EU Article 102 Guidelines

On September 3, 2026, the European Commission adopted its first Guidelines on exclusionary abuses of dominance following a three-year consultation process during which stakeholders were invited to comment on and provide feedback on the draft text. Until now, the Commission’s only guidance in this area was the 2009 Guidance Paper on enforcement priorities, a non-binding statement of enforcement policy.

The Guidelines aim to codify and systematize the evolving case law of the EU Courts on the application of Article 102 of the Treaty on the Functioning of the European Union (TFEU). At the same time, the Commission seeks to address a number of concerns raised in response to the 2024 draft, in particular criticisms that certain aspects of the proposed framework lowered the threshold for intervention or departed from established case law.

While the new Guidelines contain several changes that improve legal certainty, they do not signal a retreat from vigorous Article 102 enforcement. Key developments include the reintroduction of the 40% market-share soft safe harbor, a more nuanced approach to evidentiary presumptions, increased emphasis on digital and AI-driven markets, and further guidance on the rebuttal strategies available to dominant undertakings.*

Key changes at a glance

One of the most notable changes introduced following the consultation process is the reintroduction of the Commission’s 40% market share threshold. While this benchmark appeared in the 2009 Guidance Paper, it was absent from the 2024 draft. The Guidelines now recognize that a market share below 40% is a soft indication – not a safe harbor. Dependence, entry and expansion barriers, limited countervailing buyer power, network effects, switching costs, data and ecosystem reach can still support dominance. A near-dominant business should treat the same indicators as a compliance trigger. This reintroduction of the safe harbor is a welcome development, providing greater legal certainty than the 2024 draft, which left open the possibility of findings of dominance at comparatively low market shares.

Reflecting developments in the EU Courts’ case law and the Commission’s decisional practice, the Guidelines confirm the evolving role of the “As Efficient Competitor” (AEC) test. Traditionally, the AEC test has been used to assess whether conduct is capable of excluding a hypothetical competitor that is as efficient as the dominant undertaking, thus restricting competition. The Guidelines clarify that the AEC test remains primarily relevant to the assessment of pricing abuses, but no longer occupies the position of a generally applicable benchmark for assessing exclusionary conduct. Instead, it is presented as one analytical tool among others, with more limited relevance to non-pricing and multifaceted forms of conduct. Objective justification remains relevant.

The sliding scale of evidence and presumptions

The Commission largely retains the analytical framework set out in the 2024 draft for determining whether conduct by a dominant undertaking distorts effective competition. The assessment continues to rest on a two-step test. First, whether the conduct departs from competition on the merits, and second, whether it is capable of producing exclusionary effects.

Competition on the merits is defined both positively, i.e., as conduct that benefits consumers through lower prices, better quality, wider choice or more innovation, and negatively, i.e., through a non-exhaustive list of departure indicators such as misleading regulators or misusing legal proceedings. Recent discussions surrounding the revised Guidelines highlighted that this dual approach reflects the fact that EU case law has historically focused on identifying departures from the concept rather than defining it directly.

It was also noted that the Guidelines do not dispense with the need for a theory of harm. Rather, while proof of actual consumer harm is not required, the theory of harm and the applicable legal test remain closely linked. This distinction may be particularly relevant for novel forms of conduct, such as self-preferencing, where the body of case law remains relatively limited.

One notable refinement concerns the role of presumptions. While the 2024 draft primarily distinguished between categories of conduct subject to different evidentiary standards, the final Guidelines introduce a more nuanced, sliding-scale approach. Under this framework, the amount of case-specific evidence required from the Commission varies according to the likelihood that the conduct is capable of distorting effective competition. In practice, however, presumptions remain a central feature of the enforcement framework. Once certain factual elements are established, the Commission may rely on a presumption of distortion of effective competition, thereby shifting the evidentiary burden to the dominant undertaking to demonstrate the absence of anti-competitive effects.

