ICLE Issue Brief

Apple in Brazil’s Antitrust Orchard: When Ecosystem Theory Bears Strange Fruit

Executive Summary

This issue brief examines the Brazilian Administrative Council for Economic Defense’s (CADE) decision to uphold interim measures against Apple’s App Store rules for app distribution and in-app payments on iOS. Although the case settled in December 2025, the May 2025 decision remains CADE’s clearest statement on competition in digital ecosystems.

The brief argues that the Decision rests on fragile foundations. CADE defined the relevant market as “the non-licensable smart mobile operating system iOS,” effectively making Apple dominant within its own platform by construction. That single-brand market definition sidesteps the core antitrust inquiry: whether consumers and developers have meaningful substitutes. They do. Apple competes with Android-based devices across hardware, software, privacy, security, services, artificial intelligence, and overall user experience.

The brief also argues that “ecosystem theory” adds little beyond established doctrines of tying, leveraging, foreclosure, and discrimination. As applied here, it risks treating ordinary platform governance and vertical integration as evidence of dominance and harm, without sufficient proof of market power or consumer injury. The Decision also falls short on evidence. CADE relied heavily on foreign precedent and academic literature, but did little to test its theory against Brazilian market conditions. It also gave insufficient weight to Apple’s procompetitive justifications, including privacy, security, fraud prevention, curation, quality control, and reduced transaction costs.

The remedies underscore the concern. The interim measures and settlement required Apple to open parts of its distribution and payment architecture, while specifying fee structures and implementation terms. That resembles sector-specific regulation more than ordinary antitrust relief.

At the same time, the case undercuts calls for new ex ante digital-market regulation in Brazil. CADE’s existing tools proved powerful enough to address the alleged harms. The Apple case is therefore less a vindication of ecosystem theory than a warning about its limits: antitrust should protect competition and consumers, not redistribute platform rents.

I.                    Introduction

This brief critiques the decision by Brazil’s Administrative Council for Economic Defense (CADE) to uphold interim measures against Apple’s rules for app distribution and in-app payments on iOS. CADE found that Apple’s App Store rules—especially mandatory use of Apple’s in-app purchase (IAP) system, “anti-steering” provisions, and limits on alternative app stores—likely amounted to an abuse of dominance in a market the Tribunal defined as “the non-licensable smart mobile operating system iOS.” The interim measures required Apple to permit third-party app stores, sideloading, alternative in-app payment systems, and external links to other purchasing options. Although the case ended in a December 2025 settlement, the Tribunal’s May 2025 decision remains CADE’s clearest statement on competition in digital ecosystems. Its reasoning may shape future Brazilian enforcement and serve as a reference abroad.

The Tribunal deserves credit for engaging seriously with the literature and case law on digital ecosystems. But the Decision rests on fragile analytical foundations. Its single-brand market definition effectively makes Apple dominant within its own platform by construction. And its ecosystem theories of harm—despite their descriptive sophistication—add little to established doctrines of tying, leveraging, and foreclosure. The Decision also falls short of the evidentiary standard appropriate for abuse-of-dominance findings, even at the interim-measures stage. It leans heavily on foreign precedent while giving insufficient weight to the procompetitive benefits of Apple’s vertically integrated model, particularly for privacy, security, and quality control.

Paradoxically, the case underscores the adequacy of Brazil’s existing competition framework. Law 12,529/2011 and CADE’s current enforcement tools proved sufficient to address the alleged harms. That weakens, rather than strengthens, the case for new ex ante digital-market regulation, including Bill 4,675/2025[1] and the more recent Bill 882/2026, both of which target ecosystem access more directly.[2] The Apple case in Brazil is less a vindication of ecosystem theory than a demonstration of its limits—and of the continuing need for rigorous economic analysis in competition enforcement.

The brief proceeds as follows. Section II provides the case’s factual and procedural background. Section III examines “ecosystem theory” as a competition-law framework, arguing that its core concepts largely repackage established antitrust doctrines without adding analytical value. Section IV turns to CADE’s application of that framework, focusing on its single-brand market definition for iOS app distribution. Section v examines the Decision’s theory of harm, arguing that it falls short of the evidentiary standard required in abuse-of-dominance cases, even for interim measures, and relies heavily on foreign case law while giving limited attention to the challenged conduct’s procompetitive effects. Section VI assesses broader implications for Brazilian competition law and digital regulation. Section VII concludes.

II.                    Background of the Case

In January 2023, Brazil’s Administrative Council for Economic Defense (CADE) opened an administrative inquiry into Apple for alleged abuse of dominance in “the market for the distribution of apps for iOS devices.” The investigation focused on Apple’s rules for app distribution and in-app payments on iOS.[3]

The case arose from a complaint by Mercado Livre—the leading e-commerce platform in Latin America—which alleged that Apple’s App Store policies barred developers from using alternative payment systems and tied access to iOS users to Apple’s proprietary in-app purchase (IAP) system. Apple requires developers that offer in-app purchases to use IAP, for which it charges commissions of up to 30%.

Mercado Livre also challenged Apple’s alleged use of discriminatory contractual terms. It argued that Apple restricted the distribution and promotion of third-party digital services and imposed so-called “anti-steering” provisions that barred developers from telling users about alternative payment options outside the app.

According to Mercado Livre, these practices restricted competition in digital payments, limited developer autonomy, and reinforced Apple’s dominance. At first glance, those allegations may appear to support antitrust intervention. A broader view suggests more skepticism. These disputes often concern how platform revenues are divided,[4] rather than competitive foreclosure that harms consumers and falls within antitrust law’s traditional scope. Like Spotify in the European Union proceedings[5] or Epic Games in U.S. litigation,[6] Mercado Livre appears primarily to challenge Apple’s control over parts of the value chain through which Mercado Livre could further monetize its own complementary services. Whether antitrust law should resolve such disputes is the central question of this brief.

This section provides the case’s factual and procedural background.

In October 2023, Mercado Livre asked CADE to impose interim measures requiring Apple to allow the sale of third-party digital goods and services through iOS or, alternatively, to permit developers to tell users about external offers and link them to alternative purchasing channels.

On Nov. 25, 2024, CADE’s General Superintendence (SG) instituted Administrative Proceeding No. 24/2024[7] and granted interim measures largely reflecting Mercado Livre’s requests. The SG ordered Apple to suspend enforcement of key contractual provisions supporting its anti-steering rules, mandatory use of IAP, and App Store exclusivity. In practical terms, the order required Apple to allow developers to inform users about alternative purchasing options; include external links and call-to-action mechanisms; use competing in-app payment systems; distribute iOS apps outside the App Store, including through sideloading and third-party app stores; and decouple App Store distribution from mandatory use of IAP, even for digital goods.

The SG also barred Apple from adopting functionally equivalent restrictions, required the company to make the necessary technical changes in Brazil, and ordered Apple to notify developers and publicly disclose the decision. To enforce compliance, the SG imposed daily fines of R$250,000, or about US$46,000. Apple had 20 days to implement the required changes in Brazil.

According to the SG, the challenged conduct involved two sets of practices: first, artificial barriers to entry and expansion in app distribution and digital goods within iOS, including mandatory use of Apple’s IAP system; and second, tying app-distribution services to Apple’s payment-processing system. Central to the SG’s analysis was a narrow, ecosystem-specific market definition that closely tracked Mercado Livre’s arguments.

The SG also relied heavily on foreign precedent. It cited the European Union’s Google Android[8] and Apple Music streaming decisions,[9] as well as the U.S. cases Epic Games, Inc. v. Apple Inc.[10] and Epic Games, Inc. v. Google LLC.[11] Apple appealed administratively, arguing that its integrated system protected privacy and security and that neither law nor economics justified forcing it to redesign its platform.

Apple also sought judicial review. A civil court suspended the interim measures, finding them “disproportionate and unnecessary” and imposed “without any demonstration of urgency or concrete risk to the market.”[12] CADE appealed, and an appellate judge partially reinstated the measures pending further review.

In February 2025, CADE held a public hearing on competition issues involving Apple’s iOS and Google’s Android mobile ecosystems.[13] Representatives from industry, civil society, and academia participated.[14]

In May 2025, CADE’s Tribunal upheld the interim measures, concluding that Apple’s conduct could restrict competition in app distribution and digital payments.[15] The proceeding quickly became one of Brazil’s highest-profile digital-platform cases, in part because it mirrored disputes Apple had faced in the United States, Europe, Latin America, and Asia.[16]

The Decision drew attention not only because of its outcome—which effectively opened Apple’s “walled garden”—but also because of its reasoning. The Tribunal described Apple’s business model as a “closed and vertically integrated ecosystem,” emphasizing that developers were “compelled” to rely on the App Store and IAP, while consumers were effectively “locked in” by switching costs. On that basis, CADE defined the relevant market as “the non-licensable smart mobile operating system iOS.”

That definition built on—and extended—the European Commission’s distinction in Google Android between “licensable” and “non-licensable” mobile operating systems.[17] As discussed below, that framing largely ignores the competitive constraints imposed by broader competition in smartphone-device markets.[18] Within the Tribunal’s defined market, Apple was necessarily the sole supplier and therefore dominant by definition.

On the merits, the Tribunal concluded that:

[T]here is strong plausibility in the contention of the unlawfulness of the tying practiced by Apple through the imposition of the IAP. The analysis of economic data and the submissions made by market participants demonstrates that this practice constitutes a sophisticated strategy of intra-product discrimination that disproportionately extracts value from specific developers, notably those of games and streaming applications.[19]

The Decision also marked the first time CADE expressly adopted what it called an “ecosystem theory of harm.”[20] The Tribunal reasoned that Apple’s integration across devices, operating systems, and services could let it foreclose competition in adjacent markets, such as payments. As discussed below, the analytical shortcuts embedded in that framework raise more questions than they answer.

Commissioner Gustavo Augusto Freitas de Lima issued a concurring opinion[21] endorsing the interim measures but rejecting the Tribunal’s market definition. He rejected “from the outset the thesis of the so-called ‘monopoly of oneself,’”[22] arguing that product differentiation and exclusivity alone do not create monopoly power. Freitas de Lima instead proposed a broader market for mobile-app stores, in which Apple and Google operate as a duopoly. Under that framework, Apple’s market power would stem not from exclusivity alone, but from revenue leadership and the purchasing power of its users.

Freitas de Lima also characterized the case as a conventional tying and leveraging dispute, drawing an explicit analogy to CADE’s 1993 Xerox precedent. He further cautioned that the European Union’s Digital Markets Act “does not apply in Brazil and has not been incorporated into the Competition Act.”[23]

In July 2025, the SG formally recommended that CADE convict Apple.[24] Before the Tribunal could issue a final decision, the parties settled. On Dec. 29, 2025, Apple entered into a Termo de Compromisso de Cessação (TCC),[25]  or cease-and-desist agreement. Under the TCC, Apple agreed to allow third-party app stores on iOS devices in Brazil, permit alternative payment processors for in-app purchases, and remove anti-steering restrictions that prevented developers from linking users to external offers.

