Tenjin: Mobile Gaming Benchmark 2026... "AppLovin strengthened its lead on iOS in Q2, increasing from 39% to 44% of ad revenue share." Here are the rev-share stats from Q1 to Q2: $U - Unity: - iOS: 13% → 14% - Android: 12% → 12% $APP - AppLovin: - iOS: 39% → 44% - Android: 19% → 23% Mintegral: - iOS: 21% → 16% - Android: 17% → 12% I'm not convinced this narrative that Unity is taking market-share from AppLovin is valid. More market-share = More data = better models. $APP $U

$APP sentiment is quite interesting. People are of the belief that this is a dying company, set to report declining revenues consistently quarter after quarter. This is SO far from the truth. In the recent Edgewater Research report, they make the argument that AppLovin has effectively hit their ceiling. Specifically - and this is important... The mobile gaming MAX supply ceiling. To the uneducated reader, this may come across as dramatic. AppLovin has a 73% share monopoly on ALL top-downloaded mobile game mediation. Essentially, AppLovin is the integrated ad-space auctioneer for this overwhelming majority of mobile games. It comes at NO surprise that they have more-less hit a ceiling here. The report continued to make the argument that AppLovin's QoQ growth has started to plateau, which indicates this ceiling has neared. This is not overly bearish - not at all. I'll explain why in a second. But first, there is this sentiment that competitors are going to release superior models and displace AppLovin's monopoly. I think there is a vast technological misunderstanding here. AppLovin has the superior models achieving superior outcomes for both buyer and seller because they have the lions-share of the market - looking at this through a perhaps naive lens you may assume that one could just come and "take" market share with a superior model or product... But you're missing a critical piece of the puzzle. These game developer companies literally rely on AppLovin for revenue. There is a dependency here. Their models have become so good, that every dollar spend by a game developer on AppLovin's ecosystem, leads to an dipropionate increase of revenue generated. They accomplish this by real-time model refinement and data ingestion through each SDK install within their MAX mediation network. 1 billion DAUs playing thousands of the top-downloaded games produce a lot of 1P data in real-time. It is this very data that gives AppLovin their edge - and it's this very edge that the game-developers rely on to stay profitable. The game developers would have to gamble with solvency, just to jump-ship and divert spend/integrations with a competitor. This would be a horrible business decision. So the compounding outcome is exponential for AppLovin. The developers are locked in because their ability to generate revenue has become an arbitrage of dollars spent on platform. And this relationship only exists if they stay integrated with AppLovin, thus feeding them the data which refines their models that produces the superior outcomes each developer desires. It's a self-fulfilling flywheel. Interestingly - the party with the superior dataset will always produce the superior model. It just so happens that AppLovin has the largest market-share, the top-download apps, and real-time data ingestion for the 1 billion DAU audience it has created. Good luck displacing this juggernaut with an inferior dataset. Now, let's circle back to plateauing revenues QoQ. Yes, with 73% market-share eventually your revenue potential here will start to cap-out. Though AppLovin has started to expand self-serve demand by allowing DTC Shopify brands or consumer brands to also buy ad-space on the network - this has been a modest start. Even though we have recently seen an uptick of pixel installs on consumer sites (this is a strong indicator of the amount/velocity of new advertisers ramping). The true potential unlock for further growth and TAM expansion, is the natural unlock beyond the confined four-walls of mobile-gaming AppLovin finds themselves in. This is called "supply-side expansion". Adam, the CEO has outlined this himself... 1/2