AppLovin Q2 2026: growth stepped down to 53%, then 46-48%

AppLovin’s revenue came in roughly $16–26 million below what analysts expected (our arithmetic), and the report is still about a slowdown.

Revenue was $1.924 billion against a consensus of $1.94 billion (FactSet, as reported by MarketScreener) or $1.95 billion (StockStory), depending on whose survey you use. The shortfall is about 1% (our arithmetic). Understanding why a shortfall that small is the centre of the report is the useful part.

	Q2 2026	Consensus
Revenue	$1.924B	$1.94B (FactSet, via MarketScreener); $1.95B (StockStory)	miss
Diluted EPS	$3.76	$3.75 (StockStory)	in line
Adjusted EBITDA	$1.614B	$1.64B (StockStory)	slight miss
Q3 revenue guide	$2.055–2.085B	$2.08B (StockStory)	midpoint ~0.5% below (our arithmetic)
Q3 adj. EBITDA guide	$1.71–1.74B	$1.75B (StockStory)	below

Every growth number behind that table is superb. Revenue up 53%. Diluted EPS up 57% (our arithmetic, from $2.39). Net income up 55%. Adjusted EBITDA up 58%, at an 84% margin — a level almost no company of any size achieves.

The lines that fall short of consensus — revenue and adjusted EBITDA — do so by 1–2%.

Did AppLovin meet its own guidance in Q2 2026?

On revenue, yes; on adjusted EBITDA, narrowly not. Revenue of $1,924 million came in within the $1,915–1,945 million range AppLovin guided to in its May 6, 2026 8-K, $6 million below the midpoint (our arithmetic), and was up 53% from $1,259 million a year earlier. Adjusted EBITDA of $1,614 million ($1,613.8 million unrounded) came in just below the bottom of the $1,615–1,645 million guided range, while the 84% adjusted EBITDA margin was within the guided 84–85%. The guidance did not include EPS; diluted EPS was $3.76, against $2.39 a year earlier.

The quarter is not the story. The slope is.

Put the last four data points in a row:

Period	Revenue	YoY growth
Q4 2025	$1.658B	+66%
Q1 2026	$1.842B	+59%
Q2 2026	$1.924B	+53%
Q3 2026 (guided)	$2.055–2.085B	+46–48% (our arithmetic, on $1,405 million in Q3 2025)

Three consecutive steps down, each roughly six to seven percentage points. That is not a quarter that missed; that is a trend line with a visible gradient, and management’s own guidance extends it rather than breaking it.

A company growing 50%+ is judged on how long the rate persists, and a single quarter’s revenue says little about that. The rate of change of the growth rate is the only evidence the reported numbers offer about duration, so a revenue miss of about 1% matters mainly as the latest data point on a slope.

The explanation management gave is the real news

On the earnings call, as reported by BigGo Finance, CEO Adam Foroughi did not attribute the shortfall to weak advertiser demand. He attributed it to the cadence of AI model improvements in the gaming business — upgrades arrived later in the quarter than usual, and a meaningful upgrade went live just after the quarter closed, which he said was already powering a strong start to Q3. None of this is in the filed release.

Take that seriously, because it describes a business whose revenue is unusual in a specific way.

Most advertising companies are levered to budgets: when advertisers spend more, revenue rises. AppLovin is levered to that too, but the marginal driver management is pointing at is its own model release schedule. A better model bids more accurately, wins more auctions at better prices, and lifts revenue mechanically without a single new advertiser. Which means:

• Revenue is partly a function of internal R&D timing, not just external demand. Model shipping dates slip. Demand curves do not slip in the same lumpy way.
• The explanation is testable and it is testable soon. Foroughi committed to a specific Q3 range and said the delayed upgrade is already live. If the shortfall really was timing, Q3 catches up and this quarter is noise. If Q3 also lands at or below the low end, the timing story becomes the first draft of a structural one.
• There is a harder version of the same fact. Model-driven gains have a natural shape: the first improvements to a ranking system harvest the largest inefficiencies, and later ones harvest less. If the cadence is slowing because each successive upgrade is worth less, then “timing” and “deceleration” are the same phenomenon described optimistically.

Nothing in this quarter distinguishes those two readings. The reported numbers cannot tell them apart yet, and the guidance says they will not be able to until November.

What is going right, and it is a lot

The bearish reading above should not obscure an operating record that is close to unique.