The settlement may represent one of the broadest outcomes in global app-store antitrust enforcement to date.[26]

III.                    ‘Ecosystem Theory’ and Competition Law

The term “ecosystem” has clear rhetorical appeal. As a metaphor, it captures the interdependence among platform operators, business users, and consumers. It also helps describe how products and services can form broader networks of complementarities and dependencies. But its value in competition law remains doubtful. So far, ecosystem theory has added little to our understanding of market dynamics or to the methodological rigor of antitrust enforcement. In some respects, it may distort the analysis.

As Giuseppe Colangelo argues, the ecosystem concept has succeeded rhetorically without offering much analytical novelty.[27] Despite its frequent use in policy debates, the term lacks a precise legal definition and has not produced a coherent economic framework distinct from established theories of multi-sided markets. What it mostly offers is a convenient label for the scope and integration of large digital firms, not a new way to understand competition. The underlying issues—network effects, data synergies, interoperability, multi-homing, and indirect network effects—were already well understood through platform economics and existing antitrust doctrine.[28]

Colangelo’s review of European case law reinforces the point. He finds that “although a significant number of decisions potentially involve digital ecosystems, the term itself rarely appears and, when it does, it is not used to describe a clear or distinctive scenario.”[29] Even the European Union’s Digital Markets Act (DMA), which targets “contestability” and “fairness” in digital markets, refers to ecosystems only in passing and gives the term no operational meaning.[30]

The problem with ecosystem theory is not merely that it adds little. It can also distort the analysis. As Christopher Yoo observes, “it remains unclear what a concept of ecosystems defined in terms of complementary products adds to the principles governing vertical integration.”[31] The ecosystem framework risks obscuring what decades of economic literature have consistently shown: vertical integration and vertical restraints are often procompetitive and beneficial to consumers.[32]

Yoo further notes that ecosystem theory depends on a platform occupying a “central position” within an ecosystem—a concept related to, but distinct from, dominance.[33] That framing risks turning market success into evidence of liability:

Defining harm in terms of economic significance suffers from significant endogeneity problems. Achievement of a central or important position can be the result of either successful competition on the merits or anticompetitive activity. Using those outcomes as a basis for liability risks treating causes as effects.[34]

The concern is that the ecosystem framework may encourage suspicion toward business practices that are often efficient. Modern antitrust law has largely moved away from presumptions against vertical restraints, recognizing that such arrangements require context-specific analysis. As the Competition Committee of the Organisation for Economic Co-operation and Development’s (OECD) Directorate for Financial and Enterprise Affairs explained:

The ambiguous effect of vertical restraints, and the consequence that one cannot simply outlaw certain vertical restraints and permit others, suggests that a rule of reason (as opposed to per se rule) approach is preferred. Economic analysis in this regard can help to determine which types of agreements are capable of and likely to raise competition concerns, while it can also help to understand what types of justifications could be judged to outweigh identified anti-competitive effects.[35]

Proponents of ecosystem theory do not expressly advocate per se illegality. But when the framework relies on presumptions, circular findings of dominance, or generalized suspicion toward integration, it can produce much the same result.

Applied this way, ecosystem theories of harm may also chill innovation. Integrated design—the seamless interaction among hardware, software, and services—often creates substantial consumer benefits, especially in security, privacy, and quality control.[36] Treating that integration as suspect because it creates “lock-in” could deter firms from pursuing efficiency-enhancing complementarities. The result would revive a structural hostility toward vertical integration that modern antitrust law, after decades of economic learning, had largely abandoned.

IV.                    CADE’s Market Definition and Theory of Harm

As argued above, ecosystem theory—while descriptively useful—risks becoming a doctrinal shortcut. It can encourage excessively narrow market definitions and overstate business users’ “dependencies” on a platform, lowering the evidentiary bar for finding dominance and abuse. This section asks whether that risk materialized in CADE’s treatment of Apple. It appears to have done so. Although the Decision concerned interim measures,[37] which naturally involve a lower evidentiary threshold, CADE appears to have lowered that threshold too far,[38] adopting premises that could distort future antitrust enforcement.

Before addressing the merits, the Decision introduces a “foundational semantic framework regarding the operational structure of mobile digital ecosystems,”[39] describing ecosystems as a “‘new’ explanatory lens for the competitive dynamics of digital markets in which large technology companies exploit business models based on multiple layers of intermediation.”[40]

According to the Tribunal:

The economic logic of ecosystems has relevant implications for antitrust policies. Firstly, ecosystems are characterized by new forms of internal competition, manifested as “vertical competition” between products that are not directly substitutable. This phenomenon calls into question the centrality of the notion of the “market” in antitrust analysis, since the delimitation of spaces of substitutability where competitive forces mutually constrain each other becomes complex, as does the apprehension of anti-competitive effects in interconnected economic activities.

Secondly, there is a strengthening of economic dependencies between the complementors and the central platform orchestrator, the latter being able to unilaterally alter the conditions of competition in the various adjacent markets. While orchestration is crucial for coordination in the ecosystem, design and governance decisions can generate functional and distributive failures, resulting in a loss of value for all participants and hindering innovation. The framing of these decisions under competition laws blurs the boundaries between the exclusionary and exploitative nature of abusive conduct.[41] [Emphasis ours, citation omitted].

The Tribunal’s framing rests on two propositions. First, ecosystems supposedly involve new forms of “vertical competition” that require rethinking market definition in digital-platform cases, effectively narrowing the unit of analysis to the platform itself. Second, “economic dependencies” between complementors and the platform owner allegedly allow the latter to “unilaterally alter the conditions of competition.” Both propositions are problematic.

First, the suggestion that digital ecosystems diminish the relevance of traditional market definition is a non sequitur. A digital ecosystem—whether understood as “an organization that sells interconnected products or services” or as a multi-sided platform[42]—is not inherently more complex than a “traditional” market. Complexity depends on the conduct at issue and the market under examination. Some digital-platform cases are complex; the Google AdTech litigation is an obvious example.[43] But digital markets are not uniquely so. A case involving a social-media platform or e-commerce marketplace is not obviously more complex than one involving health insurance, payment networks, or financial exchanges. Moreover, complexity does not justify analytical shortcuts.

Rigorous competition analysis still requires the same core inquiry: who exercises market power, against whom, and whether that power can be exercised profitably and sustainably. Multi-sided-platform economics provides tools for conducting precisely that analysis. As the U.S. Supreme Court recognized in Ohio v. American Express, platforms that facilitate transactions between different user groups must be analyzed as integrated markets that account for competitive dynamics on all sides.[44] That does not require abandoning traditional market definition. It requires applying it correctly.

The purpose of market definition remains straightforward: identifying the competitive constraints a firm faces through demand substitution, supply responses, and potential entry. That inquiry remains indispensable in digital-platform cases. Without it, there is no principled basis for distinguishing firms with substantial market power from firms that operate successful businesses in competitive markets.

Second, the Tribunal’s emphasis on “economic dependencies” is equally problematic. Dependency is inherently contextual, and treating it as a starting assumption risks creating a self-fulfilling theory of market power.[45] The fact that a platform operator can set the terms for access to its ecosystem does not establish market power or meaningful dependency by business users. In many cases, business users multi-home precisely to reduce dependence on any single platform.[46]

What the Tribunal describes—a platform setting rules for access to its infrastructure—is ordinarily an exercise of property rights and contractual freedom. Every private marketplace, from shopping malls to stock exchanges, sets rules for participation. The relevant antitrust question is not whether a platform controls access to its infrastructure; it inevitably does. The question is whether users had viable alternatives when they chose to participate.

That question matters here because CADE never established the market power needed to support its theory. The Decision correctly notes that Apple users must use the App Store to install apps and Apple’s payment system for digital purchases.[47] Because iOS is a “closed” system,[48] users ostensibly have “no choice” but to use Apple’s infrastructure. Except they do have a choice: consumers can buy different smartphones, and developers can prioritize different platforms.

As Craig Conrath explains:

[I]t is necessary to distinguish between contract power and monopoly power, between post-contractual opportunism and abuse of dominant position. … Therefore, whenever there is a claim of monopolistic practice by a party in an contractual relationship with the person accused of a monopolistic practice, it is important not to ask the question, “Does this party now have any other choices?” This is a question of contract law. Instead, one should ask the relevant antimonopoly law question: When the contract was made, did this party have other choices?[49]

Earlier in the Decision, the Tribunal surveyed foreign cases involving Apple’s App Store practices, purportedly to “establish objective parameters for legal comparability.”[50] To its credit, the Tribunal reviewed precedent thoroughly, including cases that cut against its position. Even so, much of this discussion functions more as rhetoric than analysis.

The Tribunal, for example, relies heavily on Judge Yvonne Gonzalez Rogers’ opinion in Epic Games v. Apple.[51] As the Decision notes, the court rejected Epic’s Sherman Act claims while still finding certain practices problematic under California law.[52] The Tribunal then quotes portions of the opinion suggesting Apple’s conduct could be “exploitative.”

That reliance is difficult to square with the actual holding. The relevant findings arose under California’s Unfair Competition Law,[53] not under the Sherman Act. That statute imposes a substantially less demanding standard than federal antitrust law’s rule-of-reason framework. Its relevance to a Brazilian abuse-of-dominance case is therefore far from obvious.

More importantly, the Tribunal’s reliance on Epic indirectly imports the logic of Eastman Kodak Co. v. Image Technical Services,^[54] where the U.S. Supreme Court accepted that a derivative aftermarket could constitute a distinct market if consumers were “locked in.” Judge Gonzalez Rogers cited Kodak in the passage quoted by CADE:

In the context of technology markets, the open flow of information becomes even more critical. As explained above, information costs can create a lock-in for platforms, since users have no information about the lifetime costs of an ecosystem. Users may also lack the ability to attribute costs to the platform rather than the developer, which further prevents them from making informed choices. In these circumstances, the ability of developers to provide information across platforms is crucial. Although Epic Games has not met its burden of demonstrating actual blocking on this record, the Supreme Court has recognized that such information costs can create the potential for anticompetitive exploitation of consumers. Eastman Kodak, 504 U.S. at 473-75, 112 S. Ct. 2072.^[55]

Subsequent U.S. case law has narrowed the circumstances in which single-brand aftermarket claims can succeed. Judge Gonzalez Rogers herself noted that the 5th U.S. Circuit Court of Appeals held in United Farmers Agents Ass’n v. Farmers Insurance Exchange[56] that such claims fail where customers “would clearly have become aware of [the alleged anticompetitive] policy long before they faced significant switching costs.”[57]

That principle cuts strongly against CADE’s position. Apple’s integrated distribution model—exclusive use of the App Store, mandatory IAP, and restrictions on alternative distribution channels—has been publicly known since the App Store launched in 2008. Developers and consumers therefore entered the ecosystem with advance knowledge of Apple’s rules. Where there is no significant information asymmetry at the time of the initial purchasing decision, the economic logic for recognizing a single-brand aftermarket largely disappears.