Exclusive dealing practices provide the clearest example. The Guidelines consider such practices to have a particularly high potential to generate exclusionary effects and therefore subject them to a presumption of harm. A dominant undertaking may rebut that presumption by demonstrating, for example, that the conduct does not in fact amount to exclusive dealing or that the exclusivity arrangements are too short or insufficiently extensive to produce foreclosure effects.

Pricing abuses are treated similarly. As explained in the Commission’s Q&A, pricing below the relevant cost benchmark is in principle sufficient to establish a distortion of effective competition. Likewise, a margin squeeze resulting in a negative margin for an equally efficient competitor may justify the same conclusion.

The Guidelines also place particular emphasis on conduct that is “by its very nature harmful to competition”. Although the final text abandons the controversial reference to “naked restrictions” contained in the 2024 draft, it preserves the underlying notion that certain forms of conduct are inherently suspect. This is particularly the case where conduct serves no plausible economic purpose other than restricting competition. Such conduct is considered to fall outside the scope of competition on the merits and is therefore unlikely to generate consumer benefits. While the Commission does not exclude the possibility of rebuttal, the practical scope for successfully challenging such a presumption appears limited.

At the same time, the new Guidelines provide considerably more detailed guidance on how dominant undertakings may rebut presumptions of harm. In particular, an undertaking may seek to demonstrate that its conduct generates efficiencies, benefits consumers, is indispensable to achieving those benefits, and does not eliminate effective competition. The Commission also expressly recognizes that, in appropriate circumstances, sustainability benefits, resilience considerations and out-of-market efficiencies may form part of the justification analysis. This expanded guidance offers businesses greater clarity regarding the types of arguments that may be deployed in response to an Article 102 investigation.

The overall message is therefore nuanced. While the new Guidelines embrace a more flexible and effects-based approach to evidence, they simultaneously preserve powerful presumptions for conduct that the Commission views as inherently or highly likely to be exclusionary. As a result, dominant undertakings may benefit from greater clarity regarding the applicable legal framework, but will still face a substantial evidentiary burden when seeking to rebut the Commission’s theories of harm.

Rebutting presumptions and establishing objective justifications

A dominant undertaking may seek to justify conduct by reference to objective circumstances or consumer interests. In practice, a credible defense should identify a genuine efficiency, quality, safety, resilience or sustainability benefit; show that the conduct is necessary or indispensable to achieve it; demonstrate that the benefit is passed through to consumers or otherwise protects the competitive process; and show that the conduct does not eliminate effective competition. Out-of-market or ecosystem-wide benefits require especially careful evidence of causation and proportionality.

Where a presumption applies, the undertaking should address the trigger directly— for example, by showing that an arrangement is not exclusive in substance, that its duration or coverage is too limited to foreclose, or that customers and rivals retain realistic alternatives. A presumption can change the practical evidentiary burden, but it does not make unsupported assertions sufficient: preserve contemporaneous commercial, technical and economic evidence and ensure that the Commission or national authority can test the rebuttal.

Dominance in the digital era

Digital markets raise recurring Article 102 questions because the same features can support both dominance and foreclosure. Network effects, low marginal distribution costs, user lock-in, switching costs, default positions, data feedback loops, access to data or compute, and ecosystem integration may make entry or expansion more difficult. AI-related data and computational resources should be analyzed as potential sources of market power and competitive advantage, not as new standalone categories of abuse. The conduct still must be assessed under the Article 102 framework.

The Guidelines place significant emphasis on the specific characteristics of digital markets when assessing dominance. They recognize that network effects, whereby the value of a product or service increases with the number of users, can create significant barriers to entry and expansion, particularly in platform markets. The Commission also highlights the role of digital ecosystems, user lock-in, switching costs and data-driven advantages, which may reinforce market power across multiple products and services. Notably, the Guidelines identify access to large datasets and computational resources as an increasingly important source of market power in the context of artificial intelligence development. As a result, the assessment of dominance is no longer confined to the competitive dynamics of a single market but increasingly takes into account broader ecosystem effects and interconnections between markets.

