AppLovin (APP) · Deep analysis

Report previewAPP.USReport date Last updated
← Back to reports
Report contents

At a glance

  • Business model: AppLovin earns the spread between what performance advertisers pay for mobile ad impressions and in-app actions and what app publishers receive, so revenue is a take on other companies' marketing budgets rather than a subscription, license, or consumer payment. Sustaining it depends on inventory supply and auction density, and the company does not disclose its take rate.
  • Financial quality: FY2025 operating cash flow of USD 3.97bn exceeded net income of USD 3.33bn, but 1H2026 reversed that, with net income of USD 2.48bn against operating cash flow of USD 2.16bn, and Q2 2026 at 0.69x. Equity rebuilt to USD 3.16bn by 6/30/26, cash reached USD 3.05bn, and long-term debt was last disclosed at USD 3.51bn in mid-2025.
01

Business model

AppLovin sells performance advertising for mobile apps. Advertisers — mostly game studios, and more recently e-commerce and other direct-to-consumer marketers — pay for installs and in-app actions, and AppLovin keeps the spread between what those advertisers pay and what the app publishers that supply the ad space receive. It is a platform, not a media owner: no subscriptions, no licenses, no consumer payments. Revenue is a take on other companies' marketing budgets. The company is in an expansion stage built on software, and the paying customer is the app marketer, not the phone user who sees the ad.

The take rate is the whole model, and the filings don't quantify it

AppLovin's revenue is gross advertiser spend minus the share paid to publishers and other direct costs of delivering an ad. That net figure is the take rate, and it drives gross margin, EBITDA, and cash.

Two things move it. Direct cause: advertiser demand per unit of inventory — more bidders per impression means each dollar of advertiser spend leaves more behind. Structural cause: auction density — how many qualified bidders its ranking model can place in the same auction, because a denser auction clears at a higher price without requiring higher publisher payouts.

The tradeoff is standard for a two-sided platform. Raising the take rate lifts revenue now, but it weakens publisher returns, and weaker publisher economics eventually shrink the inventory that makes the auction valuable. There is no disclosed take rate. AppLovin reports revenue and cost of revenue, not gross advertiser spend, so the ratio has to be inferred and compared only on a consistent basis across periods.

The current source is the Form 10-Q for the quarter ended June 30, 2026, filed August 5, 2026. The excerpt available here is the cover page and forward-looking-statement list; it confirms the Nasdaq listing (APP), the Palo Alto headquarters, and 304.4 million Class A and 30.2 million Class B shares outstanding as of July 31, 2026. It does not contain the segment revenue, direct cost, or geographic tables, so segment mix and any advertiser or publisher metrics have to be pulled from the full filing before the trend can be judged.

Supply, not advertiser demand, is the binding constraint

Ad budgets are abundant. Good mobile ad inventory is not. AppLovin's growth depends on publishers continuing to route impressions through its mediation and auction. Direct cause: networks that fill demand and pay reliably keep the integration. Structural cause: publishers can run several networks side by side, so AppLovin's position rests on being the highest-yielding bidder rather than on a locked-in contract. Competition therefore appears first at the publisher — as higher payouts or better fill — and only later in AppLovin's reported revenue.

That is the tradeoff worth watching. Paying more for inventory protects the supply base but compresses the spread, and the financial consequence lands in software gross margin before it lands in revenue growth. A second-order effect matters too: publishers earning less from the network route fewer impressions, which reduces auction density, which lowers the take rate. The loop can run in either direction.

Non-gaming advertisers are the growth story that still needs proof

The bull case is not more game studios; it is that the same ranking model works for e-commerce, retail, and other direct-response advertisers with larger, less seasonal budgets. Direct cause: those advertisers have measurable conversion events, so a performance model can learn from their data. Structural cause: their spend is less tied to app-store release calendars than game marketing, which smooths the revenue base.

The tradeoff is expansion versus execution. New advertiser categories need measurement and creative the model was not originally trained on, and early spend is often opportunistic. If e-commerce budgets persist past the test, revenue grows without requiring more inventory. If they don't, sales effort has bought a trial rather than a franchise. Concentration cuts the other way as well: growth that still leans on a few large game publishers is more cyclical than headline growth suggests.