Benjamin Klein made a similar point in his critique of Kodak. Echoing Conrath,[58] Klein argues that market power must be assessed before users make platform-specific investments: “the level of competition should be measured before the buyer makes any specific investments. If the market at this time is competitive, then the tie is merely part of the total freely negotiated competitive price.”[59]

At this point, the Decision attempts to distance itself from foreign doctrine, emphasizing that the case rests “exclusively on the application of the Brazilian legal framework,” particularly Article 173, § 4 of the Constitution and Law 12,529/2011.[60] Formally, the alleged infringements involve familiar doctrines: tying through mandatory use of Apple’s IAP system and discriminatory application of the App Store Review Guidelines.[61]

CADE thus presents ecosystem theory not as the theory of harm itself, but as the broader analytical framework within which traditional theories operate. That characterization understates the work ecosystem theory performs in the Decision. The framework drives the single-brand market definition for “distribution of apps for iOS devices”—a market that could not exist without treating iOS as a self-contained competitive space. It also justifies remedies far broader than those adopted elsewhere, including mandatory opening of Apple’s entire distribution and payment architecture.

The relevant question should instead be whether developers and consumers had viable alternatives when they chose to develop for, or purchase, iOS devices. In conventional antitrust terms, the question is whether Apple’s App Store faces meaningful substitutes.

The Tribunal approaches this issue by examining substitutability from the perspective of device manufacturers, developers, and consumers. In each case, it concludes that iOS and Android are not substitutes. The analysis is flawed at every stage.

The Decision first observes that “for mobile device manufacturers, it would not be possible to address substitutability between licensable and non-licensable systems.”[62] That observation is true but analytically irrelevant. Apple’s refusal to license iOS reflects a business strategy built around vertical integration among hardware, software, and services. By choosing that model, Apple gives up licensing revenue to compete through differentiated products and user experience. That does not insulate it from competition.

Apple competes not only with Google’s Android ecosystem, but also with device manufacturers such as Samsung, Xiaomi, OPPO, vivo, and Huawei. As the district court observed in Epic v. Apple, “it is illogical to argue that there is a market for something that is not licensed or sold to anyone. Competition exists for smartphones, which are more than just the operating system.”[63] Consumers buy smartphones, not operating systems.

That point has direct implications the Tribunal largely ignores. Competition at the device level constrains conduct at the operating-system level. If Apple used its App Store rules to extract excessive rents from developers, the resulting deterioration in app quality, availability, or pricing would weaken the iPhone’s competitiveness against Android-based devices.

The Tribunal’s distinction between “licensable” and “non-licensable” operating systems therefore obscures the competitive dynamics at issue. Apple’s integrated model and Google’s licensing-based model are competing strategies for attracting users, developers, and manufacturers to mobile platforms. That rivalry has generated substantial innovation and consumer benefits. As David Evans explains:

Innovations by Apple and Google, in particular, have led to spread of smart mobile phones around the world, enabling billions of people to consumer Internet-based services and millions of businesses to provide mobile-app based services to them. Apple introduced the iPhone, which consisted of a powerful computer, a mobile operating system, and a standard set of applications including a mobile browser in June 2007. Google invested in developing a mobile operating system, Android, which it ran as an open-source project, and developing and organizing an ecosystem of handset makers, mobile network operators, and other technology partners. It introduced the first Android phone in October 2008. Apple and Google also stimulated the production of mobile apps by providing software tools for developing apps for their operating systems, creating a quality certification process for these apps, and creating “app stores” that provided centralized places for developers to distribute apps and for users to download them on their mobile devices.[64]

Competition between Apple and Google extends beyond operating systems. It includes hardware, software integration, AI capabilities, security, privacy, and the overall user experience. As ICLE observed in comments submitted to the UK Competition and Markets Authority:

The fierce competition between Apple and Google is waged through continuous improvements in camera technology, the race for superior processing power (with Apple’s A-series chips and Google’s Tensor chips forming the basis of their respective performance claims), and divergent—yet intensely competitive—approaches to user security and privacy, which have become central pillars of each company’s marketing and value proposition.

Most recently, this rivalry has shifted to the next frontier of innovation: the deep integration of artificial intelligence into the core user experience, with the corollary that both firms have invested massively invested in AI technology (and continue to do so). To observe this landscape of ceaseless innovation and declare it a market with “limited competition” is to misunderstand its fundamental nature.[65]

One recent example is Google’s rollout of Gemini “task automation,” initially available only on recent Pixel and Samsung devices.[66]

By treating these business models as non-substitutable, the Tribunal excludes one of the most important constraints on Apple’s conduct: interbrand competition between mobile ecosystems.

The Tribunal reaches similarly flawed conclusions from the perspective of developers. It argues that iOS and Android are not substitutes because app development differs across systems, migration is difficult, and developers typically operate on both platforms.[67] Ironically, the Tribunal treats multi-homing as evidence of non-substitutability. In reality, multi-homing reduces dependence on Apple. Developers operate on both platforms precisely because they can reach consumers through alternative channels. Android-only development remains commercially viable, especially in Brazil, where Android accounts for more than 80% of smartphones.[68]

From the consumer perspective, the Tribunal concludes that “there would be no substitutability” because operating systems are tied to devices and Android phones generally cost less than iPhones in Brazil.[69]

The Tribunal identifies real differences but draws the wrong conclusion. Consumers choose smartphones based on many characteristics, including operating systems, app ecosystems, cameras, design, connectivity, and AI features. Apple therefore competes not only with Google, but with manufacturers such as Samsung, Xiaomi, Huawei, and Oppo.[70] That fact supports a broader, not narrower, market definition.[71]

The Tribunal also emphasizes iPhones’ higher average prices.[72] Economic theory recognizes that products in different quality or price tiers can constitute separate markets where cross-elasticity of demand is low. But that conclusion requires evidence, not mere price differences.[73]

Android manufacturers increasingly compete directly with Apple in the premium-smartphone segment. Devices priced above US$600 represented 25% of global smartphone unit sales in 2024[74] and more than 60% of industry revenue.[75] Premium-segment sales continued growing through 2025,[76] with Samsung, Huawei, and Google gaining share against Apple.[77] Samsung’s Galaxy S26 Ultra, Google’s Pixel 10 Pro XL, and Apple’s iPhone 17 Pro Max launch at comparable price points,[78] undermining the notion that Apple occupies a uniquely expensive segment insulated from competition.[79]

The Decision also notes that “only a portion of consumers can afford to replace Android with iOS.”[80] But competition occurs at the margins. Effective competitive constraints do not require every Android user to switch to Apple, or vice versa. They require only enough marginal consumers to switch in response to price increases or quality degradation to discipline conduct. As Simon Baker and Laurence Wu explain:

Customers who can easily switch from one product to another and who would do so in response to a price increase are known as marginal customers, whilst those who are unlikely to do so (…) are termed infra-marginal customers. Infra-marginal customers are protected by the willingness of marginal customers to switch so long as there are enough marginal customers and so long as sellers cannot effectively identify the marginal from the infra-marginal customers…

The ability of these customers to switch just a portion of their total purchases from one product to another can be as effective in disciplining pricing as the complete switching of purchases by a few customers.[81]

The Decision also largely ignores substitution from iOS to Android, even though such switching may be easier given Android’s lower average prices. More importantly, CADE never analyzes actual switching rates in Brazil.

Available evidence from other jurisdictions strongly suggests dynamic competition between mobile ecosystems. As ICLE explained:

Consumers can readily find myriad comparisons of Android and iPhone devices online. Moreover, both Apple and Google maintain webpages that offer to help users switch from one platform to the other. The business press has extensively covered the fierce rivalry between the two companies. And numerous academic studies have reached similar conclusions about the nature of their competition. Nicolas Petit refers to Apple and Google as “moligopolists,” while David Evans has described their rivalry as “dynamic competition.” Marshall Van Alstyne and his coauthors have analyzed the strategies that both Google and Apple have deployed to outcompete one another. (Citations omitted).[82]

Dirk Auer made a similar point in criticizing the European Commission’s Google Android decision.[83] He noted that even the Commission’s own data showed substantial switching and contestability:

Take the claim that 82% of Android users stick with Android when they change phones (compared to 78% for Apple), and that 75% of new smartphones are sold to existing users. The Commission asserted, without further evidence, that these numbers prove there is little competition between Android and iOS.

But is this really so? In almost all markets consumers likely exhibit at least some loyalty to their preferred brand. At what point does this become an obstacle to interbrand competition? The Commission offered no benchmark mark against which to assess its claims.

And although inter-industry comparisons of churn rates should be taken with a pinch of salt, it is worth noting that the Commission’s implied 18% churn rate for Android is nothing out of the ordinary, including for industries that could not remotely be called anticompetitive.

To make matters worse, the Commission’s own claimed figures suggest that a large share of sales remained contestable (roughly 39%). Imagine that, every year, 100 devices are sold in Europe (75 to existing users and 25 to new users, according to the Commission’s figures). Imagine further that the installed base of users is split 76–24 in favor of Android. Under the figures cited by the Commission, it follows that at least 39% of these sales are contestable. [Emphasis added].

Likewise, a survey commissioned by the UK Competition and Markets Authority found that 8% of iOS users had switched from Android, 5% of Android users had switched from iOS, and additional users on both sides considered switching. Those are meaningful levels of substitution.[84] Yet the Decision analyzes no comparable data for Brazil.

This omission matters because operating systems are multi-sided platforms. As David Evans and Michael D. Noel explain, prices and competitive effects on one side of a platform cannot be analyzed independently from the other.[85] Even if developers face switching costs, Apple’s market power remains constrained so long as consumers can switch between ecosystems. Conversely, even if consumers face switching costs, Apple’s treatment of developers is constrained by the risk that restrictive terms could push developers toward Android-first or Android-only strategies, reducing the attractiveness of the iOS ecosystem.

The Tribunal briefly acknowledges switching costs arising from interoperability and compatibility preferences, noting that consumers prefer compatible smart devices and may therefore become “locked in.”[86] That observation has merit. But the Decision never examines actual substitution patterns in Brazil, nor does it adequately consider that interoperability itself is a dimension of competition that can enhance consumer welfare.

Later, the Tribunal appears to endorse the SG’s view that Apple’s ecosystem design itself demonstrates market power because Apple controls “all the rules and conditions of competition” within iOS.[87] This reasoning risks becoming circular. Every platform owner controls participation rules on its platform. That fact alone does not establish dominance. Market power depends on whether users had viable alternatives when deciding whether to participate. As discussed above, they did.

The circularity becomes explicit in the Tribunal’s conclusion that Apple “has market power in the form of the ability to unilaterally alter the conditions of competition in the markets related to its smart mobile digital ecosystem.”[88] In other words, Apple is dominant because it controls its own ecosystem.

This problem did not escape notice within the Tribunal itself. As discussed above,[89] Commissioner Freitas de Lima rejected “from the outset the thesis of the so-called ‘monopoly of oneself,’” arguing that product differentiation and brand exclusivity do not themselves create monopoly power. Whatever the merits of his proposed alternative market definition, his critique mirrors the central problem identified here: Apple’s market power cannot be established by definitional fiat.

