AppLovin’s 59% Revenue Print: Late-Cycle Acceleration or the Top of the Ad Spend Wave?

Revenue up 59% year-over-year. Adjusted EBITDA margin guided to 84-85%. Trailing operating margin at 77.09%. Those are the numbers AppLovin put up for the quarter ending in early May, and the market took the stock from the low 460s to 567.83 in roughly a month, a 23% move on a name already carrying a 190 billion dollar cap. The consensus target average sits at 644, the high end at 860, the low end at 340. That spread, more than the headline beat, tells you where we are.

I read those numbers as the sound of a cycle running hot, not a new floor being built.

Performance advertising, the slice AppLovin lives in through AppDiscovery and the AXON demand-side platform, has its own rhythm distinct from brand budgets. The DSP measures, retargets, attributes, and bills on outcomes, so revenue scales with advertisers’ willingness to bid CPMs higher because the ROAS math still works. When ROAS works, publisher ad-load tolerance expands, click-through holds, and CPMs rise together. We have seen this before: 2013-2015 as mobile install bidding matured, 2020-2021 as e-commerce demand collided with pandemic attention supply. Each time, operators with the best signal advantage posted 50-plus percent revenue prints and 70-plus percent incremental margins for four to six quarters. Each time, the back half of the move came not from new advertisers entering the auction but from existing advertisers bidding harder for the same impressions. That phase produces the cleanest numbers and the most fragile setup.

The supply side matters here. Mobile ad inventory is not meaningfully capacity-constrained; there is no fab to build, no mine to permit. What is constrained is advertiser confidence that a dollar in produces more than a dollar out. The consumer e-commerce vertical that drove March spend up 25% sequentially and pushed April to record levels is the marginal buyer. That cohort is rate-sensitive, inventory-sensitive, and reflexive: when its own end-demand softens, it cuts budgets within a quarter. The 84-85% EBITDA margin guide assumes that cohort keeps bidding. It is a utilization rate disguised as a margin number.

The backdrop is doing some of the work too. Short-term yields at 4.00% would normally crimp high-multiple software, yet the broader software tape held a bid through the print, and platform-level ad spend has not shown the cyclical rollover that hit programmatic peers in late 2022. More interesting is the regulatory machinery now turning. An SEC inquiry into data-collection practices, short-seller allegations around app store compliance, and emerging state-level rules on synthetic ad performers and AI disclosure all hit the same surface: the signal layer the DSP depends on. Identity and signal disruption is a multi-year margin story, not a one-quarter event, and it tends to bite hardest at operators with the highest current take-rate.

So how do I read the 23% one-month move and the 31% three-month move? As the market repricing operational momentum after the May 6 beat and the raised Q2 guide, amplified by price-target hikes: UBS to 750, Macquarie to 730, Piper to 665, Deutsche to 660. Institutional ownership at 75% and short float at 4.10% tell me this is accumulation, not a squeeze. That fits what late-cycle accumulation looks like in this sector. The shorts that were loud in early 2025 have been covered or sidelined; the long-only money is chasing the print. None of that tells me where in the cycle we are. It tells me the cycle is loud right now.

The valuation frame I care about is not the forward multiple on 15.75 dollars of expected EPS, which screens at roughly 36 times and looks digestible against the growth. It is the P/E on cyclical-adjusted EPS, earnings normalized across a full performance-ad cycle, including the down-leg that 2022 produced for every adjacent operator. On that basis, paying mid-30s forward when trailing operating margins are at 77.09% means underwriting that the current margin structure is closer to a new baseline than a cycle peak. Trailing twelve-month free cash flow of 4.43 billion supports the multiple if the run-rate holds. It does not survive a reversion to even 60% operating margins, which would still be extraordinary by any historical comparison.

The counter is structural. For this to not behave like a cycle, AXON’s signal advantage has to compound faster than the regulatory and platform-level identity erosion now in motion. Capacity rationalization on the publisher side, fewer apps and more concentrated inventory, could give a take-rate floor. A genuine technology displacement, where AI-driven bidding becomes the only viable performance channel and brand budgets migrate into it, would extend the runway. Those mechanisms are real and worth watching quarterly. They are also exactly the mechanisms invoked at every cycle peak to explain why the next leg will not look like the last one.

If consumer-vertical ad spend posts a sequential decline for two consecutive quarters while adjusted EBITDA margin compresses below 75%, the late-cycle read is wrong and I am paying for utilization that is already breaking. Until then, I am watching the auction dynamics, not the headline growth rate. The 59% print is what a hot auction looks like from the inside.

Figures reflect the most recent available data and may differ slightly from live market prices. · © Mathstock