Self-Preferencing Crackdown Transforms Platform Law in US and EU

The European Commission is finalizing a fine against Google that will run into the high hundreds of millions of euros, The Next Web reported on May 26, citing a Handelsblatt report, for alleged breaches of the Digital Markets Act's prohibition on self-preferencing in search results. The penalty, still unannounced as of July 2026, will be the largest levied under the DMA since the regulation took full effect, eclipsing earlier fines and signaling that Brussels views the platform's treatment of its own services in search rankings as a structural violation rather than a technical compliance lapse. The precise figure remains confidential, but the range being discussed inside the Commission places it well above the triple-digit-million-euro threshold, according to the Handelsblatt account. The case turns on Article 6(5) of the DMA, which requires designated gatekeepers to refrain from giving their own products more favorable ranking than similar third-party offerings.

The Google penalty is the sharpest expression to date of a regulatory theory that has, over the past three years, migrated from the margins of competition policy to its dead center: that the architecture of platforms, not merely their conduct toward competitors, constitutes the antitrust problem. The remedy most associated with this theory is the choice screen, a mandated interface that requires a platform to present users with a neutral menu of competing services at a critical decision point: when setting a default search engine, selecting a browser, or choosing a maps provider. Choice screens now sit at the intersection of at least four major enforcement actions and legislative efforts, including the DMA, the long-running Google Shopping competition case, the browser choice screen provisions imposed on Apple, and the American Innovation and Choice Online Act, which Senators Chuck Grassley and Amy Klobuchar reintroduced in the U.S. Senate on June 11, 2026.

The AICOA bill, formally reintroduced as S. 4858, would prohibit dominant platforms from favoring their own products in search results, app stores, and other digital marketplaces. Its reintroduction was immediately met with opposition from Apple, which argued through a statement to Android Headlines that the legislation would "undermine the privacy and security protections that consumers rely on." The bill is narrower than the DMA in scope but broader in remedy: it does not mandate choice screens specifically, but its prohibition on self-preferencing would effectively require platforms to implement some form of neutral intermediation between their own services and those of rivals. The key legislative question is whether that neutrality can be enforced through ex-post litigation under antitrust law or whether it requires the kind of ex-ante regulatory architecture the DMA provides.

The choice-screen remedy draws on a long and uneven regulatory record. The European Commission's first high-profile experiment with the mechanism came in 2009, when it settled an antitrust case against Microsoft by requiring the company to display a "browser ballot" screen to Windows users in the European Economic Area. The screen listed twelve browsers in randomized order and allowed users to select and install one. The remedy was widely criticized as ineffective: Microsoft's Internet Explorer share did decline, but the decline had begun before the ballot was implemented and continued along roughly the same trajectory. Academic researchers who studied the intervention found that it produced at best a marginal acceleration of a trend already underway, driven more by Chrome's technical superiority than by any neutral presentation of options. The episode taught regulators that a badly designed choice screen, one that presents too many options without meaningful differentiation or defaults to paralysis, can function as a compliance artifact rather than a genuine competitive intervention.

The DMA's approach to choice screens reflects some of those lessons but introduces new complications. Article 6(3) requires gatekeeper operating systems to allow users to uninstall pre-installed software applications and to change default settings easily, while Article 6(5) directly prohibits self-preferencing in ranking. The European Commission closed its year-long investigation into Apple's browser choice screen in March 2025, after the company revised the design and functionality of the screen it presents to iOS users in the EU during device setup. The precise changes Apple made were not disclosed in full, but regulatory filings indicate they included expanding the number of browsers displayed, randomizing their order, and providing a brief description of each option drawn from the developer's own App Store listing. The Commission's decision to close the investigation without a fine was taken as a signal that Apple had achieved sufficient compliance, but the standard was behavioral rather than outcome-based: the question was whether the screen met the DMA's design requirements, not whether it shifted market share.

That distinction, between compliance with the architecture of a choice screen and its actual competitive effects, is the central tension running through every self-preferencing case now active. The DMA imposes obligations on gatekeepers but does not set market-share targets, a design choice that makes enforcement administrable but leaves open the question of whether the remedy works. The Google search self-preferencing case illustrates the stakes. Under Article 6(5), Google is required to treat its own vertical search services, for flights, hotels, shopping, and local businesses, no more favorably than competing services from platforms like Kayak, Booking.com, or Yelp. The Commission's view, as described in the draft statement of objections reported by Handelsblatt, is that Google has failed to do so, continuing to display its own rich results, knowledge panels, and carousels in positions that outperform rival aggregators even when those rivals offer objectively more complete or more relevant information for a given query.

