberg offered a sketch of something larger: a cloud computing business aimed at enterprise customers. The detail was thin. No product description, no timeline, no named competitive context against Amazon Web Services, Google Cloud, and Microsoft Azure, each of which holds years of enterprise sales contracts and data center relationships that Meta does not. The market read the lack of specifics as a cost signal, not a revenue signal—more capital spending, less clarity on when it inflects into profit.
The investor skepticism running through Meta’s report is not unique to Menlo Park. The same pressure on the assumption that AI infrastructure investment justifies its scale on a near-term timeline has been compressing technology valuations across the sector. As the Eastern Herald earlier reported, the Philadelphia Semiconductor Index entered bear market territory on the same trading session, down more than 20 percent from its June record, as investors questioned whether the data-center buildout driving chipmaker valuations would produce the returns those prices embed.
For Meta specifically, the debate resolves around two questions no Q2 report can answer. Whether an open-source model strategy—Llama distributed freely, used widely, but not directly monetized in the conventional SaaS sense—can build durable competitive position in enterprise AI without a clear commercial architecture. And whether a company with no existing cloud infrastructure business can build one at cost while the category incumbents have already locked in most of the large enterprise contracts currently up for renewal.
The scale of the free cash flow compression—from $8.5 billion to $784 million in twelve months—tells investors the investment phase is happening now, fully, without delay. Total expense guidance for the full year was raised to between $165 billion and $169 billion, incorporating $2.4 billion in legal charges recognized in the quarter. Third-quarter revenue guidance of $61 billion to $64 billion signals confidence in ad demand through the back half of the year. What the guidance provides no line of sight on is when the capex program moderates.
Zuckerberg has not indicated it will moderate soon. The cloud business he referenced, if pursued at meaningful scale, would require capital beyond what the current guidance already absorbs. The question investors are left with after Tuesday’s report is not whether Meta’s AI thesis is real—the advertising improvement makes a credible case that it is. The question is whether the company’s spending decisions are calibrated to a market window that still exists, or to one that assumed more runway than the competition has left.
