[APP] AppLovin: Q3 2026 Earnings Test for Its New Axon Model

AppLovin is a mobile advertising platform built around an AI recommendation engine, and it earns substantially all of its revenue from its demand-side platform, AppLovin Ads[1]. AppLovin Q3 2026 earnings will cover the Third quarter 2026 (three months ended September 30, 2026), and the results are scheduled for 2026-11-04, according to the Drillr earnings calendar[2]. In the most recent disclosed quarter, the second quarter of 2026, revenue was $1.92 billion, up 53% year over year, diluted EPS was $3.76, and adjusted EBITDA was $1.61 billion at an 84% margin[3][4], but revenue landed below the midpoint of the $1.915 billion to $1.945 billion guidance[5][6]. On August 5, management guided third-quarter revenue to $2.055 billion to $2.085 billion and adjusted EBITDA to $1.71 billion to $1.74 billion (a margin of about 83%), and said full-year 2026 free cash flow should be about 75% of adjusted EBITDA[7][6]; the analyst consensus compiled by Drillr stands at $2.077 billion of third-quarter revenue and EPS of $4.04, with the revenue consensus slightly above the $2.07 billion guidance midpoint[8].

Three things matter most in this report. First, whether revenue returns to the upper half of guidance: in the second quarter, net revenue per installation rose 58% while installation volume fell 2%, and price growth had already slowed from 93% in the first quarter[9][10]; management blamed the shortfall against the midpoint on a model improvement that launched just after quarter end[11], and third-quarter numbers will test that explanation directly. Second, whether the compute cost of the new model erodes margins: second-quarter cost of revenue rose $70.7 million year over year, about 10.6% of incremental revenue and above roughly 7.6% in the first quarter[12][13], so holding a margin of about 83% decides the quality of the growth. Third, whether cash conversion recovers: second-quarter free cash flow of $863.3 million equaled only about 53% of adjusted EBITDA, and first-half cash taxes of $639.8 million were more than six times the prior-year level[4][14], which determines whether buybacks can climb back from $551.3 million in the second quarter.

Company Background and Business Structure

AppLovin has moved from a hybrid of game studios and an ad platform to a pure advertising-technology company. On May 7, 2025, it agreed to sell its Apps business to Tripledot, closed the sale on June 30, and has since reported the Apps business as discontinued operations[15]; the deal brought about $424.7 million of cash proceeds net of cash divested plus $285 million of non-cash consideration[14], and equity method investments stood at $290 million on the balance sheet at the end of June 2026[16]. Tripledot still uses AppLovin Ads as a related party, and the company recognized $18.0 million and $42.9 million of related revenue in the second quarter and first half of 2026[17].

After the sale, the company operates a single reportable segment that offers four products: AppLovin Ads (formerly Axon Ads Manager), MAX, Adjust and Wurl[18]. AppLovin Ads is the demand-side platform: advertisers set return goals for their campaigns, the Axon model matches users to those goals in auctions and charges dynamically against them, and this business supplies substantially all revenue[19]. MAX is an in-app bidding mediation tool that helps publishers monetize ad inventory while also supplying inventory and data to the platform; Adjust provides attribution analytics and Wurl provides connected-TV content distribution and advertising, and the company says revenue from other services is not material[20]. Advertisers are mainly mobile game developers, and since 2024 the company has begun adding web-based e-commerce advertisers[21].

The company reports revenue on a net basis, and its main inputs are compute and engineering staff. Because it does not control the ad inventory, AppLovin acts as an agent and recognizes advertiser spend net of what it pays publishers; impression-based campaigns are recognized when the impression is delivered and action-based campaigns when the click or install occurs[20]. Advertisers generally pay 30 days after month end, and substantially all contracts can be canceled at any time[1]. In 2025, datacenter costs of $542.7 million, personnel expenses of $207.3 million and stock-based compensation of $208.0 million were the main operating expense lines in the continuing-operations income statement[18].