A broader market definition would better align with established antitrust methodology and with cases such as Epic Games v. Apple, where the court rejected a single-brand iOS market and instead analyzed competition across mobile platforms.[90] Similarly, the U.S. Department of Justice recently defined a market for “performance smartphones” in its case against Apple.[91] Even that narrower framework—which still includes competing devices such as Samsung’s Galaxy S26 Ultra and Google’s Pixel 10 Pro XL—is substantially broader than CADE’s single-brand market.

The distinction matters because Apple’s position in Brazil looks very different under a broader market definition. Apple’s share of Brazilian smartphone sales is estimated at roughly 15-20% by units sold—well below levels typically associated with dominance under Law 12,529/2011.[92] The gap between Apple’s actual market position and the Tribunal’s dominance finding illustrates the consequences of adopting an excessively narrow market definition.

A careful reading of the Decision suggests that CADE could have pursued its case without invoking ecosystem theory at all. The problem is not merely that ecosystem theory was unnecessary. It is that the framework pushed the analysis toward a tautological finding of dominance based on Apple’s control over its own platform.

V.                    Evidentiary Standards and the Limits of Ex Post Enforcement

As discussed in the previous section, the Tribunal’s single-brand market definition for iOS app distribution and in-app payments lacks support under established antitrust methodology. That defect alone should substantially weaken the case: absent dominance, the vertical agreements governing the App Store and Apple’s IAP system cannot constitute abuse of dominance. Even so, CADE’s characterization of Apple’s conduct as exclusionary discrimination (“self-preferencing”) and unlawful tying warrants closer examination.

The Tribunal’s decision upholding interim measures against Apple combines ecosystem theory with traditional antitrust doctrines, including tying, self-preferencing, and foreclosure of rival distribution channels. Commissioner José Levi Mello do Amaral Júnior Fernandes presents these theories as mutually reinforcing: ecosystem theory supplies the descriptive framework, while conventional antitrust doctrines provide the legal structure. This section argues that the Decision falls short of the evidentiary standard that abuse-of-dominance cases should require, even at the interim-measures stage.

Interim measures are valuable enforcement tools, particularly in digital markets where proceedings often move more slowly than the markets themselves. But they still require reasonably robust evidence, especially where the remedy would radically alter market structure. Here, the Decision relies primarily on theories and evidence from foreign cases, rather than Brazil-specific evidence. It also gives remarkably little attention to the procompetitive effects of the conduct it condemns.

That omission is striking because the Decision itself acknowledges the evidentiary gap. At several points, the Tribunal concedes that definitive conclusions would require “supplementing the evidentiary investigation” and “a more detailed analysis of the competitive conditions” in the relevant markets.[93]

The reliance on foreign material is structural, not incidental. After describing Mercado Livre’s claims and testimony from the Public Hearing on Mobile Digital Ecosystems[94]—much of it from complementors raising concerns similar to Mercado Livre’s—the Tribunal simply asserts that “these practices would have the effect of impairing competition and restricting the promotion of alternatives, while Apple’s own apps do not face the same costs arising from these bans.”[95] The Decision provides no concrete evidence of actual or likely harm to competition in Brazil.

The tying analysis in Section 5.3.4 is similarly thin. Even assuming the Tribunal correctly concluded that Apple’s operating system and IAP system are “separate products,”[96] the evidence of competitive harm again comes almost entirely from theory and foreign precedent. The Decision relies on the U.S. Supreme Court’s Jefferson Parish framework, the European Commission’s Article 82 Guidance,[97] and academic literature describing how tying arrangements can reduce welfare.[98]

The Brazilian precedents cited—Comgás, PTI, Steel Placas, and the rapporteur’s opinion in Google Shopping—do little more than establish that tying and discriminatory conduct can, in principle, violate Article 36 of Law 12,529/2011. None involved vertically integrated mobile ecosystems, and none supplied the empirical record the Tribunal purported to apply.

To be clear, competition authorities may properly rely on academic literature and foreign jurisprudence. Comparative analysis is often useful, especially in cases involving novel business models or market structures. The Tribunal deserves credit for engaging seriously with international debates and precedent.

The problem is that the Decision uses those authorities mainly to establish the theoretical plausibility of its theory of harm. Plausibility is not proof that challenged conduct harms competition in the relevant market. The Decision devotes substantial attention to what courts and scholars in the United States and Europe have said about Apple’s business model, while doing far less to test those propositions against Brazil’s competitive realities.

The Decision is also notably thin on procompetitive effects. Modern antitrust systems, including Brazilian competition law, generally require effects-based analysis that weighs anticompetitive risks against efficiencies before condemning vertical restraints. The Tribunal acknowledges that principle in the abstract,[99] including CADE precedent recognizing that bundled or joint sales can “reduce the impact of double profit margins and enable economies of scale and scope arising from joint production or distribution.” Yet the analysis never meaningfully returns to those efficiencies.

Instead, the 30% commission, App Store-IAP integration, and anti-steering rules are treated only as potential sources of harm. The countervailing benefits—security investment, fraud prevention, curation, cross-subsidization of the more than 80% of apps that pay no commission,[100] and reduced transaction costs for users and developers—receive little comparable attention. An interim measure that effectively restructures a multi-sided platform should, at minimum, rest on a record that seriously evaluates both sides of the competitive ledger.

Apple’s procompetitive justifications are substantial and have been recognized even by courts skeptical of other aspects of its conduct. In Epic Games v. Apple, Judge Yvonne Gonzalez Rogers concluded that Apple’s integrated approach generated real consumer benefits in privacy, security, and curation—benefits that operate as dimensions of interbrand competition against Android’s more open model.[101]

The economic literature on vertical integration in technology markets reaches similar conclusions. Integration among hardware, software, and services often reduces transaction costs, improves quality control, and internalizes externalities that arise in modular systems.[102]

Randal Picker identifies another dimension the Tribunal largely overlooks: security competition between ecosystems. Apple and Google compete through markedly different security architectures. According to Picker, forcing Apple to permit open app distribution would reduce security competition by pushing iOS toward Android’s more permissive—and demonstrably more vulnerable—model.[103]

The Tribunal itself recognizes that Apple competes partly through privacy and security, while Google emphasizes openness and advertising monetization.[104] Yet the Decision never fully confronts the implications of that observation. If Apple’s closed architecture is a differentiated competitive offering, mandating openness may reduce, rather than increase, consumer choice at the ecosystem level.

Implicit in the Decision is the notion that digital ecosystems should operate as “open squares”—spaces where platform operators cannot meaningfully restrict access. But that is not an economic conclusion. It is a regulatory preference. Some jurisdictions, most notably the European Union through the Digital Markets Act (DMA),[105] have adopted that preference expressly through legislation.

The relevant question, then, is whether ex post antitrust enforcement can properly reach the same result through abuse-of-dominance doctrine—and whether it should.

The Tribunal’s theory of harm also understates the costs of the remedies it contemplates. Requiring Apple to permit third-party app stores and alternative payment systems on iOS would necessarily alter the platform’s security architecture.[106] The Tribunal acknowledges the sideloading debate and Apple’s claims about malware differentials between iOS and Android.[107] But it ultimately dismisses those concerns as pretextual or addressable through less restrictive alternatives, without the detailed analysis such a conclusion requires.

The literature on forced interoperability suggests those costs are not trivial. Research by the Center for European Policy Analysis (CEPA) argues that interoperability mandates create new attack surfaces and increase opportunities for malicious actors, potentially undermining the end-to-end security associated with vertically integrated systems.[108] Picker likewise suggests that mandated openness would not merely alter Apple’s business model; it would eliminate a dimension of competition consumers value.[109]

Innovation incentives are also at stake. Sven Völcker and Daniel Baker argue that Apple’s vertically integrated model encourages continuous investment in platform improvements—including security updates, APIs, and app-review infrastructure—because Apple captures part of the value the ecosystem creates through commissions.[110] If forced openness reduces or eliminates that revenue stream, investment incentives may weaken.

That concern is not merely theoretical. Early experience under the DMA suggests that compliance with interoperability and openness mandates can divert substantial engineering resources away from innovation and toward regulatory compliance.[111]

VI.                    Broader Implications for Brazilian Competition Law

The Apple proceedings carry implications well beyond the parties involved. The case tests the boundaries of Brazil’s competition framework along three dimensions. Analytically, it raises whether the Tribunal’s market-definition methodology and evidentiary standards risk overenforcement, particularly in digital-platform cases involving vertical integration and ecosystem-based theories of harm. Institutionally, it raises whether Brazil’s existing ex post competition framework is sufficient for digital markets or whether the case instead supports a new layer of ex ante regulation. Remedially, it highlights the tension between antitrust enforcement and sector-specific regulation, especially where competition remedies become detailed and prescriptive.

These issues are closely related. Lower evidentiary standards and expansive theories of dominance can facilitate increasingly intrusive remedies, blurring the line between competition enforcement and ongoing regulation. The Apple case therefore serves both as a test of the analytical discipline imposed by the consumer welfare standard and as a case study in the institutional limits of competition law in fast-moving digital ecosystems.

A.      Error Costs and the Consumer Welfare Standard

Brazilian competition law, as codified in Law 12,529/2011, uses an effects-based, efficiency-sensitive framework rather than formalistic prohibitions. The statute defines anticompetitive conduct by reference to its object or effects and recognizes that efficiencies and consumer benefits may justify conduct or transactions that otherwise raise concerns.[112] Consistent with that framework, CADE’s decisional practice has generally focused on consumer-welfare variables, asking whether challenged conduct or mergers are likely to affect prices, output, quality, or innovation, and weighing potential harms against cognizable efficiencies.[113] That orientation appears in CADE’s guidelines and in Brazil’s submissions to the Organisation for Economic Co-operation and Development (OECD), which emphasize that enforcement remains focused on conduct’s economic effects on consumers, even when broader policy considerations enter the discussion.[114]

The Apple proceedings risk departing from that tradition. By defining a single-brand market, the Tribunal collapses the distinction between property rights and market power: any firm that controls its own platform becomes dominant within it by definition. That lowers the threshold for finding dominance. By invoking ecosystem theory—even as a framing device—the Tribunal also lowers the threshold for finding competitive harm, overstating lock-in while underweighting the well-documented benefits of integration, including quality, security, and innovation.[115]

This move is not costless. As Frank Easterbrook showed in his foundational error-cost analysis, false positives in antitrust enforcement—the mistaken condemnation of procompetitive conduct—impose costs that markets cannot easily correct.[116] False negatives may eventually be disciplined by entry and innovation. False positives, by contrast, can chill efficient conduct and deter beneficial innovation across markets operating under the shadow of precedent. Joshua Wright and Murat Mungan applied this insight to digital markets, concluding that platform cases warrant a stronger evidentiary standard than preponderance of the evidence, precisely because the costs of mistaken intervention are amplified in fast-moving sectors with novel business models and rapidly shifting competitive dynamics.[117]

Geoffrey Manne has likewise argued that erroneous enforcement is both more likely and more costly in digital markets, where “antitrust decisions are made in innovative, fast-moving, poorly understood, or novel market settings” and where “erroneous interventions against innovation and the business models used to deploy it threaten to deter subsequent innovation.”[118] The CADE proceedings illustrate that risk. The most exhaustive judicial treatment of the same conduct concluded that Apple’s relatively closed ecosystem generates significant procompetitive benefits in privacy, security, and quality assurance. Condemning that model—or restructuring it through interim relief before a final merits determination—risks exactly the kind of false positive the error-cost literature warns against: suppressing an efficient vertical-integration strategy based on an unsupported theory of dominance.