An 84% adjusted EBITDA margin. $1.614 billion of adjusted EBITDA on $1.924 billion of revenue. This is the economics of a toll booth, not an ad agency: the model is built once and every incremental auction runs across it at almost no marginal cost. For reference, most excellent software businesses top out in the 30s and 40s.

Operating leverage is still improving. Revenue grew 53%; adjusted EBITDA grew 58%. Even in a quarter that missed on revenue, profit grew faster than sales.

The consumer/e-commerce engine is working. Management said on the call (per BigGo Finance) that advertiser spend in the consumer vertical set a record — 28% above the seasonally strong Q4 2025 peak; the figure is not in the filed release. Q4 is the quarter with Black Friday in it. Beating it in a June quarter by more than a quarter is a genuine structural signal, not a seasonal one.

The self-serve platform is now open. In June 2026 AppLovin opened its self-serve ads platform, AppLovin Ads Manager, to the public (per BigGo Finance; PPC Land reported the June opening as scheduled at the Q1 report). Until then, access ran through managed relationships and, since October 2025, a referral system. The quarter just reported contains only its first weeks. The growth vector is being widened at exactly the moment the gaming vertical had a timing problem.

Read those together and the mix story emerges: the new engine accelerated while the old one stumbled on release timing. That is a materially better position to be in than the reverse.

Quality of earnings: 84% margin, 45% cash

Here is the number that deserves more attention than it got.

	Q2 2026	% of revenue
Adjusted EBITDA	$1.614B	84%
Free cash flow	$863M	45%

A roughly $751 million gap (our arithmetic). The filings account for most of it: the two 10-Qs show $533 million of income taxes paid (net of refunds) and $98 million of interest paid in the quarter (our arithmetic: six-month figures of $639.8 million and $99.7 million less first-quarter figures of $106.7 million and $1.7 million). On the call, management described the tax and interest payments as timing and the constraint as temporary (per BigGo Finance).

That is a reasonable explanation and it also tells you something structural. Interest payments imply debt — AppLovin carries $3.5 billion of long-term debt while it repurchased $1.53 billion of stock in the first half — and tax is not an adjustment, it is a cost. Adjusted EBITDA excludes both. So the 84% figure, accurate as it is on its own terms, overstates what actually reaches an equity holder by roughly half in this quarter.

None of that makes AppLovin a poor business; 45% free cash flow conversion is still excellent. But anyone anchoring on “84% margins” as the reason to own it should anchor on 45% instead, and watch whether the gap narrows next quarter as management implied it would.

The targets on the record

Target	What management said
Q3 2026 revenue	$2.055–2.085 billion (8-K), implying 46–48% year-over-year growth (our arithmetic)
Q3 2026 adjusted EBITDA	$1.71–1.74 billion, an 83% margin (8-K)
The delayed model upgrade	went live just after quarter-end and, management said, was already powering a strong Q3 start (earnings call, per BigGo Finance)
Self-serve availability	AppLovin Ads Manager open to the public since June 2026 (per BigGo Finance)

The first and third of those are linked, and together they form an unusually clean test. Management named the cause, said the cause has already been removed, and put a number on the quarter that follows. There is very little room to reinterpret the result in November.

The bear case, stated fairly

1. Three consecutive decelerating quarters, with the company’s own guidance supplying the third.
2. The cause is internal, not external. A demand shortfall can be blamed on the market. A model-cadence shortfall is a statement about the company’s own execution engine.
3. Adjusted EBITDA guidance was set below consensus too — a $1.725 billion midpoint against StockStory’s $1.75 billion — and the guided margin steps down to 83%, so the deceleration is not confined to the revenue line.
4. Free cash flow is 45% of revenue against an 84% EBITDA margin. Interest and tax are real.
5. The e-commerce vertical is now carrying the growth story, and it is the newer, less-proven half of the business — in a segment where AppLovin competes with the largest ad platforms in the world rather than owning a closed loop.
6. A small miss now carries more weight. With growth stepping down three quarters running, the next small shortfall will read as one more data point on the slope.