Financial History and Current Position

On a continuing-operations basis, AppLovin's revenue nearly tripled in two years. Revenue rose from $1.84 billion in 2023 to $3.22 billion in 2024 and $5.48 billion in 2025[18]; the 75% growth in 2024 came mainly from a 50% rise in installation volume, while the 70% growth in 2025 came mainly from a 72% rise in net revenue per installation, with installations up only 3%[22]. Net income from continuing operations was $3.43 billion in 2025, up from $1.59 billion in 2024[18]; full-year operating cash flow was $3.97 billion, free cash flow was $3.95 billion and net cash used in financing activities was $2.59 billion[23].

Revenue kept growing quickly in the first two quarters of 2026, but both the growth rate and delivery against guidance declined. First-quarter revenue was $1.84 billion, up 59%, net income from continuing operations was $1.21 billion and adjusted EBITDA was $1.56 billion at an 85% margin[24], with free cash flow of $1.29 billion[25]. Second-quarter revenue was $1.92 billion, up 53% year over year and 4% sequentially[3][11], with operating income of $1.49 billion and an operating margin of 78%[13]; net income was $1.27 billion, diluted EPS was $3.76, and adjusted EBITDA was $1.61 billion at an 84% margin, just below the $1.615 billion low end of guidance[3][4][5]. Free cash flow in the same quarter was $863.3 million, about one-third less than in the first quarter[4].

The balance sheet carries little leverage, and shareholder returns come mainly through buybacks. As of June 30, 2026, the company held $3.05 billion of cash and equivalents, $3.6 billion in principal of senior unsecured notes and $3.52 billion of long-term debt on the books, with its $1.0 billion revolving credit facility undrawn[16][26]; management put net leverage at about 0.1 times trailing 12-month adjusted EBITDA, well below its long-term target of about 1 times[6]. The company repurchased and withheld 2.2 million shares for $1.0 billion in the first quarter and 1.1 million shares for $551.3 million in the second, ending June with 335 million shares outstanding and about $1.8 billion left on its repurchase authorization[24][4][6].

Operating Model

Revenue can be split into installation volume times net revenue per installation, and nearly all growth over the past two years came from the second factor. The company uses exactly these two drivers to explain revenue in its 10-K and 10-Q filings: in 2025, installations rose 3% and price per installation rose 72%, lifting revenue 70%[22]; in the second quarter of 2026, installations fell 2% and price rose 58%, lifting revenue 53%[9]. Price depends on how well the Axon model matches advertisers with high-value users, and when model improvements stall, slower price growth pushes revenue straight below guidance[27]; volume depends on advertiser budgets and inventory coverage. The company discloses only year-over-year percentages for these two drivers, not absolute values, and it does not split revenue between gaming and consumer advertisers.

On the profit side, compute cost is the key variable, while every other expense line is small relative to revenue. In the second quarter of 2026, cost of revenue was $225.8 million, about 12% of revenue; sales and marketing was $63.4 million, research and development $99.9 million and general and administrative $40.3 million, or roughly 3%, 5% and 2% of revenue, leaving an operating margin of 78%[13]. Adjusted EBITDA adds back stock-based compensation ($85.8 million in the second quarter) and depreciation and amortization, among other items, to operating income[3]. Management's framework is a long-term adjusted EBITDA margin in the low 80% range and about $0.10 of incremental compute for each incremental dollar of revenue, and the third-quarter margin guidance of about 83% already includes the higher training and inference costs of the new model[7]; separately, interest expense runs about $51 million a quarter and the second-quarter income tax provision was $239.0 million[13].

Cash flow starts from adjusted EBITDA but must absorb cash taxes, interest and working capital. Capital spending is minimal: purchases of property and equipment were only $473,000 in 2025, so free cash flow is almost equal to operating cash flow[23]. Because advertisers pay 30 days after month end, receivables grow with revenue and reached $2.17 billion at the end of June, up from $1.82 billion at the end of 2025[1][16]. First-half 2026 free cash flow totaled $2.15 billion, about 68% of the $3.17 billion of adjusted EBITDA over the same period, with roughly 83% in the first quarter and 53% in the second[25][4][3], and the cash goes mainly to buybacks.

