AppLovin Rebuilt Mobile Game Advertising After Apple’s IDFA Changes
AppLovin reported Q1 FY2026 revenue of $1.99 billion — a 36 percent year-over-year increase — with its Software Platform segment, which operates the MAX ad mediation network and the AXON machine learning advertising engine, generating nearly 90 percent of total revenue at operating margins above 75 percent. AppLovin’s Q1 FY2026 investor materials confirmed the company has become the dominant infrastructure layer for mobile game user acquisition, five years after Apple’s App Tracking Transparency changes threatened to make the entire mobile gaming advertising model non-viable. What happened between 2021 and 2026 is not a recovery story so much as a structural replacement: the IDFA-dependent advertising model that powered the 2018-2021 mobile gaming bull cycle was replaced by a fundamentally different attribution and targeting system, and AppLovin built that replacement.
Apple’s ATT framework, introduced in iOS 14.5 in April 2021, required apps to obtain explicit user consent before tracking their identifier across other apps and websites. Consent rates averaged below 30 percent, which meant the deterministic user-level tracking that mobile advertising had relied on was eliminated for roughly 70 percent of the iOS audience. The immediate impact on mobile game publishers was severe: cost-per-install efficiency collapsed across iOS as targeting precision dropped, and publishers who had scaled user acquisition operations around IDFA-dependent measurement could no longer validate which campaigns were producing paying players. The companies most exposed were those running large-scale UA teams with models built on attribution data that simply stopped being available.
What AppLovin’s AXON Engine Actually Does
AXON is AppLovin’s in-house machine learning model for advertising prediction. Rather than targeting individual users based on IDFA identifiers, AXON operates on contextual signals — the properties of the app in which an ad is being shown, the characteristics of the creative, the time of day, device type, geographic location, and aggregate behavioural patterns derived from AppLovin’s network of 1.4 billion daily active users across its portfolio of owned apps and mediated publisher apps. The prediction task AXON is solving is not “this specific user has purchased in-app items in a similar game” (which requires IDFA) but “this context has historically produced users who purchase in-app items in this type of game” — a cohort inference rather than individual tracking. The underlying privacy change Apple imposed is documented in Apple’s App Tracking Transparency framework.
The practical outcome has surprised observers who expected that removing individual-level tracking would make advertising less effective permanently. For publishers using AXON through AppLovin’s network, return on ad spend has recovered to levels that exceed the pre-ATT baseline for the top-performing creative categories. The reason is that AXON’s dataset — derived from AppLovin’s ownership of 200+ mobile games generating direct player behaviour signals — provides training data that no independent ad network can replicate. A network that only mediates third-party publishers has only aggregate signals; AppLovin’s first-party game portfolio generates the granular engagement and monetisation data that makes the cohort inference model more accurate than individual tracking on a noisy dataset. The subscription gaming model addresses a different segment of gaming monetisation; AXON’s dominance in mobile UA addresses the free-to-play sector that subscription services cannot reach.
Who Lost the IDFA Era and Who Won It
The IDFA transition created distinct winners and losers that have now fully resolved in 2026. Unity Technologies, which had built a significant advertising business through Unity Ads and its IronSource acquisition, failed to make the transition effectively. Unity’s advertising revenue declined through 2023 and 2024 as AXON’s performance superiority became apparent to publishers comparing UA efficiency across networks. By 2026, Unity’s core business is the game engine and development tools — the advertising division has been substantially restructured. The competitive consolidation that followed ATT has left AppLovin without a direct peer in mobile game advertising at its performance tier.
The mobile gaming market’s broader consolidation mirrors what happened in advertising: the top publishers who had the LTV models and monetisation depth to sustain higher UA costs have emerged with stronger market positions, while the middle tier has thinned significantly. Sensor Tower’s mid-2026 mobile gaming market analysis shows the top 50 iOS games by revenue accounting for a higher share of total market revenue than at any point before ATT — Sensor Tower’s 2026 mobile gaming market report projects total consumer spending on mobile games at $97 billion globally, with growth concentrated in the top decile of publishers who have rebuilt UA operations around AXON and Google’s Privacy Sandbox attribution alternatives.