The European Commission is preparing the largest fine yet under the Digital Markets Act, targeting Google over self-preferencing in search., The Next Web, reporting on the Handelsblatt account, May 2026

The Google case has a direct lineage to the Commission's 2017 Google Shopping decision, in which the company was fined €2.42 billion for favoring its own comparison-shopping service over rivals in general search results. That decision, upheld by the General Court in 2021, was the foundational self-preferencing precedent under traditional competition law, and it required Google to implement an auction-based remedy in which rival comparison-shopping services could bid for placement in the Shopping Unit displayed atop search results. Critics of the remedy, including several of the rival services it was meant to help, argued that the auction design merely converted Google's organic self-preferencing into a paid self-preferencing mechanism, with Google collecting fees from the very competitors it had excluded. A German court in November 2025 awarded €572 million in damages to a group of rival shopping platforms, finding that Google's conduct had caused them quantifiable harm. The judgment reinforced the view that self-preferencing cases are not merely about future conduct but about past extraction of rents that will take years of follow-on litigation to price.

The U.S. legal landscape on these questions is more fragmented but is converging on the same set of doctrinal problems. The Supreme Court is currently weighing whether to hear Apple's challenge to a contempt finding in the Epic Games v. Apple litigation, a case that revolves around whether Apple's anti-steering provisions, which prevent developers from directing users to external payment systems, constitute a form of self-preferencing that violates the original injunction. Tech Times reported on June 25, 2026, that all nine justices were preparing to vote privately on whether to grant certiorari. The procedural posture is narrow; the implications are not. A ruling that affirms the contempt finding would strengthen the hand of private plaintiffs seeking to use existing antitrust law to challenge self-preferencing, even without the passage of AICOA. A ruling that overturns it would place even greater weight on the legislative effort.

The AICOA bill itself remains in committee, and its prospects are uncertain in a Congress divided between a Republican House and a Democratic Senate. The bill's sponsors have framed it as a competition measure rather than a content-regulation measure, a distinction designed to insulate it from First Amendment challenges of the kind that have complicated state-level platform regulation in Texas and Florida. The key provisions would prohibit a covered platform from giving preference to its own products, services, or lines of business in ways that materially harm competition, and would empower the Federal Trade Commission and the Department of Justice Antitrust Division to bring enforcement actions. The bill does not spell out choice screens as a remedy, but any platform subject to a self-preferencing prohibition faces a practical question: how else, other than through some form of neutral presentation mechanism, can compliance be demonstrated to a court or an agency?

That practical question is what makes the European experience with choice screens directly relevant to U.S. policy development. If the DMA's Article 6(5) enforcement produces demonstrable shifts in user behavior, the U.S. debate will tip toward the view that structural remedies can work without structural separation. If it produces only cosmetic compliance, the argument for bolder remedies, including line-of-business separation or nondiscrimination regimes modeled on common-carriage law, will strengthen. The evidence so far is thin. Browser market-share data for the European Economic Area shows some movement among smaller browsers since Apple implemented its iOS choice screen in early 2024, but the magnitude of the shift is on the order of single-digit percentage points, and it is difficult to disentangle the effect of the choice screen from the general increase in consumer awareness of browser alternatives that accompanied the DMA's implementation more broadly.

Academic researchers tracking the DMA's effects have identified a pattern that is consistent with the Microsoft browser ballot experience: choice screens do produce an initial spike in switching behavior, concentrated among the subset of users who are already inclined to explore alternatives, but the effect decays over time as new device setups revert to familiarity. One working paper circulated by economists at the Toulouse School of Economics in early 2026, drawing on aggregated telemetry data from several European mobile carriers, found that the share of iOS users who selected a non-Safari browser during device setup peaked at roughly 19 percent in the first quarter after the choice screen was introduced and had declined to roughly 14 percent by the end of 2025. The paper has not been peer-reviewed, and the carriers that provided the data did so under nondisclosure agreements that limit independent verification, but the trend line has been cited by both supporters and critics of the DMA as evidence that choice screens are a partial solution at best.