Industry and Competitive Position

AppLovin operates in a highly fragmented mobile advertising ecosystem in which competitors are often also partners and clients. In its 10-K and 10-Q, the company names Meta, Google, Amazon, Unity Software and Liftoff Mobile as competitors, notes that advertisers typically use several platforms and shift budgets among them based on returns, and warns that clients who are also competitors may invest in their own offerings instead[28][29]. Its advantage is concentrated in mobile game user acquisition: its own MAX mediation platform brings a large pool of in-app inventory and data, and the Axon model bids against advertisers' return goals; management said MAX publisher earnings grew by double digits sequentially in the second quarter and AppLovin's share of publisher waterfalls stayed consistent[11].

Outside mobile games, AppLovin is a latecomer and depends on rules set by a few platforms. Web-based e-commerce and other consumer advertising are markets where Meta, Google and others already hold established positions, and the 10-K acknowledges that this expansion is still early[21], while management concedes that most advertisers associate the company only with gaming[27]. Mobile app distribution and data depend on platforms such as the Apple App Store, Google Play and Meta, some of which are direct competitors and can unilaterally change privacy rules, fees or how apps advertise[30]. The available material provides only a list of competitors and the company's own account, with no comparable share or pricing data, so AppLovin's relative position in e-commerce advertising cannot yet be quantified.

Core Debates

After a second quarter that landed below the midpoint of guidance, can AppLovin's newly launched model lift third-quarter revenue into the upper half of its guided range?

This debate decides whether the second-quarter shortfall was a one-time timing issue or the start of slowing momentum. AppLovin earns substantially all of its revenue from AppLovin Ads, and growth comes mainly from net revenue per installation: in 2025, installations rose only 3% while price rose 72%[22][19]. Second-quarter revenue of $1.92 billion fell below the midpoint of the $1.915 billion to $1.945 billion guidance, and adjusted EBITDA of $1.61 billion was slightly below the $1.615 billion low end[3][5]. Third-quarter guidance of $2.055 billion to $2.085 billion implies sequential growth of 7% to 8%, roughly double the second quarter's 4% sequential growth, which requires the new model to deliver a real lift in price[7][11].

Current evidence shows price is still growing fast, but the pace is converging. Second-quarter revenue rose 53% year over year, with net revenue per installation up 58% and installations down 2%; first-quarter revenue rose 59%, with price up 93% and installations down 18%[9][10]. Management described the second-quarter miss as purely a matter of model-improvement timing, said the next large improvement launched just after quarter end, said demand and the competitive landscape had not changed, and said third-quarter guidance already incorporates that improvement[6][7]. Management also said MAX publisher earnings grew by double digits sequentially and AppLovin's share of publisher waterfalls stayed consistent[11].

The financial transmission is direct, but the same numbers support another reading. When the Axon recommendation model improves, the same installation can be matched with a higher-value user, advertisers bid more at the same return goal, net revenue per installation rises and revenue grows even with installations roughly flat; when model improvements stall, slower price growth pushes revenue straight below guidance[9][27]. The alternative reading is that the slowdown in price growth from 93% in the first quarter to 58% shows the model-driven price lift is converging and gaming advertisers' budget share is near saturation, so whether the third quarter meets guidance depends on a one-time model launch rather than sustained share gains. Because the company does not disclose absolute installation volume or price, the available data cannot separate model effects from shifts in customer mix.

Three checkpoints in the third-quarter report can test this debate. The first is whether revenue is at or above the $2.07 billion guidance midpoint and whether management confirms that the new model's effect is fully reflected in the quarter; the second is whether the 10-Q shows net revenue per installation still growing above 40% year over year with any decline in installations held within 10%; the third is the sequential growth implied by fourth-quarter guidance and whether that guidance again includes only model improvements already deployed[7]. If third-quarter revenue falls below $2.055 billion and price growth drops below 40%, the second-quarter shortfall was not a one-time timing issue; if revenue reaches the upper half of guidance and installations turn positive, outside claims that share gains have stalled would be weakened.

