The Grand Chamber of the European Court of Justice’s (CJEU) recent decision in Android Auto marks a pivotal—and possibly final—chapter in the contentious evolution of the essential facility doctrine (EFD) [for a comprehensive analysis of the decision and its implications, see my working paper on “The EU essential facilities doctrine after Android Auto: a wild card without limiting principles?”].
As an exception to the general principle that businesses are free to decide whether to grant access to their facilities, the EFD debate fundamentally revolves around defining its boundaries. At its core, the doctrine seeks to balance fundamental rights with competition policy, as well as short-term and long-term competitive dynamics.
Navigating this delicate trade-off has shaped both the development and the divergent trajectories of the EFD in the U.S. and the EU. While U.S. antitrust law initially embraced the doctrine before ultimately rejecting it, EU courts have upheld it, applying it only in exceptional circumstances. Notably, when these conditions are met, competition concerns take precedence over investment incentives, obliging the owner of a physical or intangible asset to share it with third parties. As originally crafted in Magill, the exceptional circumstances arise when a refusal: (i) concerns a product that is “indispensable” for carrying out the business in question; (ii) lacks “objective justification”; (iii) is likely to eliminate “all competition in a secondary market”; and (iv) prevents the emergence of a “new product” for which there is potential consumer demand.
Over time, however, the interpretation of these circumstances has expanded the doctrine’s reach, effectively reversing the original relationship between the general rule and its exception. In this scenario, although the indispensability criterion remains central to the EFD’s legal framework, CJEU case law has progressively narrowed the situations in which it is strictly required. With Android Auto, this dilution appears to be complete.
From indispensable to convenient: The gradual departure from Bronner
The rationale behind the EFD and the broader EU effort to balance competing interests hinges on determining whether a facility or input is indispensable. Indispensability thus serves as the threshold for distinguishing between a facility essential to competition and one that is merely convenient for competitors, justifying antitrust intervention into a dominant undertaking’s freedom to conduct business. To this aim, in Bronner, the CJEU stated that, to determine whether a product or service is indispensable for an undertaking’s operation in a specific market, it must be assessed whether alternative products or services exist—albeit less advantageous—and whether technical, legal, or economic obstacles would make it impossible or unreasonably difficult for an undertaking to create such alternatives, potentially in collaboration with others. For economic obstacles to be recognized, it must be demonstrated that creating these alternatives is not economically viable on a comparable scale to the undertaking controlling the existing product or service.
However, recent CJEU case law has narrowed the scope of the Bronner conditions, arguing that their imposition was justified by the specific circumstances of that case. Bronner is viewed as a peak in antitrust law, setting a particularly high legal standard for finding an abusive practice. There is concern that broadening the Bronner criteria could undermine the EU Commission’s ability to address abusive practices effectively. As a result, it has been suggested that the Bronner criteria should be limited in scope, with the indispensability requirement not being suitable for determining abuse in all cases involving access to a dominant undertaking’s facility. Therefore, its application should be confined to refusal to deal cases similar to the specific situation in Bronner.
The first departure from Bronner can be traced to Van den Bergh Foods, where the CJEU ruled that the indispensability criterion is only necessary when antitrust intervention would force a dominant undertaking to transfer an asset or enter into agreements with parties it has not chosen to contract with. This was not the case in the decision at hand, as, unlike in Bronner, the assets were not reserved for the undertaking’s exclusive use but were voluntarily made available to independent players who paid for the right to use them.
Following this reasoning, the primary exception to the indispensability criterion came in cases of margin squeeze, which under EU competition law, constitutes a standalone abuse. Specifically, in TeliaSonera, the Court held that the indispensability condition from Bronner did not apply, distinguishing between an outright refusal to deal and a situation where a dominant undertaking grants access to its infrastructure but imposes unfair terms and conditions. In Slovak Telekom, the CJEU extended this margin squeeze treatment to non-price conduct as well. More recently, in Google Shopping, the CJEU rejected the application of the Bronner criteria, arguing that the case concerned the conditions of access to Google’s general search service rather than access to a separate infrastructure, such as the boxes at issue.