The implications for Brazilian competition law are significant. If authorities can establish dominance by defining a market in which the respondent is the sole supplier by construction, and infer competitive harm from ecosystem structure rather than rigorous economic analysis, the consumer welfare standard no longer meaningfully constrains enforcement discretion. It becomes an ex post rationalization rather than an ex ante analytical discipline—the very outcome the error-cost framework is meant to prevent.

B.       The Case Against New Ex Ante Regulation

There is, however, a second—and clearer—reading of the Apple proceedings. Whatever the Decision’s analytical shortcomings, the case demonstrates that Brazil’s competition framework—Law 12,529/2011, CADE’s institutional structure, and the procedural tools available to the General Superintendence and Tribunal—already provides the mechanisms necessary to address competition concerns in digital markets. CADE opened an investigation, developed a theory of harm, imposed interim measures, conducted adversarial proceedings with full due process, and ultimately resolved the dispute through a negotiated settlement that imposed structural remedies on a global technology firm.

That is a significant institutional achievement. It should inform Brazil’s ongoing legislative debate over digital-market regulation. Bill 4,675/2025, currently before Congress, would create a new Digital Markets Superintendence within CADE and establish an ex ante regulatory framework for firms designated as “systemically relevant” in digital markets.[119] Proponents argue that traditional ex post antitrust enforcement cannot adequately address structural features of digital platforms, such as network effects, switching costs, and data advantages, and that ex ante regulation is therefore necessary to prevent harm before it occurs.[120]

The Apple case complicates that narrative. It shows that CADE can intervene aggressively against a global platform using its current ex post tools, including interim measures that, whatever their analytical merits, produced in months what years of litigation failed to achieve elsewhere. The real question is not whether Brazil’s competition framework can address digital markets, but whether it should do so through the existing case-by-case adjudicatory model or through a new layer of ex ante regulation.

The latter approach carries substantial risks. As Dario Oliveira Neto argues, it could “subtly sweep consumer welfare under the rug” by introducing regulatory objectives—such as “protection of the competitive process” and “reduction of barriers to entry”—that may sound consistent with competition policy but, in practice, diverge from the consumer welfare standard in both logic and application.[121]

The more prudent path is to preserve the existing adjudicatory model while insisting on the analytical rigor the consumer welfare standard requires.[122] The Apple case should serve both as proof of concept and as a cautionary tale: proof that Brazil’s current framework can address digital-platform disputes, and a reminder that its legitimacy depends on the quality of the analysis applied within it.

Lowering evidentiary standards in the name of addressing novel market structures does not strengthen competition enforcement. It weakens the institutional credibility on which effective enforcement depends.

C.      Antitrust Remedies or Sectoral Regulation?

The Decision did not merely resolve a precautionary dispute. It also laid out a remedial blueprint that anticipated—and effectively invited—the highly prescriptive outcomes later incorporated into the December 2025 settlement. That matters because the analytical weaknesses identified above were not merely theoretical. They shaped the remedies ultimately imposed on Apple and may do the same in future cases involving digital platforms.

The Decision identified three “principal pillars” of relief: (i) removal of anti-steering provisions; (ii) unbundling payment processing from app distribution; and (iii) enabling alternative app distribution channels, including third-party app stores and sideloading.[123] Standing alone, none of these remedies is especially unusual. The first closely mirrors the injunction upheld in Epic Games v. Apple,[124] while the second and third resemble obligations imposed under the European Union’s DMA.[125] What distinguishes the Decision is its remedial ambition: it pursued all three simultaneously through a precautionary measure issued before any merits determination.

The Tribunal was not unaware of the tension. Commissioner José Levi Mello do Amaral Júnior Fernandes acknowledged that, if Apple’s distribution “monopoly” were maintained, “CADE would have to arbitrate the prices charged by the monopolist, which is typical of a regulatory body and outside the regime of Law 12,529/2011.”[126] He also recognized that, at the precautionary stage, “it would be premature and methodologically inappropriate to establish ex ante the definitive set of appropriate technical safeguards” for sideloading.[127]

Those admissions are revealing. The Decision simultaneously recognized that price-setting and detailed technical regulation fall outside antitrust’s proper domain while constructing a framework that almost inevitably led there.

That outcome arrived seven months later. The December 2025 settlement operationalized the Tribunal’s three principles by specifying what the Decision itself acknowledged it could not: precise commission rates of 10% or 20% for App Store transactions, depending on the payment processor; a 5% “core technology commission” for apps distributed through alternative stores; and a 15% fee on transactions initiated through steering links.[128] CADE also required any security warnings shown to users to use “neutral and objective” language, effectively dictating the content of Apple’s consumer communications.[129] The settlement imposed a 105-day implementation timeline and penalties of up to R$150 million for noncompliance. Notably, it ultimately omitted the sideloading requirement.

The specificity of these obligations inverts the conventional hierarchy of antitrust remedies. Herbert Hovenkamp argues that antitrust relief should proceed incrementally: first targeted injunctions against challenged conduct, then quasi-structural remedies such as interoperability mandates, and structural interventions only as a last resort.[130]

The logic is institutional as much as economic. Courts and competition authorities generally lack the continuous supervisory capacity characteristic of sector-specific regulators, including detailed industry expertise, access to firm-level information, and the administrative infrastructure necessary to oversee ongoing compliance. Remedies prescribing prices or business architecture may therefore produce greater welfare losses than the conduct they seek to correct.[131]

Comparisons with other jurisdictions underscore the point. In the United States, the most exhaustive judicial treatment of the same conduct—Epic Games v. Apple—produced a single, narrowly tailored injunction against anti-steering provisions, leaving Apple’s commission structure, payment system, and distribution model intact.[132] In the European Union, the DMA’s obligations emerged from a legislative process involving stakeholder participation, impact assessments, and continuing administrative oversight by the European Commission.[133] Brazil’s settlement achieved outcomes more expansive than either regime—mandating alternative app stores, alternative payment systems, and specific fee structures—through a precautionary proceeding resolved before any merits determination.

As Dennis Carlton and Randal Picker explain, antitrust and regulation reflect different institutional competencies.[134] Antitrust promotes competition through generally applicable principles without prescribing prices or mandatory dealing terms. Regulation, by contrast, can set prices and mandate access conditions, but only where supported by the ongoing administrative capacity and sector-specific expertise regulatory agencies are designed to provide. Where markets are not characterized by structural failures, such as natural monopoly or insurmountable barriers to entry, antitrust is generally preferable precisely because it avoids the costs and distortions of prescriptive regulation.

The CADE settlement blurs that distinction. By codifying specific fee percentages—including the 15% steering fee and 5% core technology commission—it performs a price-regulation function without the institutional infrastructure that price regulation requires: continuous monitoring, periodic adjustment, cost analysis, and revision as market conditions evolve.[135] The more detailed and prescriptive an antitrust remedy becomes, the greater the risk of regulatory ossification and unintended consequences—especially in digital markets, where business models evolve faster than regulatory frameworks.

None of this suggests CADE’s remedial objectives were illegitimate. If the competitive concerns identified in the proceedings were well founded, meaningful relief might indeed require substantial structural or behavioral intervention. But the path from diagnosis to remedy matters. The Tribunal adopted a contestable market definition, relied on a theory of harm without clear limiting principles, and applied a relaxed evidentiary standard. On that foundation, it constructed a remedial framework that ultimately produced one of the most prescriptive antitrust settlements in global digital-platform enforcement.

The mismatch between the rigor of the analysis and the ambition of the remedy should concern anyone interested in the institutional integrity of competition enforcement.

VII.                    Conclusion

CADE’s decision to uphold interim measures against Apple, and the case’s later settlement, mark significant moments in Brazilian competition enforcement. The proceedings show CADE’s willingness to address digital-market concerns and to engage with emerging theories of platform competition. They also show that traditional ex post antitrust tools can produce many of the changes that legislative proposals such as Bill 4,675/2025 and Bill 882/2026 would impose through ex ante regulation.

In that respect, the case weakens the argument for new conduct-specific digital-market regulation. Existing prohibitions on abuse of dominance proved capable of reaching the alleged harms involving steering, in-app payments, and app distribution. The interim measures and final settlement also confirm that Brazil’s competition regime has tools to prevent market harm when the evidence and circumstances warrant intervention.

But the case’s importance should not be confused with analytical soundness. The Decision rests on a single-brand market definition for iOS app distribution that is difficult to reconcile with established antitrust methodology, even though some decisions in the United States and, especially, the European Union have adopted similar approaches. The Tribunal’s ecosystem theory of harm may offer useful descriptive vocabulary, but it has not shown that it adds meaningful analytical value beyond traditional doctrines of tying, leveraging, and foreclosure.

The characterization of CADE’s decision as the “first true ecosystem theory of harm” may be accurate as a matter of terminology. It may be the first case in which a competition authority explicitly used the language of the academic ecosystem literature as the organizing framework for its analysis. But if the underlying economics are the same—if the real issues remain tying, leveraging, and foreclosure—then the innovation is rhetorical rather than substantive. The Brazilian experience thus appears to confirm what the European experience suggests: ecosystem theory has so far added little to antitrust analysis, and where it produces different outcomes, it often does so by being misunderstood or misapplied rather than by expanding the analytical toolkit.

Analytical frameworks matter. Competition enforcement protects consumers when it rests on rigorous economic analysis, properly defined markets, and falsifiable theories of harm. As applied here, the ecosystem framework does not yet meet that standard. It exaggerates Apple’s market power, gives too little weight to legitimate business justifications for vertical integration, and contributes to an unwarranted finding of anticompetitive conduct.

The deeper lesson is institutional. Brazil’s competition framework is flexible enough to address digital-market disputes without abandoning the consumer welfare standard or importing regulatory presumptions through antitrust procedure. That flexibility is a strength only if paired with discipline. Otherwise, ex post enforcement risks becoming sectoral regulation by another name.

Developers may understandably want a larger share of the ecosystem’s returns. But dividing platform rents is not—and should not become—the purpose of antitrust law.

[1] Projeto de Lei No. 4.675, de 2025, Câmara dos Deputados (Braz.), https://www.camara.leg.br/proposicoesWeb/fichadetramitacao?idProposicao=2562481.

[2] Projeto de Lei No. 882/2026, Câmara dos Deputados (introduced Mar. 3, 2026) (Braz.), https://www.camara.leg.br/proposicoesWeb/fichadetramitacao?idProposicao=2606171. The bill would amend Law No. 12,529/2011 and Law No. 8,078/1990 to classify restricting a competitor’s access to an application-distribution system as an abuse-of-dominance offense. As of May 2026, the bill remained pending before the Consumer Defense Commission. Deputy Clodoaldo Magalhães (PV-PE) was designated rapporteur on Apr. 17, 2026, and the amendment period closed May 4, 2026, with no amendments filed.