With self-serve just opened to the public and creative tools still incomplete, can AppLovin's spending from e-commerce and other consumer advertisers set another record in the third quarter?

Consumer advertisers are the only new customer group outside gaming that has already scaled, and they are the main support for the company's long-term growth goal. Management treats the consumer vertical as a long-term revenue source and has set a long-term target of 30% compound annual growth for the whole business[6]. The second quarter is seasonally slow, yet consumer advertiser spend still came in 28% above the seasonal peak of the fourth quarter of 2025[11]; the company does not disclose revenue or customer counts for this business, however, so outsiders cannot verify its size and must infer it from management's relative descriptions and marketing expense.

Current evidence on the consumer business centers on product launches and customer-acquisition spending. In the first quarter, management said March spend grew roughly 25% more than January and April set a monthly record[5]; in the second quarter, the self-service AppLovin Ads Manager launched publicly, prioritizing mid-market advertisers, with plans to open it to long-tail advertisers as data scale and product capabilities improve[6]. The customer base is currently concentrated in a smaller number of larger mid-market advertisers, and the company draws them in through e-commerce data partners such as Triple Whale to reach higher-spending clients[31]. Second-quarter sales and marketing expense was $63.4 million, up 35% year over year, driven mainly by a $17.2 million increase in advertising and marketing program costs; first-half spending on this line was $124.1 million, up 17%[32][13].

Consumer growth must pass two gates, customer acquisition and creative capability, before it becomes net revenue. Referral programs, partner channels and the self-service Ads Manager lower the barrier for mid-market e-commerce advertisers to join, and as new advertisers test and scale, consumer spend net of publisher payouts becomes incremental net revenue, while acquisition marketing is booked first in sales and marketing expense[32][1]. Management acknowledges that its most effective ad units require a 30-60 second video plus an interactive end card and that the platform cannot yet consistently generate high-quality long-form video automatically, which limits onboarding conversion for long-tail advertisers[31]. The alternative reading is that consumer spend growth is concentrated in a small number of mid-market clients brought in through partners, so the second-quarter record could reflect concentrated test budgets rather than sustainable advertiser expansion; until the company discloses customer counts or revenue, the two cases cannot be told apart.

The third quarter offers three ways to track the consumer expansion. The first is whether management reports another record in consumer spend and whether it gives advertiser counts, retention or revenue for the first time; the second is whether sales and marketing expense grows faster than revenue year over year; the third is progress in opening self-serve to long-tail advertisers and in automated video creative[31]. If third-quarter consumer spend falls while marketing expense keeps growing fast, acquisition efficiency is falling short; if spend sets consecutive records and the company begins disclosing advertiser counts, the view that consumer growth is spreading beyond a few large clients would gain support.

As its new model demands more compute, can AppLovin hold its adjusted EBITDA margin at around 83%?

Compute is the largest variable in AppLovin's cost base, and it decides whether high growth can keep converting at margins above 80%. Datacenter costs of $542.7 million in 2025 were the largest single operating expense in the continuing-operations income statement[18]; in the second quarter of 2026, the operating margin was 78% and the adjusted EBITDA margin was 84%[13][3]. Management's third-quarter margin guidance of about 83% is below the 84% to 85% it guided for the second quarter, and it says it approves compute investment only when it is tied to clear incremental revenue[7][5][31]; if compute grows faster than revenue, margins will be hit before revenue is.

Second-quarter cost data show the incremental cost ratio rising. Cost of revenue was $225.8 million, up $70.7 million (46%) year over year, including a $52.1 million increase in network infrastructure expense[12][13]. Dividing the increase in cost of revenue by the $664.9 million increase in revenue gives an incremental cost ratio of about 10.6%, above roughly 7.6% in the first quarter and 6.4% for full-year 2025[9][10][33]. Research and development expense rose $55.9 million (127%), mainly from stock-based compensation-related personnel costs[34], while the stock-based compensation added back in the adjusted EBITDA reconciliation rose from $34.6 million to $85.8 million over the same period[3]; third-quarter adjusted EBITDA guidance of $1.71 billion to $1.74 billion already includes the higher training and inference costs of the new model[7].

