EC Recalibrates Guidance for Abuse of Dominance Under Article 102 TFEU

On September 3, 2026, the European Commission (EC) adopted Guidelines on the application of Article 102 TFEU to abusive exclusionary conduct by dominant undertakings (the Guidelines).

The Guidelines, which follow a public consultation on draft guidelines published in July 2024 (the Draft Guidelines[1]), replace the EC’s 2008 Guidance on its enforcement priorities. They are designed to codify EU Courts’ case-law and the EC’s decisional practice, provide legal certainty, guide national competition agencies, and allow dominant companies to assess whether specific types of conduct are unlawful.

The Guidelines take account of certain criticisms levelled against the Draft Guidelines. Among other things, they recognize that market shares below 40% are unlikely to support a dominance finding, narrow the types of conduct where the EC may presume unlawful exclusion, maintain the as-efficient competitor (AEC) test in certain situations, and provide detailed guidance on business justifications.

The Guidelines do not suggest any softening in the EC’s treatment of exclusionary conduct by dominant companies. On the contrary, they confirm the EC’s resolve to intervene in respect of such conduct, particularly in digital and innovation-driven markets.

I. Dominance: Safe Harbour Reinstated, Digital Framework Advanced

Market-share thresholds. The Draft Guidelines had not maintained the 2008 Guidance’s view that dominance was unlikely below 40%, noting only that shares below 10% would exclude dominance save in exceptional circumstances. The Guidelines restore the safe harbor, acknowledging that “dominance is generally unlikely” where market shares are below 40%, but 50% or higher market shares held for a significant period are – save in exceptional circumstances – evidence of dominance.[2]  

The Guidelines clarify that dominance and market power are not synonyms. Dominance is a “binary concept” – an undertaking either is or is not dominant – whereas market power exists on a spectrum. Once dominance has been established, the dominant undertaking’s degree of market power is not decisive for the applicability of Article 102 TFEU, although it may be relevant for the assessment of whether conduct can have exclusionary effects.[3]

Digital markets and barriers to entry. The Guidelines underline the importance of barriers to entry in data-intensive and digital markets. They expressly address data-driven advantages (including access to unique datasets, data accumulation, and data-related network effects), digital ecosystems of interlinked products and platforms, and in the AI context, access to large, high-quality datasets and sufficient computational power.[4] The final text adds talent recruitment and retention, and capital strength (including the ability to acquire competitors and innovators) as potential barriers to entry and expansion.[5] Companies with lower market shares may be found dominant where network effects, ecosystem dynamics, and high entry barriers are present.

Aftermarkets. The Guidelines add a new standalone aftermarket framework, explaining that primary-market competition constrains aftermarket power, and therefore excludes dominance in the aftermarket, only if four cumulative conditions are met: (i) primary market customers can make an informed choice taking into account lifecycle pricing; (ii) they are likely to do so; (iii) a sufficient number would adapt their primary-market purchasing in response to significant aftermarket price increases; and (iv) they would do so within a reasonable time. [6]

Collective dominance. The Guidelines extend the scope of Article 102 TFEU to collective dominance, namely situations where legally independent entities act collectively. The EC acknowledges that Article 102 TFEU has been “enforced only rarely” in such circumstances, but notes that the “increasing use of algorithms may lead to more findings” of collective dominance based on tacit coordination, which algorithms can facilitate by improving market observability and helping firms reach and maintain coordination terms.[7]

II. Legal Test: Presumptions Weakened

Two limbs, but no fixed sequence. Under the Draft Guidelines, the EC envisaged asking two questions: (i) whether conduct departed from competition on the merits and (ii) whether conduct was capable of producing exclusionary effects.[8] The Guidelines retain the two-limb analysis, while clarifying that the two limbs need not be assessed separately or in a fixed order.[9] The final text identifies scenarios where a two-limb analysis is unnecessary: conduct that is subject to a specific analytical framework recognised by the EU Courts, conduct that may exclude an as-efficient competitor, and conduct that is harmful by its very nature.[10] 

Codification of types of conduct. The Guidelines codify the law in respect of conduct that has been assessed by the EU Courts, including predatory pricing, margin squeeze, rebates, tying and bundling, exclusive dealing, access restrictions, and refusal to supply.[11]

Weakening of presumptions. The Draft Guidelines divided conduct into three categories: (i) conduct requiring specific evidence of exclusionary capability; (ii) conduct subject to a rebuttable presumption concerning their capability to produce exclusionary effects; and (iii) ‘naked restrictions’.[12] The Draft Guidelines’ reliance on presumptions was much criticised, including because of the burden it placed on dominant companies.