Beyond the assessment of dominance, the new Guidelines identify access restrictions as a distinct category of potentially abusive conduct. In particular, the Commission recognizes that restrictions on access to key functionalities, interfaces, interoperability solutions or other critical inputs may hinder rivals’ ability to enter, compete or expand in digital markets. This reflects the growing importance of access-related theories of harm in technology and platform cases, where control over infrastructure, data or interoperability mechanisms may be as significant as traditional forms of market power.

Finally, while the Guidelines acknowledge that the regulatory environment may form part of the factual context in which conduct is assessed, they stop short of creating any form of regulatory safe harbor. In particular, this silence should not be read as suggesting that compliance with the Digital Market Act (DMA) is sufficient to shield conduct from scrutiny under Article 102 TFEU. Rather, the two regimes are complementary and capable of applying in parallel. Gatekeepers and other dominant undertakings should therefore continue to assess their conduct separately under both frameworks and should not assume that compliance with one regime will automatically ensure compliance with the other.

The DMA imposes ex ante obligations on designated gatekeepers, while Article 102 applies to abuse by any dominant undertaking where the Treaty conditions are met. The regimes may address overlapping conduct, but they ask different legal questions and have different enforcement tools. DMA compliance is relevant context, not a safe harbor: it neither immunizes conduct under Article 102 nor removes the need to assess dominance, departure from competition on the merits and exclusionary capability.

Recent Commission practice illustrates the enforcement focus on defaults, tying, self-preferencing, interoperability, data advantages and conduct spanning adjacent markets. The same product decision should therefore be reviewed under both regimes where the facts warrant it, without assuming that a DMA analysis answers the Article 102 question.

More broadly, the new Guidelines confirm the Commission’s intention to adapt Article 102 enforcement to the realities of digital and AI-driven markets, where data, ecosystems, interoperability and access to key inputs increasingly shape competitive dynamics and may give rise to novel forms of exclusionary conduct.

High-risk conduct: Where compliance review matters most

The Guidelines confirm that certain categories of conduct are likely to attract particularly close scrutiny under Article 102 TFEU and therefore warrant careful assessment by dominant undertakings.

Exclusivity rebates: The presumption discussed above is not limited to explicit exclusivity clauses. It may also apply to de facto exclusivity created through volume thresholds, stocking requirements or other mechanisms that incentivize customers to concentrate purchases with the dominant undertaking. Businesses should therefore look beyond contractual language and assess whether rebate or purchasing schemes could, in practice, produce similar effects. Particular attention should be paid to the coverage and duration of such arrangements, as these factors will be relevant when seeking to rebut a presumption of foreclosure. The revised Guidelines also indicate that the absence of reliable cost data, including situations where such data is not made available by the dominant undertaking, will not necessarily preclude an assessment. In such circumstances, greater weight may be placed on the qualitative evidence available.

Self-preferencing and leveraging. The Guidelines express recognize self-preferencing as a standalone abuse rather than a variant of refusal to supply. As a form of non-pricing conduct, self-preferencing generally falls outside the AEC and price-cost framework. Accordingly, ranking decisions, platform design choices or interoperability measures that favor a dominant undertaking’s own offering on adjacent markets should be assessed under the general two-step test of whether the conduct departs from competition on the merits and is capable of producing exclusionary effects.

Access restrictions, interoperability and refusal to supply. The new Guidelines identify access restrictions as a distinct category of potentially abusive conduct. Restricting, degrading or delaying access to key inputs, including interoperability information, may be caught under Article 102 even where the input is not indispensable, although indispensability remains an important factor weighing in favor of intervention. A stricter standard applies where a dominant undertaking is asked, for the first time, to grant access to an input developed for its own use. In those circumstances, the established requirements of indispensability, meaning that there is no actual or potential substitute, and the elimination of all effective competition downstream continue to apply, together with the new-product criterion where the input is protected by intellectual property rights. That framework should not be assumed for every access dispute. Discriminatory access, delays, degraded interoperability, selective technical restrictions, margin squeeze, or the removal or destruction of infrastructure may require a general exclusionary-effects analysis instead.