In addition, the CJEU in Slovak Telekom and in Lithuanian Railways suggested that enforcers are relieved from proving indispensability when access to the facility is granted due to a regulatory obligation. In such a case, the rationale for departing from Bronner rests on the assumption that the trade-off between the benefits and drawbacks of intervention has already been evaluated within the applicable regulatory framework.
Moreover, in Lithuanian Railways, the CJEU identified two additional exceptions to Bronner, which pertain to cases where the infrastructure in question was financed not by investments specific to the dominant undertaking but by public funds, and situations where a dominant undertaking destroys an infrastructure. These circumstances justify disregarding the indispensability requirement established in Bronner as no significant trade-offs arise when the facility has been destroyed by its owner or was developed using public funds. In both cases, the owner cannot reasonably argue that regulating access terms through competition law would undermine investment incentives, as it either opted to relinquish its asset or did not assume significant entrepreneurial risks in its development. However, even in cases involving public support, the investments made may be relevant as shown by the pending request for a preliminary ruling in LUKOIL Bulgaria.
Finally, a further narrowing of Bronner is again linked to the specific circumstances of that case as it is argued that the indispensability criterion is required only when a dominant undertaking refuses to grant rivals access to an infrastructure it has developed for the needs of its own business. While the reference to “its own use” first appeared in Van den Bergh Foods, the original formulation is found in Advocate General (AG) Jacobs’s opinion in Bronner, where the conduct in question is described as a refusal to allow another newspaper publisher access to a distribution system developed for the purposes of its own newspaper business. The rationale behind this limitation on the application of the indispensability requirement is to protect investment incentives against the risk of free riding: an undertaking’s incentive to invest in facilities it has developed is preserved only if it can exploit those assets exclusively for its own use.
Such an interpretation could have significant implications, particularly in digital markets, as it may suggest that platforms designed to be open to and used by third parties—due to their nature and business model—cannot invoke the lack of indispensability to justify denying access. This line of reasoning was specifically put forward by AG Kokott in Google Shopping and AG Melina in Android Auto.
Android Auto: The Italian proceedings
The case at issue revolves around Google’s refusal to integrate Enel X’s Recharge app (JuicePass) into Android Auto, an infotainment system that brings various Android device features to a car’s dashboard. JuicePass offers services for recharging electric vehicles, such as locating charging stations, managing charging sessions, and reserving charging slots. Therefore, it competes with Google Maps, which provides similar functionalities but lacks reservation and payment services. Google rejected Enel’s request to make JuicePass compatible with Android Auto, arguing that only media and messaging apps were allowed for third-party integrations. Google also cited security concerns and the need to efficiently allocate development resources as reasons for its decision.
However, the Italian Competition Authority (ICA) argued that by obstructing and delaying the availability of JuicePass on Android Auto, Google was attempting to favor its own app, effectively reserving the full range of recharging services for Google Maps. The ICA’s reasoning was based on the fact that Android Auto constitutes a “competitive space,” where service apps compete against the additional functionalities offered by Google’s proprietary navigation app, either directly or potentially. As a result, Google’s actions were seen as a refusal to allow interoperability, violating the principle of a level playing field and granting Google’s app an unfair advantage over Enel X’s. The ICA also determined that Android Auto is an essential facility, despite the fact that drivers can access JuicePass on smartphones via both Google Play and the App Store.
The ICA addressed the indispensability requirement in relation to Android Auto by departing from the definition set by the CJEU in Bronner. According to the decision, the indispensability criterion is satisfied because no alternatives are as convenient and safe as Android Auto, even though less advantageous options could achieve similar outcomes. In the ICA’s view, protecting competition in digital markets requires consideration of their unique characteristics and dynamics. Therefore, to ensure effective competition protection and enhance consumer choice, the legal criteria typically applied in such cases should be used with flexibility.
