Why Google Play Became the Gatekeeper
The subject here is the dispute between Epic Games and Google over who controls access to Android users and how transactions inside Android apps are processed. The useful question is not simply why Fortnite was removed from Google Play. It is how a conflict that began with one game’s payment route exposed a larger difference between Android’s technical openness and the practical power of Google Play. Reconstructing that sequence lets us judge the later injunction more accurately: Epic’s own breach of Google’s contract can be true at the same time that the jury found Google’s platform practices unlawful.
Fortnite turned a platform rule into a public confrontation
Epic Games released Fortnite in 2018 as a free, cross-platform mobile game. The game did not depend on an upfront download price. Epic earned revenue when players bought special features inside it. For roughly two years, Epic sought to distribute the Android version through direct downloads from its own website and through Samsung’s Galaxy Store. That history matters because it shows that Epic was not unaware of alternatives to Google Play. It tried to use them.
In 2020, however, Epic decided that Google Play was necessary to reach Android users at scale. The Ninth Circuit’s opinion quotes Epic’s conclusion that Google Play was the company’s only realistic hope of reaching users. Epic then offered Fortnite on Play as well as on Apple’s App Store. The choice was commercially understandable and legally consequential: once Fortnite entered Play, the game was subject to Google’s Developer Distribution Agreement and its payment rules.
Soon after the launch, Epic embedded secret code in Fortnite so that players making in-app purchases could bypass the payment-processing systems required by Google and Apple. Epic called the campaign “Project Liberty.” The immediate economic target was the commission that the stores charged on in-app transactions; the broader public argument was that mainstream app stores were restricting developers’ and users’ choices about distribution and billing.
Google and Apple removed Fortnite for violating their terms of service. Epic responded with antitrust suits against both companies. “Antitrust” here means law aimed at protecting competition and limiting conduct that maintains market power through unlawful restraints—not a general rule that every platform contract must be favorable to every developer. That distinction became important because the case had two different layers from the start: what Epic had agreed to do under Google’s contract, and whether Google was using its platform position in a way that harmed competition.
The later trial record made the separation concrete. Epic stipulated that it had violated the Developer Distribution Agreement by putting its own payment solution into Fortnite and owed Google $398,931.23 in fees that Google otherwise would have received. The parties later settled Google’s counterclaims for damages. That resolution did not answer the separate antitrust question. The jury still had to decide whether Google had monopoly power in the relevant Android markets, whether its conduct restrained trade, and whether tying Play distribution to Google Play Billing was unlawful.
The sequence is not a morality play in which one company’s action explains everything that followed. Epic’s payment bypass supplied the immediate trigger and gave Google a contractual counterclaim. Google’s removal supplied the event that brought the dispute into court. The antitrust case then examined the surrounding system: how developers reach users, how users obtain apps, and how payment rules operate when one company controls the most important route between them.
Android was open in design, but alternatives were not equally usable
To understand why Google’s distribution position mattered, it helps to distinguish an operating system from an app store. Android is publicly available and free for others to access, modify, and distribute. Google licenses it to many phone manufacturers, including companies such as Samsung and Motorola. Apps are installed separately from the operating system, and Android phones are not all Google-branded devices. In that sense, Android is structurally different from Apple’s iOS, which is tied to Apple hardware and does not support third-party app stores in the same way.
But an open operating system does not automatically create an open distribution market. A developer can make an app available outside Play only if users can find the download, understand the warnings, complete the steps, and trust the destination. The court’s account of the trial evidence describes Android’s default settings as disabling direct downloads. Users who changed those settings could then face a series of “scare screens,” sometimes as many as 14, warning that the software came from an unknown source and could harm the device. The same warnings appeared whether the destination was a trusted developer website or a plainly malicious one.
The issue was therefore not whether an alternative technically existed. It was whether the operating system made the alternative practically unattractive. The opinion reports that Android users attempting to download Fortnite directly had to navigate more than 15 steps, and that 35% of users who started the process abandoned it after encountering Google’s warning messages. Those figures are historical findings from the trial record, not a current measurement of every Android device or download flow. Their importance is the relationship they illustrate: friction at the user’s end can protect a dominant store even when the platform formally permits another route.
The same difference between formal and practical access applied to rival app stores. Google’s Play Store had a catalog of more than two million apps, according to the opinion. Users were attracted to a large catalog, while developers were attracted to the resulting user base. This is a network effect: each side makes the platform more valuable to the other. A new store could exist, but it would face a difficult initial bargain. Without enough apps, users had little reason to install it; without enough users, developers had little reason to supply it.
