Meta's Revenue Rose 28% and Left $784 Million in Cash

The easiest way to spot a bad quarter is the revenue line. With the numbers Meta put out on the evening of July 29, that shortcut failed.

Revenue came in at $60.8 billion — up 28% year over year, and ahead of consensus. Daily users reached 3.6 billion.

The next day the stock fell 9%.

After paying every expense and funding every investment in the quarter, Meta had $784 million left; a year earlier that figure was $8.5 billion. That gap is why a line item small next to $60 billion of revenue repriced an entire company.

A Month Earlier, the Same Company Knocked Down Chip Stocks

In early July, Meta announced "Meta Compute," a plan to sell rented GPU capacity to outside customers. That wounded the scarcity thesis which had carried semiconductor stocks for three years, and more than a trillion dollars of sector value was erased: Meta's One Announcement.

Meta's own stock barely flinched; the market read it charitably, as capacity that eventually turns into revenue.

The July 29 report made that harder to hold. The capacity was real: $31.1 billion of capex in the quarter, $50.9 billion in the first half. The revenue attached to it was not.

On the call, Mark Zuckerberg said Meta was fielding offers "meaningfully above what we're paying" for compute, then added that selling it all for near-term profit "would be unwise."

What investors heard: the cash is not coming back this year either.

Both Sides of an Ad Company Got Heavier at Once

Three things happened simultaneously in Meta's earnings that quarter.

Costs grew faster than revenue. Total expenses rose 55% to $42 billion, R&D up 68%; operating income fell 8% and net income 14%, to $15.85 billion. Earnings per share came in at $6.18 against a $7.17 estimate.

Part of that miss is non-recurring: a $2.4 billion legal provision and $1.18 billion for a roughly 8,000-person workforce reduction. Only part.

The cash side emptied out. Operating cash flow was $31.9 billion and capital expenditure $31.1 billion; what remained was $784 million.

The balance sheet took on debt. Long-term borrowings climbed from $58.7 billion to $83.7 billion and cash equivalents shrank 57%. A company that had repurchased shares every quarter for years bought none in the first half.

The Arithmetic of Depreciation: Paid Today, Booked for Five Years

The fragility sits in the five-year assumption itself; the industry generally assumes a five- to six-year useful life.

If the true life is three years, the same $140 billion books not $28 billion but roughly $47 billion a year. Nobody spent an extra dollar; one assumption changed, and a third of the profit went with it.

Capital expenditure is not a one-year question: an order signed today rewrites profit estimates for five years, which is why everyone running a valuation rebuilt the model this quarter.

Same Week, Four Companies, Three Different Reactions

The last ten days of July were big-tech earnings season; all four said they would spend more, and the market did not punish them equally.

Amazon confirmed the group's largest 2026 budget — more than $200 billion — and rose 15%, because cloud revenue grew 37% and beat expectations.

Alphabet fell when it raised its budget on July 22, then recovered after Amazon's numbers; Google Cloud revenue had grown 82%.

Meta has no such line: the return shows up only indirectly, through an advertising business that works somewhat better.

The phrase that recurred across price-target notes: limited visibility into how the AI spending gets monetized outside ads.

The price-target cuts averaged 10% to 20%: the banks were not lowering profit forecasts so much as the multiple they would pay for those profits.

The Other Side: Two Readings of the Same Numbers

The certainty between them: Meta is no longer an advertising company that generates cash but one that routes its cash into a hardware program.

What Is Left: An Ad Company That Borrows

Meta trades roughly 26% below its August 2025 peak and is down 11% year to date, but the durable change is what kind of company it now is.

For years Meta fit a familiar description: low capital intensity, heavy cash generation, steady buybacks — a profile markets pay high prices for, because growth is close to free.

The description is different now: $130 billion to $145 billion of annual spending, part of it financed with debt, and buybacks halted.

That is the profile of an airline or a utility, and markets have never paid the same price for it.

The Lesson: Same Growth, Now Worth Less

Meta's revenue grew 28% this quarter, much as the year before; what changed is not the rate of growth but what it costs.

The second lesson: calling a company profitable from the income statement alone is a verdict handed down without looking at the bank account. Meta reported $15.85 billion of profit and had $784 million left — both true, answers to different questions.

This article is based on reporting from CNBC, Fortune, The Motley Fool and market data providers as of August 2, 2026. Meta's quarterly figures come from the company's own release, and price-target changes from bank notes. Sources differ on the lower bound of the 2026 capex budget, placing it between $130 billion and $135 billion, and on Alphabet's one-day loss of July 22, which is reported as between 5% and 7%. The depreciation calculation is this article's own simplified illustration, not the company's accounting assumption.

Correction (September 13, 2026): prose tightened for rhythm; figures and sources unchanged.