[3] The description of the facts and procedural history below derives, unless otherwise noted, from the English-language version of the CADE Tribunal’s May 14, 2025 decision upholding the interim measure issued by the agency’s General Superintendency [hereinafter Decision]. The full English-language decision is available at https://www.gov.br/cade/en/matters/news/cade-upholds-interim-measure-against-apple (last visited Apr. 4, 2026). The case file is available in Portuguese at https://sei.cade.gov.br/sei/modulos/pesquisa/md_pesq_processo_exibir.php?1MQnTNkPQ_sX_bghfgNtnzTLgP9Ehbk5UOJvmzyesnbE-Rf6Pd6hBcedDS_xdwMQMK6_PgwPd2GFLljH0OLyFUsJpu5C1ProV1zLrv9dPsgwllnq3_Q7zJg7Cr5_JvAc.

[4] CADE’s Tribunal appeared to acknowledge the dispute’s redistributive character in observing that Apple’s alleged conduct could be “exploitative”: “While orchestration is crucial for coordination in the ecosystem, design and governance decisions can generate functional and distributive failures, resulting in a loss of value for all participants and hindering innovation. The framing of these decisions under competition laws blurs the boundaries between the exclusionary and exploitative nature of abusive conduct.” Decision, supra note 2, ¶ 50.

[5] See Commission Decision C(2024) 1307 final of 4 March 2024 relating to a proceeding under Article 102 of the Treaty on the Functioning of the European Union and Article 54 of the EEA Agreement (AT.40437 – Apple – App Store Practices (music streaming)), 2024 O.J. (C/2024/3556), https://ec.europa.eu/competition/antitrust/cases1/202419/AT_40437_10026012_3547_4.pdf [hereinafter Eur. Comm’n Apple Music Streaming Decision].

[6] See Epic Games, Inc. v. Apple Inc., 559 F. Supp. 3d 898 (N.D. Cal. 2021) (Gonzalez Rogers, J.), https://www.courtlistener.com/docket/17442392/epic-games-inc-v-apple-inc.

[7] SG Order No. 24/2024, SEI 1476083 (Nov. 25, 2024), https://sei.cade.gov.br/sei/modulos/pesquisa/md_pesq_documento_consulta_externa.php?HJ7F4wnIPj2Y8B7Bj80h1lskjh7ohC8yMfhLoDBLddbbVtapRDlsIRSGO8EfHGTjlAvVixwsF8RWAxY7K1lOCwlw9mGBV9xj89bT7n1aVU7T0oE2YGB0rHOHC39OrFWc. The order adopted the reasoning set out in a technical note issued by the SG staff. See Technical Note No. 63/2024/CGAA11/SGA1/SG/CADE, SEI 1475850 (Nov. 25, 2024), https://sei.cade.gov.br/sei/modulos/pesquisa/md_pesq_documento_consulta_externa.php?HJ7F4wnIPj2Y8B7Bj80h1lskjh7ohC8yMfhLoDBLddZGjmDYkx3_EXIVLLLrA_C3ojklC750gYvLk4Wjzp2CQAzNjE5yiDgT6lb0_1xdyihsVVs3J1xFcXVJMWUJOcf9.

[8] See Commission Decision C(2018) 4761 final of 18 July 2018 relating to a proceeding under Article 102 of the Treaty on the Functioning of the European Union and Article 54 of the EEA Agreement (AT.40099 – Google Android), https://ec.europa.eu/competition/antitrust/cases/dec_docs/40099/40099_9993_3.pdf [hereinafter Eur. Comm’n Google Android Decision].

[9] See Eur. Comm’n Apple Music Streaming Decision, supra note 5.

[10] See Epic Games v. Apple, supra note 6.

[11] See Epic Games, Inc. v. Google LLC, No. 3:20-cv-05671-JD (N.D. Cal. Dec. 11, 2023) (Donato, J.), https://www.courtlistener.com/docket/17443962/epic-games-inc-v-google-llc.

[12] 14th Civil Court of the Judicial Section of the Federal District, Civil Writ of Mandamus No. 1097967-08.2024.4.01.3400, (Dec. 4, 2024)

[13] CADE, Press Release, CADE Discusses Competition in Mobile Ecosystems (Feb. 5, 2025), https://www.gov.br/cade/en/matters/news/cade-discusses-competition-in-mobile-ecosystems-1. A video of the hearing is available at CADE, Audiência Pública: Aspectos Concorrenciais dos Ecossistemas Digitais de Sistemas Operacionais Móveis (Feb. 19, 2025), https://www.youtube.com/watch?v=ehpGbkb8reA.

[14] CADE also requested written submissions. ICLE’s submission is available at Geoffrey A. Manne, Dirk Auer & Mario A. Zúñiga, ICLE Comments to Brazil’s CADE on Competition in Digital Ecosystems of Mobile Devices (Int’l Ctr. for L. & Econ., Feb. 2025), https://laweconcenter.org/resources/icle-comments-to-brazils-cade-on-competition-in-digital-ecosystems-of-mobile-devices.

[15] Decision, supra note 3.

[16] Marcela Mattiuzzo & Barbara De Sa Neves, Beyond Anti-Steering: Brazil’s Landmark Case Against Anticompetitive Conduct in Apple’s App Store, ABA Antitrust Newsletter (Oct. 1, 2025), https://www.americanbar.org/groups/antitrust_law/resources/newsletters/beyond-anti-steering.

[17] See Eur. Comm’n Google Android Decision, supra note 8.

[18] See Section III infra, particularly notes 61-63 and accompanying text.

[19] Decision, supra note 3, ¶ 374.

[20] See Bruno Carballa-Smichowski, Brazil’s Competition Authority Introduced the First True Ecosystem Theory of Harm, Kluwer Competition L. Blog (June 30, 2025), https://competitionlawblog.kluwercompetitionlaw.com/2025/06/30/brazils-competition-authority-introduced-the-first-true-ecosystem-theory-of-harm.

[21] CADE, Appeal No. 08700.009932/2024-18, Concurring Opinion of Commissioner Gustavo Augusto Freitas de Lima (May 2025), https://sei.cade.gov.br/sei/modulos/pesquisa/md_pesq_documento_consulta_externa.php?HJ7F4wnIPj2Y8B7Bj80h1lskjh7ohC8yMfhLoDBLdda_nlCzwUp7VFa9nLtVGuLOhLXyYKIuYhgQTUoVw8sFn9ghwmMNdUEzhJUqKTiW4ZBpKBg8O6ocATHg0QfSgGGn [hereinafter Freitas Concurring Opinion].

[22] Id. ¶ 7.

[23] Id. ¶ 14.

[24] CADE, Press Release, SG Recommends Apple’s Conviction for Anticompetitive Conduct (Jul. 7, 2025), https://www.gov.br/cade/en/matters/news/sg-recommends-apple2019s-conviction-for-anticompetitive-conduct. The SG’s technical note largely follows the Decision, although it does not rely on the “ecosystem theory of harm.” See Technical Note No. 51/2025/CGAA11/SGA1/SG/CADE, Administrative Proceeding No. 08700.009531/2022-04, SEI 1583153 (June 30, 2025), https://sei.cade.gov.br/sei/modulos/pesquisa/md_pesq_documento_consulta_externa.php?HJ7F4wnIPj2Y8B7Bj80h1lskjh7ohC8yMfhLoDBLddY0IIJxEsPG8Jm55kZvdBeZPpUBM3OXRcEQOClOvvDdT0EyPZqE632Y4bjFovY5rBb1Pw-jErxa5dpYf8mlN7XL.

[25] CADE, Cease and Desist Agreement, Administrative Proceeding No. 08700.007177/2022-04 (Dec. 23, 2025), https://sei.cade.gov.br/sei/modulos/pesquisa/md_pesq_documento_consulta_externa.php?HJ7F4wnIPj2Y8B7Bj80h1lskjh7ohC8yMfhLoDBLddagMC0PGtHHNJ5OLjWmXqqKoNDHq57WbWDJKa4eAKK0enGmrJphBHiCJbWXKXvI8K_s-573DMQSRm2YO5d-Z_Bw.

[26] See Dario Oliveira Neto & Mario Zúñiga, Apple in Brazil: Ex Post Antitrust Meets Ex Ante Ambitions, Truth on the Mkt. (Feb. 4, 2026), https://truthonthemarket.com/2026/02/04/apple-in-brazil-ex-post-antitrust-meets-ex-ante-ambitions.

[27] Giuseppe Colangelo, Do Ecosystems Exist in EU Competition Law? 2-4 (Nov. 6, 2025) (forthcoming in Eur. L. Rev.), https://ssrn.com/abstract=5408647.

[28] Id. at 9-12.

[29] Id. at 6.

[30] Id. at 7.

[31] Christopher S. Yoo, Ecosystems in Competition Law: A U.S. Perspective, Network L. Rev. (Spring 2025), https://www.networklawreview.org/yoo-ecosystems.

[32] See generally Francine Lafontaine & Margaret E. Slade, Transaction Cost Economics and Vertical Market Restrictions—Evidence, 55 Antitrust Bull. 587 (2010). (“… the empirical evidence is consistent and convincing. Combined with evidence on where VRs are found, as described above, we conclude that the empirical literature supports Williamson’s contention that ‘[c]ontrary to the inhospitability tradition, contractual constraints can and often do serve legitimate economic purposes. Specifically, vertical constraints may be needed lest subgoal pursuit by the individual parts destroy the viability of the system.’ The logical implication, as he suggests, is not that their use should never be questioned, but that the presumption should not be that they are detrimental to consumers.”) Id. at 608-09 (quoting Oliver E. Williamson, Assessing Vertical Market Restrictions: Antitrust Ramifications of the Transaction Cost Approach, 127 U. Pa. L. Rev. 953, 992 (1979)).

[33] See, e.g., Michael G. Jacobides & Ioannis Lianos, Ecosystems and Competition Law in Theory and Practice, 30 Indus. & Corp. Change 1199, 1209 (Oct. 2021), https://academic.oup.com/icc/article/30/5/1199/6428760.

[34] Yoo, supra note 31.

[35] Org. for Econ. Cooperation & Dev. (OECD), Efficiency Analysis in Vertical Restraints—Background Note by the Secretariat 9 (Latin Am. & Caribbean Competition F., Sept. 9, 2021), https://www.oecd.org/content/dam/oecd/en/publications/reports/2021/09/efficiency-analysis-in-vertical-restraints_7fe5fb55/7108965c-en.pdf.

[36] See, e.g., Copenhagen Econ., The Economic Rationale for Vertical Integration in the Tech Sector (2021), https://copenhageneconomics.com/wp-content/uploads/2021/12/copenhagen-economics-the-economic-rationale-for-vertical-integration-in-tech.pdf; Mario Zúñiga, Ms Vestager: Do Not Tear Down This Wall, Truth on the Mkt. (Aug. 5, 2024), https://truthonthemarket.com/2024/08/05/ms-vestager-do-not-tear-down-this-wall.