Working capital and geography are reporting effects, not the business

Two mechanics can mislead. Revenue is recognized when an ad is served, while cash arrives on advertiser payment terms and publishers are paid on their own schedule, so cash timing can diverge from reported revenue before anything about the business changes. And because a mobile ad platform bills advertisers where the contracting entity sits, geographic revenue splits describe where marketers are based more than where users see ads. Cash conversion and balance-sheet detail belong in their own module.

What would change the judgment

Gross margin compression alongside steady revenue would say the company is buying growth from publishers rather than earning it. Slower revenue with stable margins would say the market, not the model, is the constraint. Any disclosure of advertiser counts or spend cohorts would settle the volume-versus-price question that current reporting leaves open.

Questions worth following

  • Can AppLovin turn e-commerce and other non-gaming advertisers into a durable second demand pool?

    If these advertisers keep spending past the initial test, revenue grows without needing more ad inventory and depends less on game-marketing cycles. If the budgets fade, growth stays tied to a handful of large game publishers and their release calendars.

  • Can AppLovin hold its mediation position as publishers gain more network alternatives?

    Publisher supply is the input AppLovin cannot manufacture. If rival networks win placements, auction density falls and the take rate follows, however much advertiser demand exists. That would hit revenue growth and margins at the same time.

  • Does AppLovin grow ad revenue mainly by adding advertisers or by charging more per impression?

    Volume-led growth is more durable because it comes from a widening customer base; price-led growth can stall once advertiser returns fall below their hurdle. The two look identical in reported revenue, which is why the distinction shapes how long growth can run.

02

Financial quality

AppLovin's earnings are large, cash-backed, and produced almost without fixed investment. Two things make them harder to grade than the headline numbers suggest: equity is small relative to the earnings it supports, and cash conversion fell below one times earnings in the first half of 2026. On the supplied evidence, neither looks like deterioration in the underlying business yet, but each needs a specific check.

Five years that end at an 88% gross margin
Metric (USD bn unless noted)FY2021FY2022FY2023FY2024FY20251H2026 (six months to 6/30/26)Assessment
Revenue2.792.821.843.225.483.77Strong
Gross margin64.6%55.4%67.7%75.2%88.0%88.7%Strong
Operating margin5.4%-1.8%41.8%59.3%75.7%77.8%Exceptional
Net margin1.1%-6.7%19.6%49.1%60.8%65.6%Exceptional
ROE, as supplied3.6%-9.5%22.6%134.7%Not disclosedNot disclosedDistorted by a small equity base
ROA, as supplied0.9%-3.2%6.4%28.1%Not disclosedNot disclosedStrong
Liabilities / assets65.4%67.5%76.6%81.4%70.5%61.8%Adequate, improving
Operating cash flow0.360.411.062.103.972.16Exceptional
Free cash flow0.360.411.062.093.942.16Exceptional
Operating cash flow / net income12.0xn.m.2.9x1.33x1.19x0.87xStrong, weakening
Goodwill0.971.821.841.801.541.52A large share of equity
Total equity2.141.901.261.092.143.16Improving

The series is not comparable throughout. Revenue falls 35% in FY2023 while gross margin rises about 12 points, which points to a change in what the reported revenue line contains rather than an organic move. Treat FY2022-to-FY2023 comparisons cautiously.

Six quarters, and the cash conversion that slipped

Single-quarter figures, as reported.

USD bn unless notedQ1'25Q2'25Q3'25Q4'25Q1'26Q2'26Assessment | Trend
Revenue1.161.261.411.661.841.92Strong | Accelerating
Gross margin87.1%87.3%87.2%88.6%89.1%88.5%Strong | Flat at a high level
Net income0.580.820.841.101.211.27Exceptional | Rising
Total assets5.715.966.347.267.718.27Growing with cash, not capex
Total liabilities5.134.794.875.125.345.11Adequate | Stable
Cash0.551.191.672.492.763.05Strong | Building
Operating cash flow0.830.771.051.311.290.87Weakening relative to earnings
Financing cash flow-1.00-0.54-0.56-0.49-1.01-0.57Persistent outflow
Operating cash flow / net income1.430.941.251.191.070.69Weak | Diverging

Only six quarters are supplied; a full eight-quarter block would require FY2024 quarters that are not in the data. Net PP&E is not disclosed in any period, so fixed assets cannot be shown.