A time gap separates compute investment from revenue payback, and that gap is the core of the margin debate. More complex recommendation models need more training and inference compute, which raises network infrastructure expense; if the revenue lift lags the compute build, the incremental cost ratio rises and the adjusted EBITDA margin falls, whereas once the revenue lift arrives the incremental cost ratio falls back and margins hold in the low 80% range[27][35]. The alternative reading is that the higher incremental cost ratio reflects concentrated training spend ahead of the model launch and will fall back below 10% once the revenue lift is realized, making the margin dip a short-term timing gap. Quarterly reports do not break out datacenter costs, so using the change in cost of revenue as a proxy for compute also captures other cost movements.

Three numbers in the third quarter can test whether the margin holds. The first is whether adjusted EBITDA lands between $1.71 billion and $1.74 billion with a margin of at least 82%; the second is the year-over-year increase in cost of revenue and the change in network infrastructure expense in the 10-Q, and whether the incremental cost ratio exceeds 15%; the third is whether management revises its long-term framework of margins in the low 80% range or about $0.10 of compute per dollar of revenue[7]. If the incremental cost ratio exceeds 15% while revenue does not beat guidance, compute spending is running ahead of returns; if the margin returns above 84% and revenue meets guidance, compute cost pressure was overstated.

With second-quarter cash conversion at only about 53%, are AppLovin's higher cash taxes a timing issue or the new normal?

Cash conversion decides how much of AppLovin's profit turns into buybacks. The company has almost no capital spending: free cash flow was $3.95 billion in 2025, and net cash used in financing activities was $2.59 billion[23]. Second-quarter 2026 free cash flow fell to $863.3 million, and repurchases and withholding fell from $1.0 billion in the first quarter to $551.3 million[4][24]. If cash taxes become the norm, the gap between adjusted EBITDA and usable cash will persist.

The first-half cash flow statement shows cash taxes as the biggest reason conversion fell. Operating cash flow was $2.16 billion in the first half of 2026, and cash paid for income taxes was $639.8 million, more than six times the $100.6 million a year earlier; cash interest of $99.7 million was essentially flat; receivables rose $352.6 million and accrued and other liabilities fell $125.0 million[14]. By quarter, free cash flow equaled about 53% of adjusted EBITDA in the second quarter, about 83% in the first quarter and about 68% for the half[4][3][25][24]. Management attributed the weak second-quarter conversion to the timing of international cash tax and interest payments, said conversion would improve in the third quarter and reach about 75% of adjusted EBITDA for the year, and said the slower buyback pace reflected only temporarily lower free cash flow[6].

Between adjusted EBITDA and buybacks, cash taxes, cash interest and receivables growth are deducted in turn. If any link in that chain stays high for long, cash available for buybacks will stay below the level adjusted EBITDA suggests[14]. The alternative reading is that, after a large jump in profit, historical loss carryforwards and deferred tax items are being used up, and with the income tax provision rising from $112.1 million in the second quarter of 2025 to $239.0 million, higher cash taxes are a consequence of profit scale rather than just payment timing[13]. Because the company discloses only year-to-date cumulative cash taxes, the third-quarter figure must be derived by subtracting the first-half amount from the nine-month total.

Three numbers in the third quarter can show whether this is a timing issue or the new normal. The first is whether third-quarter free cash flow returns to at least 75% of adjusted EBITDA; the second is whether nine-month cash taxes in the 10-Q, minus the $639.8 million paid in the first half, show a clear drop in the third-quarter amount; the third is whether third-quarter buybacks return above $750 million and whether receivables grow faster than revenue[14]. If third-quarter conversion stays below 68%, the full-year figure of about 75% will be hard to reach; if conversion returns above 80% and buybacks recover to near $1.0 billion, the second quarter was only a timing issue.