[37] Interim measures are undoubtedly a valuable antitrust-enforcement tool in digital markets, given both the typical length of antitrust proceedings and the speed at which such markets evolve. Even so, authorities should apply them only when there is a significant degree of certainty, both to ensure procedural fairness and to minimize the risk of costly enforcement errors, particularly in abuse-of-dominance cases.

[38] At least as a matter of policy. This brief does not address whether the standard applied is consistent with CADE precedent or Brazilian law more generally.

[39] Decision, supra note 3, ¶ 45.

[40] Id. ¶ 46.

[41] Id. ¶¶ 49-50.

[42] Colangelo, supra note 27, at 6.

[43] See United States v. Google LLC, No. 1:23-cv-00108, slip op. (E.D. Va. Apr. 17, 2025).

[44] Ohio v. Am. Express Co., 585 U.S. 529, 541-45 (2018) (holding that multisided transaction platforms must be evaluated as a whole, rather than by examining one side in isolation). See also David S. Evans & Richard Schmalensee, The Antitrust Analysis of Multi-Sided Platform Businesses 18 (Coase-Sandor Inst. for L. & Econ. Working Paper No. 623, 2012).

[45] See Lafontaine & Slade, supra note 32, and accompanying text.

[46] See, e.g., Eur. Comm’n, Decision of 18 July 2018, Case AT.40099—Google Android ¶ 557 (acknowledging multihoming, while concluding that “it [is] not sufficient to make Google Android and iOS devices homogenous”).

[47] Decision, supra note 3, ¶ 197.

[48] Apple’s system is relatively closed. The original iPhone employed a fully closed model: it lacked both an application store and any means to install additional applications (then commonly referred to as “programs”). Apple later relaxed that approach. In June 2007, it allowed Web 2.0 applications that looked and behaved like native iPhone applications, and in 2008 it launched the App Store. With the App Store, the iPhone effectively became a platform connecting consumers and application developers. Apple sells devices, software, and software updates to consumers, while selling developers access to users.

[49] Craig Conrath, Practical Handbook of Antimonopoly Law Enforcement for an Economy in Transition § 5, at 5 (1995).

[50] Decision, supra note 3, ¶¶ 100-139.

[51] Id. ¶¶ 102-106.

[52] Id. ¶ 104.

[53] Cal. Bus. & Prof. Code §§ 17200-17210.

[54] Epic Games v. Apple, supra 6, at 164, quoted in Decision, supra note 3, ¶ 104.

[55] Id. at 164.

[56] United Farmers Agents Ass’n v. Farmers Ins. Exch., 89 F.3d 832 (5th Cir. 1996) (table).

[57] Epic Games v. Apple, supra note 6, at 128.

[58] See Conrath, supra note 49.

[59] Benjamin Klein, Market Power in Antitrust: Economic Analysis After Kodak, 3 Sup. Ct. Econ. Rev. 53 (1993).

[60] Decision, supra note 3, ¶ 139.

[61] Id. ¶¶ 5.3.3 (analyzing anticompetitive discrimination), 5.3.4 (analyzing tying of the App Store and IAP).

[62] Id. ¶ 205.

[63] Epic Games v. Apple, supra note 6, at 46.

[64] David Evans, Multisided Platforms, Dynamic Competition, and the Assessment of Market Power for Internet-Based Firms 12 (Univ. of Chi. Coase-Sandor Inst. for L. & Econ. Research Paper No. 753, Mar. 10, 2016), https://ssrn.com/abstract=2746095.

[65] Geoffrey A. Manne, Dirk Auer & Mario Zúñiga, Comments of the International Center for Law & Economics on CMA’s Proposal to Designate Apple and Google with Strategic Market Status 2 (Aug. 20, 2025), https://laweconcenter.org/wp-content/uploads/2025/08/ICLE-CMA-Apple-Google-Designation-comments.pdf.

[66] “It’s now possible, if you’ve got its latest Pixel or Samsung phones, to ask the Gemini app to order an Uber or a meal on DoorDash. I tried it today and it worked flawlessly, if slowly. (I ordered paper towels via DoorDash, and once Gemini had arranged the transaction, it sent me to DoorDash to click the order button). OpenAI’s ChatGPT can’t yet offer quite the same thing. Score one for Google. This feature—what it calls task automations—is an example of an AI agent actually fulfilling its promise. Notably, Google hasn’t made a big splash about this so far, other than this low-key announcement last month.” Martin Peers, Google’s Gemini Steals a March on OpenAI, The Information (Mar. 23, 2026), https://www.theinformation.com/newsletters/the-briefing/googles-gemini-steals-march-openai. With this product, Google competes primarily with ChatGPT, Anthropic, and other artificial-intelligence services. It also competes with Apple by offering Android devices an at-least-temporarily exclusive feature.

[67] Decision, supra note 3, ¶ 210.

[68] See Statcounter Glob. Stats, Mobile Operating System Market Share Brazil (Feb. 2025-Feb. 2026), https://gs.statcounter.com/os-market-share/mobile/brazil (last visited Mar. 29, 2026). The Decision specifically notes that Apple’s market share is below 20%, which cuts against a presumption of dominance.

[69] Decision, supra note 3, ¶¶ 211-212.

[70] See IDC: Smartphone Shipments Up 2.3% in Q4 2025, Apple Is the Clear Winner, GSMArena (Jan. 14, 2026), https://www.gsmarena.com/idc_smartphone_shipments_up_23_in_q4_2025_apple_is_the_clear_winner_-news-71096.php (reporting global smartphone shipments of 1.26 billion units, with Apple accounting for 24.2%, Samsung for 18.2%, and Xiaomi for 11.2%); see also Malcolm Owen, Apple Dominated 2025 Smartphone Market with a 20% Share, AppleInsider (Jan. 12, 2026), https://appleinsider.com/articles/26/01/12/apple-dominated-2025-smartphone-market-with-a-20-share.

[71] Apple and Google, of course, compete differently with the aforementioned device manufacturers because most manufacturers use Google’s Android—or a forked version of Android—as their operating system.

[72] Decision, supra note 3, ¶ 212.

[73] See Gregory J. Werden, Market Delineation and the Justice Department’s Merger Guidelines, 1983 Duke L.J. 514 (1983) (explaining that substitution patterns, rather than surface-level product similarities or differences, determine market boundaries under the hypothetical-monopolist test, and that low cross-elasticity of demand between different tiers of the same product may justify separate markets).

[74] Counterpoint Rsch., Global Premium Smartphone Share Climbs to 25% in 2024 as Premiumization Continues (Feb. 17, 2025), https://counterpointresearch.com/en/insights/post-insight-research-notes-blogs-global-premium-smartphone-share-climbs-to-25-in-2024-as-premiumization-continues.

[75] Mobile World Live, Premium Smartphone Market Hits Record in H1 (Sept. 8, 2025), https://www.mobileworldlive.com/devices/premium-smartphone-market-hits-record-in-h1.

[76] Malcolm Owen, Apple Dominated 2025 Smartphone Market with a 20% Share, AppleInsider (Jan. 12, 2026), https://appleinsider.com/articles/26/01/12/apple-dominated-2025-smartphone-market-with-a-20-share.

[77] Mobile World Live, supra note 75.

[78] From US$1,199 to US$1,299 and above, depending on specifications. See Aleksander Anastasov, Samsung Galaxy S25 Ultra vs Google Pixel 9 Pro XL: A Friendly but Heated Rivalry, PhoneArena (Mar. 26, 2025), https://www.phonearena.com/reviews/galaxy-s25-ultra-vs-pixel-9-pro-xl_id6440; Google Store (U.S.), Pixel 10 Pro, https://store.google.com/us/config/pixel_10_pro?hl=en-US&selections=eyJwcm9kdWN0RmFtaWx5IjoiY0dsNFpXeGZNVEJmY0hKdiJ9 (last visited Apr. 1, 2026); Apple Store, Buy iPhone 17 Pro, https://www.apple.com/shop/buy-iphone/iphone-17-pro/6.9-inch-display-256gb-silver-unlocked (last visited Apr. 1, 2026); Samsung Store, Galaxy S26 Ultra, https://www.samsung.com/us/smartphones/galaxy-s26-ultra/buy/galaxy-s26-ultra-1tb-unlocked-sku-sm-s948uzvfxaa (last visited Apr. 1, 2026). Prices in Brazil are often higher, and the price gap between the iPhone and Samsung’s premium models is somewhat larger. Vivo offers the Galaxy S26+ at R$9,899 (approximately US$1,920), https://store.vivo.com.br/lancamento-s26/c (last visited Apr. 1, 2026), while the iPhone 17 Pro starts at R$11,499 (approximately US$2,230), https://www.apple.com/br/shop/buy-iphone/iphone-17-pro. The premium-smartphone segment, however, continues to grow. See Daniela Braun, Premium Smartphones Gain Ground in Brazil as Unit Sales Decline, Valor Int’l (May 6, 2025), https://valorinternational.globo.com/business/news/2025/06/05/premium-smartphones-gain-ground-in-brazil-as-unit-sales-decline.ghtml (“In the world’s fourth-largest smartphone market, manufacturers are betting that Brazilian consumers are willing to spend more for advanced devices with premium features and eye-catching design.”). Samsung, moreover, increasingly relies on AI features to compete in the premium-smartphone segment. See João Luiz Rosa, Samsung Doubles Down on AI in Brazil’s Crowded Smartphone Market, Valor Int’l (Feb. 26, 2026), https://valorinternational.globo.com/business/news/2026/02/26/samsung-doubles-down-on-ai-in-brazils-crowded-smartphone-market.ghtml (“The main appeal of the Galaxy S26 series is its advanced integration of artificial intelligence.”).

[79] Counterpoint Rsch., supra note 74.

[80] Decision, supra note 3, ¶ 214.

[81] Simon Baker & Laurence Wu, Applying the Market Definition Guidelines of the European Commission, NERA Topics 21 (Feb. 1998), https://www.cea.fi/course/material/NERA.pdf.

[82] Geoffrey A. Manne, Dirk Auer & Mario A. Zúñiga, Comments of the International Center for Law & Economics: Competition in Digital Ecosystems of Mobile Devices (iOS and Android) 4 (Int’l Ctr. for L. & Econ., Feb. 12, 2025), https://laweconcenter.org/resources/icle-comments-to-brazils-cade-on-competition-in-digital-ecosystems-of-mobile-devices.

[83] Dirk Auer, Making Sense of the Google Android Decision 20 (Int’l Ctr. for L. & Econ., Feb. 25, 2020), https://laweconcenter.org/wp-content/uploads/2020/02/Auer-Making-Sense-of-the-Google-Android-Decision-White-Paper.pdf (the Commission’s decision presaged CADE’s position that Apple holds a monopoly over nonlicensable mobile operating systems).

[84] See Accent, Consumer Purchasing Behaviour in the UK Smartphone Market for the CMA’s Mobile Ecosystems Market Study: Final Report iii (June 2022), https://assets.publishing.service.gov.uk/media/62a1cb0b8fa8f50395c0a0e7/Consumer_purchasing_behaviour_in_the_UK_smartphone_market_-_CMA_research_report_new.pdf.