What is actually driving these numbers

ROE is not the signal; ROA and cash are. The FY2024 ROE of 134.7% sits on equity of $1.09bn that buybacks had shrunk from $1.26bn the year before. Earnings grew while the denominator fell, so ROE inflates mechanically. ROA of 28.1% and $3.94bn of FY2025 free cash flow are the better evidence of returns. By 6/30/26 equity had rebuilt to $3.16bn and liabilities had fallen to 61.8% of assets, so this leverage story is unwinding, not worsening, assuming the debt line is intact.

Growth is absorbing working capital. In Q2 2026, revenue rose 53% year over year to $1.92bn while operating cash flow of $869M fell short of net income of $1,267M. The last disclosed receivable balance — $1.58bn at 6/30/25 against that quarter's $1.26bn of revenue — implies roughly 114 days of sales outstanding. If collections run on similar terms at today's revenue, receivables are near $2.4bn and each quarter of growth ties up more cash. That is temporary if it is timing; it is a quality problem if days outstanding are lengthening.

Margins are real, but the headline metric is wider than GAAP. Q2 2026 adjusted EBITDA of $1.61bn was 84% of revenue against a 66% GAAP net margin. The gap is interest, taxes, depreciation, and stock-based compensation — and SBC is a genuine cost that the adjusted figure excludes. Guidance of an 83% adjusted EBITDA margin for Q3 2026 keeps that spread in place. Repurchases then partly recycle the compensation cost back into the share count.

Capital returns run on operating cash, not borrowing. Financing outflows were $2.59bn in FY2025 and $1.58bn in 1H2026 — 65% and 73% of operating cash flow — with no evidence they were debt-funded. The open question is price: 1.1 million shares for $551.3M in Q2 2026 implies roughly $500 per share. At that level repurchases only create per-share value if they exceed the equity issued to employees.

Red flags worth checking
  • Cash conversion. Q2 2026 operating cash flow covered 0.69x net income, the weakest of the six quarters shown. Check the receivables balance and days outstanding in the Q2 2026 10-Q before concluding this is timing.
  • The debt line. Long-term debt was $3.51bn at 6/30/25 and is not disclosed in any later period supplied. It matters whether that reflects repayment or a data gap.
  • Goodwill. $1.52bn equals about 48% of the $3.16bn equity at 6/30/26, so any impairment lands on a thin cushion.
  • Basis of presentation. The supplied ratio table shows FY2024 net margin of 33.5% where the income statement implies 49.1%, and FY2023 10.9% against 19.6%. The press release reports net income and net income from continuing operations separately, so a continuing-operations distinction likely explains the difference. Confirm which basis the ratios use before reading a trend into them.

Questions worth following

  • Can AppLovin's operating cash flow keep pace with net income as advertising revenue grows more than 50%?

    Cash conversion is the cleanest test of whether 65% net margins are real money. Operating cash flow ran at 1.19 times net income in FY2025 but only 0.87 times in 1H2026 and 0.69 times in Q2 2026, so sustained sub-1x readings would mean growth is outrunning collections.

  • How much of AppLovin's 88% gross margin survives if competing platforms push advertising prices down?

    An 88% gross margin on $1.92bn of quarterly revenue leaves a very wide spread over direct costs, and rival pricing is the fastest way to erode it. The answer decides whether FY2025's 60.8% net margin and $3.94bn of free cash flow are a plateau or a peak.

  • Do AppLovin's repurchases shrink the share count faster than equity awards add to it?

    AppLovin spent $551.3M repurchasing and withholding 1.1 million shares in Q2 2026 and ended the quarter with 335 million shares outstanding. If the count falls, buybacks are returning capital; if it stays flat, the cash is largely financing compensation and per-share gains are smaller.

Generated by DeepVal AI. Data and models may contain errors or be outdated. For research only; not personalized investment advice.

Go beyond the preview

Continue reading the full report

  1. Key risksWhat could break the investment case
  2. Valuation and margin of safetyWhat the current price assumes, and how much room for error it leaves
  3. SummaryOur overall judgement and what to watch next

Free to sign up · 20 research points · unlocking this report uses 10

20 points cover 2 full reports. No card needed.

Research reflects its generation date. Preview the first two modules.

More on Applovin Corp