Due to Google’s gatekeeping position and the conflicts of interest arising from its dual role, the ICA required the company to ensure a fair level playing field for all service apps offering recharge services. Consequently, Google was mandated to develop and maintain a standardized template that would accommodate the needs of third-party recharge applications, enabling their interoperability with Android Auto.
Given the circumstances of the case, the Italian higher administrative court (Council of State) referred a request for a preliminary ruling to the CJEU, seeking clarification on the obligations of dominant players in digital markets. Notably, while most of the questions focus on whether mandating interoperability would require a redesign of the product, the first question (a prerequisite for addressing the technical challenges of mandatory interoperability) concerns the potential adaptation of the EFD to the features of digital markets. Specifically, the referring court asks whether, in a refusal to supply case, it would be sufficient for access to be indispensable “for a more convenient use” of the product or service offered by the undertaking requesting access, particularly when the essential function of the product subject to the refusal is to facilitate and enhance the use of existing products or services.
The CJEU decision
Both the opinion delivered by AG Medina and the decision of the CJEU supported the findings of the ICA. However, while the Italian Council of State’s question concerned the interpretation of the indispensability requirement under Bronner, AG Medina and the Court focused on the applicability of Bronner to the case at hand. In doing so, they confirmed the ongoing trend of limiting the application of the Bronner test to the specific circumstances of that case.
In particular, when evaluating whether a refusal to deal may be considered anticompetitive, the CJEU reaffirmed that preserving freedom of contract, property rights, and long-term incentives to innovate justifies applying Bronner’s indispensability requirement only when the dominant company develops infrastructure for its own business needs and reserves it exclusively for its use. This does not apply when the infrastructure is developed “with a view to enabling third-party undertakings to use it.” In such cases, requiring the company to provide access to third parties “does not fundamentally alter the economic model that applied to the development of that infrastructure.” Consequently, in response to the first question from the referring court, the CJEU held that when a digital platform is designed to be open to third-party undertakings, a refusal to ensure interoperability with a third-party app may constitute an abuse of a dominant position, even if the platform is not indispensable for the commercial operation of that app in the downstream market, but can make the app “more attractive to consumers.”
As a result, except in cases of technical impossibility or harm to the platform’s integrity or security, a dominant platform is obligated to develop (within a reasonable time frame and for appropriate financial compensation) a template to ensure interoperability with third-party apps. The absence of a template for a specific category of apps or the development challenges faced by the dominant undertaking cannot, by itself, serve as an objective justification for refusing to grant access.
The European EFD after Android Auto
In the aftermath of Android Auto, European case law highlights a growing disconnect between the guiding principles of the EFD and its actual application.
Both the U.S. and EU case law have been shaped by similar limiting principles—chiefly, the need for a cautious approach to any duty to deal, which balances competitive concerns with the protection of fundamental rights, as well as the long-term benefits of competition and innovation against the risks of free riding. In the U.S., this balancing act was effectively ruled out by Trinko, making it almost impossible to impose a sharing obligation on a dominant firm. In contrast, the EU’s Magill sought to achieve a fair balance by defining specific exceptional circumstances that would justify such an obligation. Within this framework, the EFD was intended as a last-resort measure, applicable only when those exceptional circumstances were met. Over time, however, the gradual erosion of these criteria—especially the indispensability requirement—has reversed the rule-exception relationship, making the non-application of the EFD the exception rather than the norm. This shift raises doubts about whether the current configuration of the EFD still aligns with its original rationale.
By definitively discarding the Bronner indispensability criterion, Android Auto casts doubt on the very existence of the EFD. Indeed, all the requirements of the exceptional circumstances test established in Magill have been progressively diluted. The only remaining safeguard against rivals’ unfettered access to dominant firms’ facilities is the objective justification criterion. Furthermore, Android Auto leaves us with a deeply divided antitrust landscape. While dominant firms in the U.S. are never compelled to provide access to their infrastructure, the EU has formally embraced the convenient facilities doctrine. In both jurisdictions, however, neither achieves a fair balance between fundamental rights and competition, nor between short-term and long-term competitive benefits.
