At Gamescom in 2023, AppLovin was quietly shopping its game studios. According to mobilegamer.biz reporting on the eventual sale, the asking price at the time was around $2.4 billion.
Two years later, all ten studios went to Tripledot Studios for $800 million total: $400 million in cash plus a 20% stake in the buyer. A source told mobilegamer.biz the portfolio went essentially at cost.
Most coverage read that as a distressed sale. It reads differently once you look at what AppLovin’s ad business was doing in the same quarter, and at what those studios had already given the company.
CEO Adam Foroughi said the quiet part out loud when the deal was announced: “We’ve never been a game developer at heart.”
He was right. That is the story.
An ad company that kept buying games
AppLovin was founded in 2012 in Palo Alto by Adam Foroughi, John Krystynak, and Andrew Karam as a mobile advertising company. Not a publisher. An ad company.
Then in 2018 it did two things in the same year. In July it launched Lion Studios to publish mobile games. In September it acquired MAX, an in-app bidding platform. One purchase put it on the supply side as a game maker. The other put it in the middle, as the mediation layer that decides which network gets to serve each impression.
The buying continued. PeopleFun, Firecraft, and Belka Games in 2019. Geewa, Redemption Games, and Machine Zone in February 2020.
By the time AppLovin filed to go public, the games were not a side project. Of its $1.45 billion in 2020 revenue, 51% came from consumer in-app purchases. On the revenue split alone, AppLovin was more than half a game company when it listed on Nasdaq in April 2021 at $70 a share, valued around $24 billion.
What the studios were actually for
A studio portfolio gives an ad network two things it cannot buy anywhere else.
The first is inventory it controls completely. The second matters more: first-party data on what real players do after they see an ad, install a game, open it, and spend. That is the exact signal a machine learning bidder needs to predict whether an impression is worth $2 or twenty cents.
The timing lines up with the industry’s biggest measurement problem. Apple’s App Tracking Transparency arrived in 2021 and broke the third-party attribution most of mobile ran on. Everyone else lost signal. A company that owned games, owned mediation, and owned an ad network still had a closed loop it could see end to end.
Deconstructor of Fun made the same read in its 2025 piece on the company, framing the publishing arm as “data collection” that fed the bidding models rather than a business in its own right.
Owning the pipe
The mediation side got more aggressive than the studio side ever did.
In October 2021 AppLovin agreed to buy MoPub from Twitter for $1.1 billion, closing in January 2022. MoPub was a major independent mediation platform, and it was wound down. That inventory needed a new home, and MAX was standing there.
In August 2022 AppLovin made a $17.54 billion stock bid for Unity, which Unity’s board rejected in favour of the ironSource merger. That bid was for the other half of the mobile ads pipe.
Where MAX ended up is worth sitting with. Analysis covered by mobilegamer.biz put MAX at 73.2% of the 175 most-downloaded ad-supported games and 55% of the 60 highest-grossing games carrying ads.
If you run a mobile game with ads, there is a strong chance the company deciding your auction is also bidding in it. We broke down what that means for stack choice in our comparison of ironSource vs AppLovin MAX vs AdMob.
The part that should bother monetization teams
Owning both sides of an auction is not illegal and not unusual. Google does it. The question for your revenue is what the overlap actually does to your numbers.
The GameBiz Consulting analysis found AppLovin’s network earning roughly four times more on its own mediation than on competing mediation platforms, while serving around half of all impressions on MAX.
AppLovin’s demand simply performs better when it has full visibility into the auction it is bidding into. The mediator and the bidder sharing a parent company changes the auction. Publicly available data does not settle which it is, and AppLovin has not published anything that would.
What is documented is the pressure around staying inside the system. The same analysis reported publishers being migrated to bidding-only instances from July 2025, features being removed rather than added, and studios saying AppLovin declined to run ROAS campaigns for them unless they were on MAX mediation. That last one ties your user acquisition budget to your monetization stack, which are two decisions that should be made separately.
If you have never pressure-tested the setup, our breakdown of in-app bidding versus waterfall covers the mechanics of what changes when the auction structure moves under you.
The quarter the games stopped being worth it
By early 2025 the two halves of AppLovin were pointing in opposite directions.
The apps segment had declined for three straight years. The studios were roughly break-even. Global mobile game downloads fell 6% the prior year according to Sensor Tower data, so the category was not about to rescue them.
The warning shots came earlier. AppLovin cut 120 staff at Machine Zone in October 2024 after previous attempts to sell the division went nowhere.
The sale closed on 30 June 2025. Ten studios went across: Athena, Belka Games, Clipwire Games, Leyi, Lion Studios, Machine Zone, Magic Tavern, PeopleFun, Zenlife Games, and Zeroo Gravity. Magic Tavern and Zeroo Gravity alone accounted for more than half the group’s revenue.