That feedback loop also explains why a distribution dispute could not be reduced to a question about commissions. A fee may affect a developer’s economics, but a store’s power can also come from controlling discovery, defaults, trust signals, and the path by which an app reaches a phone. If alternatives are harder to find or use, developers may remain dependent on Play even when they dislike its payment rules. The court’s market analysis treated distribution and billing as distinct markets, but the two were connected by this dependence.
Google’s position was reinforced from both sides of the market
The Ninth Circuit described several practices that, in combination, strengthened Play’s position. Google’s mobile contract effectively required Android manufacturers to preinstall Play on the default home screen. Revenue-sharing arrangements with a “premier tier” of manufacturers could make Play the only preinstalled app store on their phones. This did not mean that every OEM arrangement was unlawful. It meant that placement and default status could affect a rival’s ability to reach users before any developer made a distribution choice.
The opinion also discusses Project Banyan, a proposal involving Samsung’s Galaxy Store. Samsung representatives understood the proposal’s purpose as preventing unnecessary competition with Play. The court’s account further describes Project Hug, a set of special agreements with 22 leading game developers under which those developers received cash and other benefits not to launch on Android stores other than Play. These named projects matter because they show different levers acting on the same system: one concerns the phone’s starting point, another a rival store, and another the supply of popular apps.
By 2020, the opinion reports, Play accounted for 95% of Android app downloads in the United States and more than 80% worldwide excluding China. Those are historical figures presented in the court’s account, not a claim about the market today. They do, however, give scale to the problem the jury was asked to evaluate. A developer could theoretically choose a website or another store while still facing a market in which most users were accustomed to Play.
Google also required developers offering apps on Play to process in-app purchases through Google Play Billing and pay a commission on nearly all such transactions. That rule connected distribution power to payment power. The store was not merely a place where users found software; it was also the required channel for many digital purchases made after the user opened that software. The jury found that Google unlawfully tied use of the Play Store to Google Play Billing, alongside findings about monopoly power and restraint of trade.
The mechanism can be stated without assuming that every Google safety explanation was pretextual or that every Epic proposal was automatically procompetitive. Play’s catalog and user base reinforced one another. Defaults and agreements helped protect the user base and app supply from rivals. Mandatory billing then allowed Google to capture value from transactions occurring inside apps distributed through the protected channel. The court later had to decide which parts of that system could be addressed by an injunction and how far the remedy could reach.
Why the Apple case could not answer the Google case
Epic sued Apple at the same time, but the two cases did not present identical markets. The Apple litigation involved a vertically integrated company whose hardware, operating system, and app store were tied together. Apple’s iPhones did not support third-party app stores, and iOS disabled direct app downloads from the web. Google’s Android system, by contrast, was licensed to other OEMs and had a different relationship with Samsung, Motorola, and other manufacturers. Android also had a history of direct downloads and competing stores, even if the court found that Google’s conduct made those alternatives difficult to use.
The theories of harm differed as well. Claims about Google’s requirements that OEMs place Play on a device’s home screen had no direct equivalent in the Apple case because Apple did not license iOS to outside phone manufacturers. Claims about Google’s deals to keep rival stores off OEM home screens were likewise specific to Android. The Google litigation also included evidence about Google’s alleged efforts to deter direct downloads and agreements with developers not to offer apps on other stores.
The Ninth Circuit therefore rejected Google’s argument that a market finding from Epic’s Apple case prevented Epic from defining Android markets separately. Two companies can compete in a broad market for mobile-game downloads or transactions while competing against different firms in a narrower Android-only market. Market definition depends on the commercial realities and the theory of harm in the case being decided. That is a legal point with practical value: a judgment about one platform’s architecture cannot simply be copied onto another platform whose distribution channels and relationships are different.
The result of this history is a disciplined starting point for the injunction. Fortnite did not prove by itself that every alternative payment or download route was beneficial, safe, or legally protected. Nor did Epic’s contract breach erase the evidence about Google’s control of Android distribution and billing. What the case made visible was a set of reinforcing dependencies: users were accustomed to Play, developers needed those users, rival stores lacked both catalog and reach, and billing rules linked access to monetization.
The next question is how the court translated those findings into relief. The answer is not a single order to “be more open.” It is a combination of prohibitions and affirmative access measures aimed at different points in the system.