Public or privatized infrastructure. Public funding or later privatization does not, by itself, displace the Bronner framework. The Court has previously treated the question as turning on the nature of the access refusal, the terms of the asset acquisition and the undertaking’s decision-making autonomy over access. Businesses operating infrastructure with a public or regulated history should therefore analyze both the ownership and governance facts and the precise conduct challenged.

Predation and margin squeeze. The Guidelines largely codify existing case law in this area. Pricing below average variable cost continues to be regarded as particularly problematic and may be presumed to lack any legitimate commercial justification other than eliminating rivals, leaving little room for defense. Pricing between average variable cost and average total cost may also be abusive where it forms part of a plan to eliminate a competitor.

Similarly, the established framework for margin squeeze remains intact, with the Commission continuing to rely on the long-run average incremental cost benchmark established in prior case law when assessing whether an equally efficient competitor could compete profitably.

Taken together, these categories illustrate that the Commission’s enforcement priorities increasingly focus on conduct capable of foreclosing competitors through exclusivity mechanisms, access restrictions, ecosystem advantages or pricing strategies. Businesses operating in digital and platform markets are likely to face particular scrutiny in this respect.

Translating the Guidelines into compliance priorities for businesses

For dominant and potentially dominant undertakings, the Guidelines should be viewed as a trigger for a documented, cross-functional review, not as a checklist for labels. They provide a roadmap of the Commission’s enforcement priorities and therefore present a timely opportunity to revisit internal compliance procedures.

At a minimum, companies should consider:

• Mapping market power before major launches, renewals or product changes: assess market definition, shares and trends, customer dependence, buyer power, entry and expansion constraints, network effects, switching costs, data, compute and ecosystem reach. Market shares below 40% may be a useful indicator, but should not be treated as a safe harbor from Article 102 scrutiny.
• Reviewing existing distribution, supply and rebate arrangements to identify explicit or de facto exclusivity mechanisms, including purchasing thresholds and stocking requirements and retroactive rebates. Businesses should also evaluate whether the factual circumstances of those arrangements could support a rebuttal if challenged. Separate pure volume discounts from loyalty-inducing mechanisms and obtain legal and economic review before launch.
• Assessing self-preferencing and default-ranking risks in platforms, marketplaces and broader digital ecosystems, particularly where the undertaking operates both the upstream platform and downstream services that compete with third parties. Document neutral criteria, quality improvements, testing and governance for any preferential treatment.
• Re-evaluating access, interoperability and data-sharing policies, bearing in mind that compliance with sector-specific regulation, including the DMA, does not shield an undertaking from Article 102 scrutiny. Denials of access, delays, discriminatory terms, API degradation, infrastructure removal or destruction, and changes affecting multi-homing or switching should be flagged for legal review before implementation, with a documented assessment of their objective justification and competitive impact. Analyse Bronner only where the conduct is a true refusal-to-supply case.
• Maintaining contemporaneous evidence of legitimate business rationales and efficiencies, including pricing analyses, internal decision-making documents, investment incentives and evidence demonstrating consumer benefits. Such evidence may prove critical where the burden of proof shifts to the dominant undertaking to rebut a presumption of harm.

Ultimately, the new Guidelines signal continuity in many respects, but they also reflect the Commission’s determination to adapt Article 102 enforcement to increasingly digital, data-driven and ecosystem-based markets. Companies that rely on rebate schemes, platform intermediation, privileged access to data or control over key inputs should therefore take the opportunity to reassess existing practices against the Commission’s updated framework.

*Trainees Cian Clehane and Clémence Gomez also contributed to this article.