Tripledot came out of it with 12 studios, 2,500-plus employees, 25 million daily active users, and close to $2 billion in expected annual gross revenue, which puts it in serious company among the largest mobile publishers by revenue.
AppLovin came out of it with the model those studios helped train.
What AppLovin looks like now
The ad business did not slow down after the games left.
In Q2 2026 the company reported $1.92 billion in revenue, up 53% year over year, with adjusted EBITDA of $1.61 billion, up 58%. It was added to the S&P 500 in September 2025. Axon opened up as a self-serve product for e-commerce advertisers in June 2026, taking the same engine outside gaming entirely.
The road there was not smooth. In February 2025, short sellers Fuzzy Panda Research and Culper Research published reports alleging ad fraud and improper data practices, and roughly $13.7 billion in shareholder value came off in the days that followed. AppLovin denied the allegations. Bloomberg reported in October 2025 that the SEC was looking at the company’s data collection practices, which the SEC has not confirmed.
Neither dented the trajectory. Gaming inventory now feeds an advertising machine whose biggest growth story is selling shoes and supplements.
The uncomfortable version for your studio
AppLovin owned games, used the behavioural data from those games to train a bidding model, and sold the games once the model was good enough to run without them. That is a clean, rational sequence.
Every studio on MAX is contributing to the same kind of loop right now. Your install events, your session data, your spend patterns, and your ad response all move through infrastructure you do not own, into models you do not see, that are then sold back to you as performance.
None of that is a scandal. It is the deal. The problem is how many studios have never priced the deal, because the reporting they use to evaluate it comes from the same platform they are evaluating.
We wrote more about this shift in the move from filling ad slots to owning your impressions.
What to actually check this quarter
Five things, in rough order of how fast they pay off:
- Map the overlap. Write down every demand partner in your stack and mark which of them also owns your mediation layer. If one company appears in both columns, note what share of your revenue it touches.
- Get an outside read on your eCPM. Comparing your numbers to a benchmark published by the platform serving your ads tells you very little. Compare against studios in your genre and geo instead.
- Test a real second bidder. Not a token line item at a floor nobody clears. Enough traffic, at competitive floors, for long enough to see what happens to fill and eCPM together.
- Separate your UA decision from your monetization decision. If a partner will only run your acquisition campaigns while you use their mediation, that is a cost. Put a number on it.
- Know your exit. Understand what a mediation migration would take before you need one. Our list of AppLovin MAX alternatives is a starting point, and our roundup of AI-powered ad networks for mobile games covers what else is bidding.
None of this means leaving MAX. Plenty of studios run it and do well. It means knowing what the arrangement costs you, so you are choosing it rather than defaulting into it.
FAQ
Did AppLovin sell its games because they were losing money?
Not exactly. Reporting indicated the studios were roughly break-even. The issue was direction: the apps segment declined 14% year over year in Q1 2025 while advertising grew 71%. A flat business attached to a fast-growing one drags the blended growth rate down, which matters a great deal to a public company.
Who owns Machine Zone and Lion Studios now?
Tripledot Studios. Both were part of the ten-studio package that closed on 30 June 2025. AppLovin retained a 20% stake in Tripledot as part of the consideration, so it still holds indirect exposure to those games.
Does using MAX mean AppLovin gets access to my game’s data?
Any mediation platform processes impression, user, and revenue signals to run the auction. What varies between providers is retention, use in model training, and what the contract permits. Read the data terms on your specific agreement rather than assuming the default, especially if the same company also sells you user acquisition.
Is AppLovin still worth using now that it does not make games?
For a lot of studios, yes. Its demand is strong and its share of top-grossing games with ads sits at 55% for a reason. The argument here is about measurement discipline, not avoidance. Run it, and verify it against something the platform does not control.
What should I do if my only mediation option is the one bidding against me?
Start by proving the gap exists in your own data. Run a controlled test with a competing bidder at real volume and compare fill, eCPM, and total revenue rather than any single metric. If the incumbent still wins on total revenue, you have your answer and you have it from your own numbers.
Why does a company with no games care so much about game inventory?
Games generate high-frequency, high-signal ad impressions, and the behavioural data from them is unusually predictive. AppLovin took that engine into e-commerce with the Axon self-serve launch in June 2026. The inventory stays in gaming; the buyers increasingly do not.
Find out what your stack is actually costing you
If you have never had someone outside your mediation provider look at your auction, you do not know what you are leaving on the table. UndrAds runs independent mediation and demand audits for mobile game studios, covering demand overlap, floor pricing, and fill against genre benchmarks we do not sell you.