The Mobile Gaming Market Structure in 2026
The mobile gaming market in 2026 has a bifurcated structure that ATT accelerated but did not create. High-monetisation genres — 4X strategy, match-3 with live service economies, role-playing games with gacha mechanics, casino/social casino — have LTVs high enough to support UA costs even at reduced targeting efficiency. These genres have consolidated around a small number of globally scaled publishers: Scopely (now part of Savvy Games Group after Saudi Arabia acquisition), King (Activision Blizzard / Microsoft), Zynga (Take-Two), and a handful of Asian publishers with strong live-service operations. Publishers in these categories are the primary buyers of AppLovin’s AXON-powered inventory, and their economics have strengthened as mid-tier competition declined. In crypto-adjacent verticals, the equivalent shift is wallet-based targeting replacing demographic ad models.
The casualty tier — puzzle games without strong live-service economies, hyper-casual games that monetised almost entirely through advertising rather than in-app purchase, mid-core games with insufficient LTV to justify AXON CPMs — has contracted substantially. Hyper-casual as a format has effectively ceased to be economically viable at scale; the CPMs available for hyper-casual ad inventory do not cover the UA cost of acquiring players in a post-IDFA environment where broad targeting is more expensive and less efficient than narrow targeting. AppLovin’s dominance has therefore produced a market where the infrastructure is strong and the beneficiaries are the publishers with the monetisation depth to access it.
The Competitive Structure of Mobile Advertising After IDFA
Apple’s ATT framework, implemented in iOS 14.5 in April 2021, did not simply remove an advertising identifier. It restructured the competitive dynamics of mobile advertising in a way that Michael Porter’s five-forces model describes precisely. The removal of the IDFA raised the barrier to entry for any advertising platform that had been relying on cross-app tracking to build user profiles — a barrier already high due to data-network-effects advantages enjoyed by incumbents. For new entrants to the post-IDFA mobile advertising market, the technical requirement is not just building an ad delivery system. It is building an on-device attribution model capable of predicting conversion probability from contextual signals alone, without persistent cross-app user identifiers. That is a machine-learning problem of sufficient complexity that only companies with access to large proprietary datasets and multi-year engineering investment can compete effectively. AppLovin’s AXON engine is the commercial manifestation of that investment.
The five-forces picture in the post-IDFA landscape has the structure of a narrowing duopoly rather than a competitive market. The threat of new entrants is low: the technical barriers to building a competitive attribution model from scratch are prohibitive for any company without AppLovin’s or Meta’s existing scale, proprietary behavioral signal libraries, and model-training infrastructure. Supplier power — Apple controls the operating system and determines the data access rules — is essentially absolute; there is no negotiating with Apple’s ATT implementation, and every mobile advertising platform operates on Apple’s terms regardless of revenue scale. Buyer power is moderate, because the mobile game developers who purchase user acquisition advertising from AppLovin have a meaningful but limited set of alternatives. They can shift budget to Meta’s advertising ecosystem, reduce overall UA spend, or experiment with emerging platforms — but the performance gap between AppLovin’s AXON model and alternatives is large enough that serious mobile game publishers cannot exit AppLovin entirely without accepting a material reduction in paid user acquisition efficiency.
The primary substitute for AppLovin’s mobile game advertising is Meta’s advertising ecosystem, which survived the IDFA changes with its own first-party data moat intact — Facebook login provides the persistent identity signal that IDFA removal denied to third-party trackers. What the post-IDFA market produced is not fragmentation but consolidation: AppLovin and Meta as the two structurally durable mobile advertising platforms, separated from a tier of smaller players who lacked the proprietary data density to maintain competitive attribution accuracy. This is the market structure Apple’s privacy policy created — one in which the entities with the largest existing behavioral data libraries were structurally advantaged to survive, and the entities most dependent on the IDFA were eliminated. AppLovin’s position is not the result of building better technology in an open market. It is the result of entering the post-IDFA regime with the data depth and model maturity to fill the vacuum that the IDFA’s removal created, and building a revenue engine in the space where smaller competitors used to operate.