Risks and Falsifiers

Changes in platform rules are the risk AppLovin can least control on its own. The mobile app ecosystem relies on a small number of platforms such as the Apple App Store, Google Play and Meta, some of which are direct competitors and can change privacy rules, fees or advertising methods; the company relies on data such as IDFA to train its recommendation models, and reduced data availability would weaken ad performance and directly lower net revenue per installation[30]. Because substantially all revenue comes from AppLovin Ads, the company has no other revenue source to cushion such a shock[19]. If Apple or Google announce no new rules affecting mobile ad attribution or data sharing during the reporting period, and the company does not cite platform changes as a factor in quarterly results, this risk did not materialize in the quarter.

Legal and regulatory proceedings could bring costs and could force changes to products. In its 10-Q, the company warns that litigation, government investigations and settlements could result in substantial costs, civil and criminal liability, or requirements to change business practices, products or technologies[36]; the previously disclosed SEC inquiry has concluded with no recommended action[6]. Legal costs are booked in general and administrative expense, and a settlement or judgment requiring changes to ad targeting practices could affect product features and revenue. If the third-quarter 10-Q adds no material legal contingency and discloses no product change made for regulatory reasons, the risk's effect on the quarter is limited.

An uneven cadence of model improvements can push quarterly revenue directly away from guidance. The company acknowledges that improvements in testing do not guarantee uplift and that delays in material improvements can cause quarterly results to miss guidance, which already happened once in the second quarter[27]. The third-quarter guidance range is $30 million wide; each percentage point of revenue below the $2.07 billion midpoint is about $20.7 million, which at an adjusted EBITDA margin of about 83% equals roughly $17 million of adjusted EBITDA[7]. If revenue in both the third and fourth quarters is at or above each quarter's guidance midpoint, the model-cadence risk did not materialize.

The consumer expansion is limited by both brand awareness and product capability. Most advertisers associate the company only with gaming, automated creative generation is still lacking, and management says consumer growth will compound over years rather than months[27]. The consumer business is not yet large enough to offset quarterly swings in gaming[11], and acquisition marketing is booked first in sales and marketing expense, where advertising and marketing program costs already rose $17.2 million year over year in the second quarter[32]. If consumer spend sets records in both the third and fourth quarters while marketing expense stays at or below 3.5% of revenue, this risk is falsified.

Compute investment may come before revenue, creating short-term margin pressure. Management acknowledges that more complex models need more training and inference compute, which can weigh on margins before the related revenue lift arrives[27]. At the third-quarter revenue guidance midpoint of about $2.07 billion, each percentage point of adjusted EBITDA margin equals about $21 million of adjusted EBITDA, and second-quarter cost of revenue already rose $70.7 million year over year[7][12]. If the adjusted EBITDA margin is at least 82% in both the third and fourth quarters and the incremental cost ratio is no higher than 12%, this risk is falsified.

Cash taxes and working capital may stay elevated and squeeze buyback capacity. First-half cash taxes were more than six times the prior-year level, and receivables keep rising with revenue[14]. Second-quarter free cash flow was about $750 million below adjusted EBITDA; each 10 percentage points of lower conversion, applied to the roughly $1.725 billion third-quarter adjusted EBITDA guidance midpoint, means about $170 million less cash in the quarter and correspondingly less money for buybacks[4][3][7]. If conversion is at least 75% in both the third and fourth quarters, this risk is falsified.

What to Watch Next

The four core debates reduce to a short list of metrics that can be checked directly in the third-quarter report.

