AppLovin reported second-quarter results after Wednesday’s close. Revenue was $1.92 billion, up about 53% from $1.26 billion a year ago. Adjusted EBITDA was $1.614 billion, up 58%, at an 84% margin. Earnings came in at $3.76 a share against roughly $3.67 expected. The stock fell as much as 21%, to about $331.57 after hours.
Find the disaster in that paragraph. There isn’t one. Revenue missed a consensus of roughly $1.94 billion by about $16 million — eight tenths of one percent — and every other line either met the plan or beat it. Third-quarter guidance of $2.06–$2.09 billion sits about 0.6% under the Street, with adjusted EBITDA guided to $1.71–$1.74 billion against a $1.74 billion estimate. A rounding error in the revenue line, a rounding error in the guide, and a fifth of the market value gone in a single session.
The detail worth keeping is where that revenue landed. In May, AppLovin told investors to expect $1.915–$1.945 billion for the quarter, with adjusted EBITDA of $1.615–$1.645 billion at an 84–85% margin. It delivered $1.92 billion — inside the range, in the bottom half — and $1.614 billion of EBITDA, one million dollars below the low end of its own forecast. Management kept its word almost to the decimal. That turned out to be the problem.
Our take: AppLovin wasn’t priced on its guidance. It was priced on its habit of shredding it. Last quarter the company put up $1.84 billion, 59% growth, and blew past its own range; investors extrapolated the beat, not the guide. When a stock carries a multiple built on serial upside, the published range stops being a target and becomes the floor — the number below which the story changes. Land on it and you haven’t missed a forecast, you’ve failed to produce the surprise the price required. That’s the whole 21%. The revenue shortfall is $16 million; the repricing is the market marking down the odds of every future beat at once. Anyone holding a high-multiple compounder should read the reaction, not the release: the question is never whether the quarter was good, it’s how much better than good was already in the stock.
The streak was the asset
Growth of 53% on a business throwing off 84% adjusted EBITDA margins is, by any normal standard, exceptional. AppLovin’s ad engine converts revenue to cash at a rate almost nothing else at this scale matches, and the sequential move — up about 4% quarter over quarter — is the arithmetic of a very large base, not deterioration.
But acceleration and deceleration are what get paid, and the guide says the second half looks like the first. A $2.07 billion midpoint against $1.92 billion delivered is roughly 8% sequential growth — healthy, and roughly in line with what the Street already had. Nothing in the release promises the next surprise. For a stock that had already fallen about 24% in the first half of 2026, the absence of a fresh catalyst was itself the catalyst.
What to watch
- Whether the guide gets beaten in Q3. $2.06–$2.09 billion is now the most important number the company has published. Landing inside it again would confirm that in-line is the new normal — and confirm the lower multiple with it.
- The 84% margin line. Q2 EBITDA came in a hair under management’s own floor. Margin steady at 84% means the model is intact; a print starting with an 8 and a smaller second digit would mean the mix is changing.
- Sequential revenue, not year-over-year. The 53% headline is flattered by an easy base. Quarter-over-quarter growth is the honest read on momentum.
- The re-rating itself. Watch where the stock settles over the next few sessions. A snap-back says the drop was positioning; a drift lower says investors have permanently downgraded what they’ll pay per dollar of AppLovin’s earnings.
AppLovin grew revenue more than half in a quarter, printed an 84% EBITDA margin, beat on earnings, and did exactly what it said it would do. It lost a fifth of its value for it. That is not a verdict on the business. It is a reminder that at a certain price, meeting expectations is the bad news — and that you buy the multiple as surely as you buy the company.