news

Meta Q2 2026: Revenue Beat, Profit Fell 14%, Free Cash Flow Cratered — Stock Down 9%

Meta grew revenue 28% but profit fell 14% and free cash flow collapsed to $784M as AI capex hit $31B for the quarter. Zuckerberg fielded questions about starting a cloud business live on the call. We break down the numbers and the AI-spending standoff with Wall Street.

7/29/2026

Meta's Q2 2026 report is the clearest example yet of a pattern showing up across the hyperscalers: strong core business, brutal capex bill. Revenue grew 28% to $60.8 billion, beating estimates. Net income fell 14% to $15.8 billion. Free cash flow — $8.5 billion a year ago — came in at $784 million, a rounding error from zero. The stock fell as much as 10% in after-hours trading, settling around a 9% decline the next day, wiping out roughly $15 billion of Mark Zuckerberg's own net worth in the process.

Every free-cash-flow figure in this article is Meta's own reported number, which is what the company published and what the market reacted to. Meta arrives at it by subtracting $962 million of finance-lease principal payments as well as capital expenditure. Our own series uses operating cash flow minus capex for every company, so the same quarter reads $1,746 million on /stocks/meta — $31,862M of operating cash flow less $30,116M of capex. Both are correct; they are answering different questions, and this piece is answering Meta's.

The Headline Numbers

Metric Q2 2026 Q2 2025 YoY
Revenue $60.8B $47.5B +28%
Operating income $18.8B $20.4B −8%
Net income $15.8B $18.3B −14%
Diluted EPS $6.18 $7.14 −13%
Operating margin 30.9% 43.0% −1,214 bps
Free cash flow $784M $8,549M −91%

Revenue growing 28% while net income falls 14% is the whole story in one line. This wasn't a demand problem — it was a spending problem, and Meta made no attempt to hide it.

Where the Money Went

Capex hit $31.1 billion for the quarter alone — nearly double a year ago — against operating cash flow of $31.9 billion. Meta spent almost every dollar its business generated on servers, data centers, networking, and chips. Depreciation is already climbing fast behind it: up 46% YoY to $6.4 billion, a preview of the expense drag still to come as this quarter's spending starts depreciating.

Two one-off items also hit the P&L: a $2.40 billion legal-proceedings charge and $1.18 billion of severance tied to the May 2026 headcount reduction (~8,000 people, mostly rolling off headcount by Q3). Strip those out and the margin story is somewhat less dramatic — but even a "clean" comparison still shows real compression, because the capex-driven depreciation and opex growth are structural, not one-time.

The financing side tells the same story a different way: Meta issued $24.9 billion of new long-term debt this quarter and repurchased zero stock, versus $10.2 billion in buybacks a year ago. CFO Susan Li was explicit that Meta is deliberately shifting toward "a greater mix of debt" to fund infrastructure with long useful lives. That's a defensible capital-structure argument — long-lived assets are reasonable things to debt-finance — but it's also a tell: the company that used to hand cash back to shareholders is now borrowing to keep building.

The Live Question Nobody Had a Clean Answer To

The most interesting moment of the call, by every account, wasn't in the press release — it was Zuckerberg fielding analyst questions about why Meta doesn't just do what Microsoft, Amazon, and Google have done: rent out the infrastructure it's building to outside customers and generate immediate cloud revenue from the buildout, instead of eating the full capex cost with no offsetting revenue stream.

Zuckerberg didn't dodge it. He confirmed Meta has plans to get into the cloud business and said an update is coming, adding: "We're getting a lot of offers for compute at a significant premium for what we paid for it." But he framed it as a side opportunity, not a strategic pivot — arguing "it would be foolish to basically just sell all of the compute and take a short-term profit" when the company expects "a significantly higher margin on selling intelligence rather than selling compute directly." When asked what 2027 capex would look like, CFO Susan Li declined to give a number, saying only that Meta expects to remain "demand constrained" — more profitable uses for compute than compute to satisfy them.

Our take: that's a coherent long-term bet, but it's also a very expensive one to ask public-market investors to sit through without a number attached. Wolfe Research is already modeling 2027 capex at $233 billion — nearly double this year's raised range. Every hyperscaler is currently asking investors to trust that AI infrastructure spend converts to durable revenue eventually; Meta is the one hyperscaler that doesn't have a cloud business to point to as the near-term proof point while everyone waits.

Wall Street's Verdict: Still Bullish, Just Less Sure of the Price

The analyst reaction after the print was unusually uniform: six of seven major firms cut their price targets, and every single one kept a Buy/Outperform rating. Average price target fell to about $795 from $856 — still roughly 50% above the post-earnings price near $531. Nobody thinks the business is broken. Several specifically cited the same concern: little clarity on 2027 infrastructure spending, and no clear timeline for when new AI products convert into revenue that shows up on the income statement rather than the capex line.

That's a fundamentally different critique than "the ad business is slowing." It isn't — average price per ad rose 12% YoY against ad impressions up 14% and audience (Family DAP) up only 3%, meaning monetization efficiency, not user growth, is driving the ad engine right now, and it's still working. The market's real complaint is about capital allocation and disclosure, not demand.

The Sector Backdrop Makes This Sharper

Meta didn't report into a vacuum. The same week, Microsoft's stock rose roughly 8–9% on an Azure reacceleration story, and Alphabet had already raised its own capex guidance the week before Meta reported. Every hyperscaler is running the same AI-infrastructure arms race; Microsoft and Google are just further along in showing investors a cloud-revenue offset for the spend. Meta's stock move looks less like "AI capex is bad" and more like "AI capex without a monetization story investors can see yet is punished, and AI capex with one gets rewarded" — a valuable contrast for anyone trying to read across this earnings season rather than ticker by ticker.

What to Watch

The Bottom Line

Meta's advertising business is healthy — arguably healthier than the headline profit numbers suggest, since price-driven monetization is exactly the kind of growth that compounds. What the market punished wasn't demand; it was the absence of a near-term answer to "how does this capex turn into revenue we can see." Zuckerberg all but confirmed a cloud business is coming, which should eventually give Meta the same investor-facing offset Microsoft and Google already have — but "eventually" is doing a lot of work in that sentence, and Wall Street's own price targets (down, but still implying ~50% upside) suggest this reads more as a patience problem than a broken thesis. The next few quarters — and whatever number Meta eventually attaches to 2027 capex — will decide which read was right.


Meta Platforms (NASDAQ: META) reported Q2 2026 revenue of $60.8B (+28% YoY) and diluted EPS of $6.18 (−13% YoY). Free cash flow fell to $784M from $8.5B a year earlier. Shares fell roughly 9% following the report, trading near $531.