• Model-driven revenue recovery, revenue against guidance: second-quarter revenue was $1.92 billion, up 53%, and the third-quarter guidance midpoint is $2.07 billion[3][7]. Watch whether revenue reaches the midpoint and whether management confirms the new model's effect is fully reflected. Revenue below $2.055 billion with price growth under 40% would weaken the timing explanation; the upper half of guidance with positive installations would strengthen it.
• Model-driven revenue recovery, price and volume: second-quarter net revenue per installation rose 58% and installations fell 2%, versus 93% and minus 18% in the first quarter[9][10]. Watch whether price growth stays above 40% and any volume decline stays within 10%. Price growth below 40% would signal a converging lift.
• Consumer records, spend and marketing expense: second-quarter consumer spend was 28% above the fourth-quarter 2025 peak, and sales and marketing expense was $63.4 million, up 35%[11][32]. Watch for another record, first-time disclosure of advertiser counts or revenue, and whether marketing grows faster than revenue. Falling spend with fast-rising marketing would weaken the case; consecutive records with disclosed advertiser counts would strengthen it.
• Holding an 83% margin, margin and incremental cost ratio: the second-quarter margin was 84% with an incremental cost ratio of about 10.6%, and third-quarter guidance is about 83%[3][12][7]. Watch whether the margin stays at or above 82% and whether the incremental cost ratio exceeds 15%. A ratio above 15% without a revenue beat would weaken the case; a margin back above 84% would strengthen it.
• Cash taxes as the new normal, conversion and quarterly cash taxes: second-quarter conversion was about 53%, first-half cash taxes were $639.8 million and buybacks were $551.3 million[4][14]. Watch whether conversion returns above 75%, whether third-quarter cash taxes fall and whether buybacks return above $750 million. Conversion below 68% would weaken the timing explanation; above 80% with buybacks near $1.0 billion would strengthen it.

Conclusion

AppLovin's business is dominated by one variable: whether the Axon model can keep raising net revenue per installation. In 2025, a 72% rise in price drove 70% revenue growth to $5.48 billion[22][18]; in the second quarter of 2026, price growth slowed to 58%, and revenue of $1.92 billion, up 53%, fell below the guidance midpoint[9][3]. The financial position remains strong, with an 84% adjusted EBITDA margin in the second quarter[3], $3.05 billion of cash at the end of June and net leverage of about 0.1 times[16][6]; but second-quarter free cash flow equaled only about 53% of adjusted EBITDA, and the incremental cost ratio rose to about 10.6%[4][12]. The central unresolved relationship is whether the price lift from the new model can cover both higher compute costs and higher cash taxes at the same time.

Outside views split sharply after the second-quarter results. Benzinga reported on September 23 that Edgewater Research analyst Joe Wittine believes AppLovin's market share expansion in mobile advertising has effectively stalled, calling it a "show-me story"[37]; if that view is right, the second-quarter shortfall was more than a model-timing issue, and installation and price growth would keep slowing in the third quarter. Proactive Investors reported on October 14 that Wedbush Securities' channel checks indicate AppLovin is not losing market share, that advertisers held spending steady in the third quarter, and that advertisers anticipate an e-commerce-led lift in the fourth quarter[38]; that view is consistent with management's timing explanation and pins fourth-quarter upside on consumer advertisers. Both are outside interpretations rather than facts or a vote, and only third-quarter price, volume and consumer-spend data can show which is closer to reality.

The combination that would materially strengthen the current understanding is third-quarter revenue at or above the $2.07 billion guidance midpoint, price growth above 40% with a narrower decline in installations, an adjusted EBITDA margin of at least 82%, free cash flow back above 75% of adjusted EBITDA, and management reporting another record in consumer spend[7][6]. Conversely, revenue below $2.055 billion, price growth under 40%, an incremental cost ratio above 15% and conversion still below 68% would show model returns converging while costs and taxes rise, materially weakening the current understanding.