The DMA's next expansion will test the choice-screen model in new market contexts. The European Commission is expected to designate Amazon Web Services and Microsoft Azure as gatekeepers under the DMA before the end of 2026, a finding that would subject cloud infrastructure providers to interoperability and data-portability obligations that are in some respects more demanding than those applied to search engines or app stores. The cloud gatekeeper designation does not trigger a classic choice screen, since cloud services are not consumer-facing in the same way, but it raises an analogous question about what a neutral intermediation obligation looks like when the platform's own products are deeply integrated into the infrastructure layer. A cloud provider that must offer equal treatment to third-party machine-learning services running on its infrastructure faces a different technical problem than a search engine that must rank rival flight aggregators fairly, but the underlying competition-policy logic is the same: the platform controls an input that rivals need, and it has an incentive to degrade that input for competitors while optimizing it for itself.

Italy's competition authority opened a parallel front on June 16, 2026, launching an investigation into Apple's compliance with DMA interoperability obligations as they relate to iCloud and third-party cloud storage services. The Italian probe is notable because it exercises a power the DMA grants to national competition authorities to enforce certain provisions of the regulation within their own jurisdictions, creating the possibility of a multi-speed enforcement landscape in which different member states pursue different theories of the same underlying obligation. For platforms operating across the entire European market, a fragmented enforcement environment may prove a more powerful constraint than any single Commission-level fine, because it multiplies the compliance burden and eliminates the ability to negotiate a single global remedy.

What unifies these disparate enforcement actions is a single unresolved question: when a platform designs the architecture through which users discover, select, and switch between services, can any remedy short of structural separation ever produce genuinely neutral intermediation? The choice screen represents the regulatory answer to that question in its most optimistic form. It assumes that users, given a fair presentation of options, will make choices that discipline platform power. The alternative view, advanced by a growing body of scholarship in both the U.S. and Europe, holds that defaults are sticky for reasons that choice screens cannot fully overcome: cognitive inertia, the transaction costs of evaluating multiple alternatives, and the platform's continuing ability to signal preference through integration, branding, and friction even when a choice screen is formally neutral. Reuters reported that the Italian probe would examine precisely this question of whether Apple's implementation of DMA-mandated data portability tools for iCloud was genuinely neutral or had been designed to impose enough friction to deter switching.

The Commission's forthcoming Google fine will be a test of more than Google's search architecture. It will be a test of whether the DMA's drafters got the calibration right: heavy enough fines to deter noncompliance, fast enough proceedings to avoid the decade-long timelines that plagued competition-law enforcement under Article 102 TFEU, and remedies specific enough to produce measurable market effects. If the fine is announced this summer, as the Handelsblatt report and subsequent Commission statements suggest, the European Court of Justice will have an opportunity to review the DMA's self-preferencing provisions far sooner than most observers expected when the regulation was adopted in 2022. The appeal, which Google has indicated it will file, will likely turn on the question of whether Article 6(5) requires the Commission to prove actual competitive harm or merely to demonstrate that the platform's ranking practices treat its own services more favorably in a way that is capable of producing such harm. The answer to that question will determine how many choice screens regulators can actually impose.

In Washington, the AICOA bill faces a narrower path but a clearer ideological alignment. The bill has bipartisan sponsorship, a rarity in a Congress otherwise fractured along party lines on technology policy, and its proponents argue that U.S. antitrust law needs an explicit self-preferencing prohibition to complement the Sherman Act's general prohibition on monopolization. The House Judiciary Committee's 2020 investigation into competition in digital markets, which produced a 450-page majority staff report, documented extensive evidence of self-preferencing by Google, Apple, Amazon, and Facebook, and recommended legislative action as the only durable remedy. Four years later, the legislative vehicle has been refined but the underlying theory remains unchanged. The question is whether the political window for major antitrust legislation, which seemed wide open in 2022, has narrowed as the Supreme Court's composition has shifted and as the platforms have successfully framed competition regulation as a threat to innovation in artificial intelligence and other emerging technologies.

The self-preferencing debate will not be resolved by any single fine, ruling, or bill. It is a structural argument about the organization of digital markets, and structural arguments take decades to play out. The choice screen is the most visible remedy because it is the most administrable, but its effectiveness depends on details that regulators are still learning how to specify: how many options, in what order, with what descriptions, at what point in the user journey, and with what default if the user declines to choose. The DMA provides a regulatory framework for iterating on those details; AICOA would provide a litigation framework for doing the same in U.S. courts. The European Commission's decision on the Google fine will be read carefully not just for the euro amount but for the design requirements it imposes on search result pages that a billion people use every day. Those design requirements, more than any fine, will shape what the next generation of choice screens actually looks like.