[85] David S. Evans & Michael D. Noel, Defining Antitrust Markets When Firms Operate Two-Sided Platforms, 2005 Colum. Bus. L. Rev. 667, 681-90 (2005); see also Herbert Hovenkamp, Antitrust and Platform Monopoly, 130 Yale L.J. 1952, 1962 (2021).

[86] Decision, supra note 3, ¶ 213.

[87] Id. ¶ 255.

[88] Id. ¶ 271.

[89] See Freitas Concurring Opinion, supra note 21.

[90] Epic Games v. Apple, supra note 6, at 132 (“As a result, neither consumers nor developers are ‘locked-in’ to the App Store for digital mobile game transactions—they can and do pursue game transactions on a variety of other mobile platforms and increasingly other game platforms. Although the state of the wider gaming market is not at a level where the entirety of these gaming platforms can truly be characterized as competing for purposes of antitrust law (e.g., substitutes), the continued rise of cross-platform games, technologies, and innovative ways in which to reach consumers only demonstrate that these differing platforms are converging and ever intertwining. In sum, with seasoned antitrust counsel at the helm, Epic Games created a market definition which theoretically made a strong showing within the Newcal and Eastman Kodak framework. For the reasons explained above, the market definition was fundamentally flawed, and in any event, does not satisfy all four of the Newcal factors.”).

[91] Complaint ¶¶ 165-181, United States v. Apple Inc., No. 2:24-cv-04055 (D.N.J. Mar. 21, 2024).

[92] Apple advanced this argument, but the SG rejected it “because (…) the relevant market in which the conduct originated was defined as the non-licensable smart mobile operating system iOS, in which Apple is a monopolist.” Decision, supra note 3, ¶ 254.

[93] Decision, supra note 3, ¶¶ 317, 363.

[94] Id. ¶¶ 298-322.

[95] Id. ¶ 323.

[96] Id. ¶¶ 333-350. The Tribunal’s conclusion that Apple’s IAP system and app-distribution platform constitute “separate products” for purposes of tying doctrine rests primarily on their technical separability—i.e., that app distribution can exist without mandatory in-app payment processing. The Decision quotes the U.S. Supreme Court’s statement in Jefferson Parish that whether two items are separate products “turns not on the functional relation between them, but rather on the character of demand for the two items.” Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 19 (1984). The Court also explained, however, that “[t]he requirement that two distinguishable product markets be involved follows from the underlying rationale of the rule against tying.” Id. at 20. In other words, the relevant question is whether the arrangement forecloses competition in a distinct product market in which separate provision would be efficient. Here, it has not been established that it would be efficient for Apple to offer its IAP separately, particularly given that Apple does not charge developers merely to distribute applications through iOS; commissions apply only when developers sell digital goods or services through the platform.

[97] Id. ¶ 104.

[98] Id. ¶ 364.

[99] Id. ¶ 326.

[100] Apple Inc., Addressing Spotify’s Claims (Mar. 15, 2019), https://www.apple.com/newsroom/2019/03/addressing-spotifys-claims (“A full 84 percent of the apps in the App Store pay nothing to Apple when you download or use the app.”).

[101] Epic Games v. Apple, supra note 6, at 105-10 (finding cognizable procompetitive justifications for Apple’s restrictions based on security and privacy).

[102] Copenhagen Econ., supra note 36, at 42-47 (finding that vertical integration in digital markets is associated with efficiency gains and enhanced consumer welfare).

[103] Randal C. Picker, Security Competition and App Stores, Network L. Rev. (Aug. 2021), https://www.networklawreview.org/picker-app-stores (arguing that open-distribution mandates risk exposing iPhone users to malware levels comparable to Android, where sideloading is permitted).

[104] Decision, supra note 3, ¶¶ 58-59 (contrasting Apple’s and Google’s business models and ecosystem strategies).

[105] Id. ¶¶ 132, 348, 389 (expressly citing the comparison as a benchmark).

[106] Apple, Building a Trusted Ecosystem for Millions of Apps: A Threat Analysis of Sideloading (Oct. 2021), https://www.apple.com/privacy/docs/Building_a_Trusted_Ecosystem_for_Millions_of_Apps_A_Threat_Analysis_of_Sideloading.pdf (reporting that Android devices permitting sideloading experience 15 to 47 times more malware infections than iPhones).

[107] Decision, supra note 3, ¶¶ 63-68 (discussing sideloading risks and the debate over whether security concerns justify exclusive app-store distribution).

[108] Heather West, Europe’s DMA: A Cybercriminal’s Paradise?, Ctr. for Eur. Pol’y Analysis (Nov. 13, 2025), https://cepa.org/article/europes-dma-a-cybercriminals-paradise (analyzing security vulnerabilities created by forced interoperability in messaging and payment systems).

[109] Picker, supra note 103.

[110] Sven Völcker & Daniel Baker, Why There Is No Antitrust Case Against Apple’s App Store: A Response to Geradin & Katsifis 23-24 (SSRN Working Paper, 2020), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3660896 (arguing that Apple’s commercial model creates procompetitive incentives for platform investment).

[111] Mikolaj Barczentewicz, The Digital Markets Act as an EU Digital Tax: When Compliance Costs Dwarf Regulatory Estimates, Truth on the Mkt. (July 8, 2025), https://truthonthemarket.com/2025/07/08/the-digital-markets-act-as-an-eu-digital-tax-when-compliance-costs-dwarf-regulatory-estimates (documenting how DMA compliance diverts resources from innovation toward regulatory negotiation).

[112] Act No. 12.529 (Nov. 30, 2011), arts. 36, 88, § 6. An English-language version is available at CADE, https://cdn.cade.gov.br/portal-ingles/topics/leniency%20program/Applicable%20Laws/law-no-12529-2011-english-version-from-18-05-2012.pdf; see also Dario Oliveira Neto, Will Brazil Subtly Sweep Consumer Welfare Under the Rug?, Truth on the Mkt. (Oct. 28, 2025), https://truthonthemarket.com/2025/10/28/will-brazil-subtly-sweep-consumer-welfare-under-the-rug.

[113] CADE, Guide for Horizontal Merger Review 7-8 (2016), https://cdn.cade.gov.br/Portal/centrais-de-conteudo/publicacoes/guias-do-cade/GUIDE%20FOR%20HORIZONTAL%20MERGER%20REVIEW.pdf (referring to consumer welfare as a criterion for evaluating the potential anticompetitive effects of both unilateral and coordinated conduct in horizontal mergers); CADE, Guidelines for the Analysis of Non-Horizontal Mergers 7-8 (2024), https://cdn.cade.gov.br/Portal/centrais-de-conteudo/publicacoes/guias-do-cade/V%2B%20Guide%20in%20English%20-%20Final%20version%202.pdf (requiring analysis of harm to final consumers when evaluating potential competitive harm from vertical mergers).

[114] Org. for Econ. Cooperation & Dev. (OECD), Advantages and Disadvantages of Competition Welfare Standards—Note by Brazil, OECD Doc. DAF/COMP/WD(2023)27, at 2-5 (June 15, 2023), https://one.oecd.org/document/DAF/COMP/WD%282023%2927/en/pdf (“stating that consumer welfare drives antitrust policy in Brazil”); see also Oliveira Neto, supra note 112.

[115] See infra Sections III-IV (analyzing the Tribunal’s market definition and theory of harm).

[116] Frank H. Easterbrook, The Limits of Antitrust, 63 Tex. L. Rev. 1, 2-3 (1984) (“If the court errs by condemning a beneficial practice, the benefits may be lost for good. Any other firm that uses the practice faces sanctions in the name of stare decisis, no matter the benefits. If the court errs by permitting a deleterious practice, though, the welfare loss decreases over time. Monopoly is self-destructive. Monopoly prices eventually attract entry.”).

[117] Joshua D. Wright & Murat C. Mungan, The Easterbrook Theorem: An Application to Digital Markets, 130 Yale L.J.F. 622, 636-43 (2021).

[118] Geoffrey A. Manne, Error Costs in Digital Markets, in Report on the Digital Economy 33, 35 (Glob. Antitrust Inst. ed., 2020), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3733662.

[119] See supra note 1. For a critical analysis of the bill, see Geoffrey A. Manne, Dario Oliveira Neto & Dirk Auer, Digital Overreach: A Premature Turn to Ex Ante Regulation in Brazil (Int’l Ctr. for L. & Econ. Apr. 7, 2026), https://laweconcenter.org/resources/digital-overreach-a-premature-turn-to-ex-ante-regulation-in-brazil.

[120] See Victor Oliveira Fernandes, Brazil’s Calibrated Revolution in Digital Competition, ProMarket (Nov. 12, 2025), https://www.promarket.org/2025/11/12/brazils-calibrated-revolution-in-digital-competition.

[121] Oliveira Neto, supra note 112.

[122] Dario Oliveira Neto & Mario Zúñiga, Beyond the Bark: Brazil’s Prudent Path in Digital Regulation, Truth on the Mkt. (Sept. 19, 2025), https://truthonthemarket.com/2025/09/19/beyond-the-bark-brazils-prudent-path-in-digital-regulation.

[123] Decision, supra note 3, ¶ 407.

[124] Epic Games v. Apple, supra note 6.

[125] Regulation (EU) 2022/1925 of the European Parliament and of the Council of 14 Sept. 2022 on Contestable and Fair Markets in the Digital Sector and Amending Directives (EU) 2019/1937 and (EU) 2020/1828 (Digital Markets Act), arts. 5(4)-(5), 6(4), 6(7), 2022 O.J. (L 265) 1, https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32022R1925 [hereinafter Digital Markets Act].

[126] Decision, supra note 3, ¶ 423.

[127] Id. ¶ 433.

[128] CADE, Cease and Desist Agreement, supra note 25; see also Marcus Mendes, Apple Settles Brazilian Antitrust Case with App Store Policy Overhaul, 9to5Mac (Dec. 23, 2025), https://9to5mac.com/2025/12/23/apple-settles-brazilian-antitrust-case-with-app-store-policy-overhaul.

[129] CADE, Cease and Desist Agreement, supra note 25.

[130] Herbert J. Hovenkamp, Structural Antitrust Relief Against Digital Platforms, 7 J.L. & Innovation 1, 47 (2024).

[131] Abbott B. Lipsky, Jr. & J. Gregory Sidak, Essential Facilities, 51 Stan. L. Rev. 1187, 1195 (1999) (“Mandatory access remedies . . . are the stuff of regulatory bodies, not courts.”)

[132] Epic Games, Inc. v. Apple Inc., No. 4:20-cv-05640 (N.D. Cal. Sept. 10, 2021) (permanent injunction), aff’d in relevant part, 67 F.4th 946 (9th Cir. 2023).

[133] Digital Markets Act, supra note 125, arts. 5–7.

[134] Dennis W. Carlton & Randal C. Picker, Antitrust and Regulation (Nat’l Bureau of Econ. Rsch., Working Paper No. 12902, 2007), https://www.nber.org/papers/w12902.

[135] See Oliveira Neto & Zúñiga, supra note 26.