What to check at the Q3 report

• Revenue against $2.055–2.085 billion, and where in the range. This is a credibility test as much as a number. Management attributed a miss to timing and said the timing has resolved; the low end would contradict that.
• The growth rate itself. 46–48% guided. Landing above it breaks the three-quarter slope. Landing below extends it to four.
• The gaming-versus-consumer split. Whether gaming recovers as the delayed model upgrade annualises, or whether e-commerce is quietly becoming the only growth engine.
• Free cash flow conversion. Whether the gap to adjusted EBITDA narrows toward the 60s, as the “temporary” framing implies.
• Whether “model cadence” appears in the language again. Once is timing. Twice is a business model characteristic.
• Adjusted EBITDA against $1.71–1.74 billion, and whether the margin holds at the guided 83% while the self-serve platform onboards a much broader, likely lower-quality advertiser base.

The read

AppLovin reported a quarter that would be the best in most companies’ histories and missed consensus revenue by about 1%. Both halves of that sentence are accurate.

The operating business is exceptional: 53% growth, 84% adjusted EBITDA margins, a record consumer vertical, and a self-serve platform that just opened to the public. The open question none of those facts answer is how many more years of this are there? — and the four data points now visible all point the same way.

The useful thing to track is therefore not the beat or the miss. It is whether Q3 breaks the 66% → 59% → 53% → 46–48% ladder or extends it, and whether management’s explanation — that this was a release calendar, not a ceiling — survives a quarter in which it can be checked.

---

Sources and method. Company figures — revenue of $1,923.7 million and its 53% growth, diluted EPS of $3.76 (and $2.39 a year earlier), net income of $1.267 billion, adjusted EBITDA of $1,613.8 million and the 84% margin, free cash flow of $863.3 million, long-term debt, first-half share repurchases, and the Q3 revenue, adjusted EBITDA and 83% margin guidance — are from AppLovin’s Form 8-K filed 2026-08-05 (Exhibit 99.1, accession 0001751008-26-000057) for the quarter ended 2026-06-30, read directly from SEC EDGAR. Cash paid for income taxes and interest is from the Forms 10-Q filed 2026-08-05 (accession 0001751008-26-000059) and 2026-05-06 (accession 0001751008-26-000044). Prior-quarter revenue and growth rates are from the Forms 8-K filed 2026-05-06 (Q1 2026, accession 0001751008-26-000042, also the source of the Q2 guidance), 2026-02-11 (Q4 2025, accession 0001751008-26-000005) and 2025-11-05 (Q3 2025, accession 0001751008-25-000079).

Not company filings, and attributed accordingly: consensus estimates — FactSet revenue consensus of $1.94 billion as reported by MarketScreener, and StockStory’s consensus for revenue ($1.95 billion), EPS ($3.75), adjusted EBITDA ($1.64 billion) and the Q3 revenue ($2.08 billion) and adjusted EBITDA ($1.75 billion) guides; and statements made on the earnings call that are not in the filed release — the model-cadence explanation, the post-quarter model upgrade, the record consumer-vertical spend 28% above the Q4 2025 peak, the timing explanation for free cash flow, and the June Ads Manager launch — which are taken from BigGo Finance’s call coverage, with PPC Land for the platform-opening background. This piece makes no claim about the share-price reaction. Consensus revenue is given as two figures because FactSet and the survey StockStory used did not agree, which is why the shortfall is stated as a range.

Whether the quarter met AppLovin’s own guidance is judged only against that previously issued guidance (Form 8-K, filed 2026-05-06) and against the year-ago quarter; no consensus estimate enters that verdict. Our arithmetic: revenue landing $6 million below the $1,930 million midpoint of the guided range; the $16–26 million and about 1% revenue shortfall against the two consensus figures; the 57% diluted EPS growth; the 46–48% growth implied by the Q3 guide on $1,405 million of Q3 2025 revenue; the $1.725 billion Q3 adjusted EBITDA midpoint; the roughly $751 million gap between adjusted EBITDA and free cash flow; and the second-quarter tax ($533 million) and interest ($98 million) payments, each the six-month figure less the first-quarter figure.

The interpretation — that the slope matters more than the quarter, and that model-cadence risk is structurally different from demand risk — is our analysis, not a company statement, and is not a forecast.

Related: Duolingo Q2 2026 and Spotify Q2 2026, two more consumer-tech companies whose guidance, rather than the quarter, is the open question, and Shopify Q2 2026, a platform balancing growth against free cash flow.

Tradune shows what you actually hold across every account, and whether a position has beaten a plain index fund since you bought it. Educational analysis, not investment advice.