Sources

[1] APP 10-Q filed 2026-08-05 · revenue recognition AppLovin Ads · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[2] APP earnings calendar updated 2026-10-25 · 2026-10-25 · Drillr earnings calendar

[3] APP 8-K filed 2026-08-05 · Q2 2026 results and adjusted EBITDA · 2026-08-05 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=8-K&dateb=&owner=include&count=40

[4] APP 8-K filed 2026-08-05 · Q2 2026 cash flow and buybacks · 2026-08-05 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=8-K&dateb=&owner=include&count=40

[5] APP Q1 2026 earnings call 2026-05-06 · consumer vertical and Q2 guidance · 2026-05-06 · earnings-call · https://investors.applovin.com/

[6] APP Q2 2026 earnings call 2026-08-05 · management highlights · 2026-08-05 · earnings-call · https://investors.applovin.com/

[7] APP Q2 2026 earnings call 2026-08-05 · Q3 guidance and margin framework · 2026-08-05 · earnings-call · https://investors.applovin.com/

[8] Drillr analyst estimates APP Q3 2026 retrieved 2026-10-25 · 2026-10-25 · Drillr analyst estimates · https://gateway.drillr.ai/mcp/private

[9] APP 10-Q filed 2026-08-05 · Q2 2026 revenue drivers · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[10] APP 10-Q filed 2026-05-06 · Q1 2026 revenue drivers · 2026-05-06 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[11] APP Q2 2026 earnings call 2026-08-05 · gaming and consumer · 2026-08-05 · earnings-call · https://investors.applovin.com/

[12] APP 10-Q filed 2026-08-05 · Q2 2026 cost of revenue · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[13] APP 10-Q filed 2026-08-05 · Q2 2026 results of operations · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[14] APP 10-Q filed 2026-08-05 · H1 2026 cash flows · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[15] APP 10-K filed 2026-02-19 · Apps business divestiture · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[16] APP 10-Q filed 2026-08-05 · balance sheet June 30 2026 · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[17] APP 10-Q filed 2026-08-05 · Tripledot related-party revenue · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[18] APP 10-K filed 2026-02-19 · single segment net income 2025 · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[19] APP 10-K filed 2026-02-19 · Axon Ads Manager · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[20] APP 10-K filed 2026-02-19 · revenue recognition · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[21] APP 10-K filed 2026-02-19 · new verticals and e-commerce · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[22] APP 10-K filed 2026-02-19 · revenue drivers 2025 and 2024 · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[23] APP 10-K filed 2026-02-19 · free cash flow 2023-2025 · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[24] APP 8-K filed 2026-05-06 · Q1 2026 results · 2026-05-06 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=8-K&dateb=&owner=include&count=40

[25] APP 8-K filed 2026-05-06 · Q1 2026 free cash flow · 2026-05-06 · 8-K · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=8-K&dateb=&owner=include&count=40

[26] APP 10-Q filed 2026-08-05 · liquidity and debt · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[27] APP Q2 2026 earnings call 2026-08-05 · risks · 2026-08-05 · earnings-call · https://investors.applovin.com/

[28] APP 10-K filed 2026-02-19 · competition · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[29] APP 10-Q filed 2026-08-05 · competition · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[30] APP 10-K filed 2026-02-19 · platform policy risk · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[31] APP Q2 2026 earnings call 2026-08-05 · Q&A self-serve and creative · 2026-08-05 · earnings-call · https://investors.applovin.com/

[32] APP 10-Q filed 2026-08-05 · Q2 2026 sales and marketing · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[33] APP 10-K filed 2026-02-19 · cost of revenue 2025 · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[34] APP 10-Q filed 2026-08-05 · Q2 2026 research and development · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[35] APP 10-K filed 2026-02-19 · investment in Axon AI and datacenter · 2026-02-19 · 10-K · https://www.sec.gov/Archives/edgar/data/1751008/000175100826000010/

[36] APP 10-Q filed 2026-08-05 · legal and regulatory proceedings · 2026-08-05 · 10-Q · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001751008&type=10-Q&dateb=&owner=include&count=40

[37] Benzinga 2026-09-23 · Edgewater Research on AppLovin share gains · 2026-09-23 · Benzinga · https://www.benzinga.com/trading-ideas/movers/26/09/61943731/applovin-stock-drops-after-analyst-calls-it-a-show-me-story

[38] Proactive Investors 2026-10-14 · Wedbush channel checks on AppLovin · 2026-10-14 · Proactive Investors · https://www.proactiveinvestors.com/companies/news/1100121