The Guidelines replace that classification with a sliding scale: “the more a given conduct is considered generally likely to distort effective competition, the less case-specific evidence is required to prove that this is the case, and the other way around”.[13] Presumptions remain for exclusive dealing and conduct that is considered harmful by its very nature.[14] In such cases, companies must bear the evidentiary burden of rebutting those presumptions. 

In other cases, the Guidelines state that, for conduct to harm effective competition, it must be capable of producing exclusionary effects in light of the relevant economic context.[15] These effects must be more than hypothetical, although the EC need not prove actual exclusionary effects or that the conduct would be profitable.[16] There is no de minimis threshold, and conduct need only increase the likelihood of exclusionary effects rather than be their sole cause.[17] The EC must articulate a theory of harm, but need not systematically use a formal counterfactual or any other single methodology.[18]

III. The AEC Test: Clarified and Maintained

The 2008 Guidance used the AEC test to assess price-based conduct as one factor in the EC’s broader foreclosure analysis.[19] The Guidelines note that “as a rule” an analysis of whether conduct is exclusionary should examine whether it is capable of excluding or hampering the growth of an equally efficient competitor.[20] Accordingly, the assessment of pricing conduct (predatory pricing, margin squeeze, and non-exclusive rebates) will typically be based on the AEC test. 

The AEC test typically uses the dominant undertaking’s own prices and costs at the time of the conduct in question, although dynamic considerations may also be taken into account.[21] Where pricing conduct can exclude an equally efficient competitor, the EC need not prove separately that it departs from competition on the merits.[22] 

In relation to non-pricing conduct, the Guidelines state that the AEC test may be applied, while noting that the test may have difficulty capturing non-price parameters of competition such as quality, innovation, and choice. Moreover, in digital markets or ecosystems, conduct or market features including innovation, access to data, multi-sidedness, user behavior, or network effects may impede entry by rivals at an early stage, preventing the emergence of as-efficient competitors, in which case the AEC test is not relevant.[23]

In practice, the Guidelines acknowledge that a dominant undertaking may submit its own AEC evidence, which the EC is required to examine.[24]

IV. Digital Markets: Rules Against Self-Preferencing Explained

The Guidelines treat in significant detail the circumstances in which self-preferencing by a dominant company may be abusive. They confirm that there is no general rule against self-preferencing by a dominant undertaking,[25] and explain that product integration, promotion of an undertaking’s own services, and differentiated treatment are common across many industries and are not inherently abusive. A dominant firm treating its own products more favorably than those of its competitors may be problematic only where it departs from competition on the merits and the conduct in question is capable of producing exclusionary effects.[26]

The Guidelines distinguish offensive leveraging (where the undertaking uses its dominance in one market to strengthen its position in a related market) from defensive leveraging (where it protects the dominated market against a competitive threat arising elsewhere). Concerns are more likely where the dominant undertaking controls the conditions of access to a neighboring market or to an ecosystem of interlinked products, services, or platforms.[27]

Self-preferencing may involve rankings, positioning, display, defaults, steering, auction design, selection mechanisms, or a combination of practices over time.[28] The EC will examine whether the leveraging market is important and irreplaceable for rivals, whether the difference in treatment lacks an objective business rationale, whether there was an expectation of neutrality or openness that the undertaking frustrated through non-objective or non-transparent conduct, and whether the practice is likely to influence users irrespective of the intrinsic quality of the favored product.[29]

The input or access point need not be indispensable. The fact that competitors cannot effectively replace the relevant source of business may support a finding of exclusionary capability even where that source is not indispensable.[30] In digital markets, network effects, user lock-in, switching costs, and data-driven advantages may form part of the broader market context in which self-preferencing is assessed.[31] 

In practice, platform and ecosystem operators should document the product, quality, security, and efficiency reasons for rankings, defaults, access conditions, and integration decisions. They should test whether similarly situated third parties receive materially different treatment, whether published openness commitments create expectations, how users respond to interface design, and whether less restrictive alternatives could achieve the same objective.

V. Conduct That Is “by its very nature harmful”: Presumption Maintained

The Draft Guidelines’ identification of “naked restrictions” is maintained but termed “conduct that is by its very nature harmful to competition.”[32] The category covers conduct that a dominant undertaking would have no commercial interest in implementing other than to harm competition (including pay-for-delay, dismantling infrastructure on which a rival depends, conditioning payments on customers not selling, or delaying the sale of a competitor’s products, withdrawing discounts where a customer replaces a dominant undertaking’s product with a rival’s product, and using regulation to impose unfair and discriminatory conditions).[33] Such conduct may not be considered competition on the merits and is unlikely to produce consumer benefits.[34] Once the EC has established that conduct is “by its very nature” harmful to competition, such conduct is deemed to distort effective competition.[35]

The EU Courts have not established whether a dominant undertaking can rebut an adverse finding by proving that its conduct was incapable of producing exclusionary effects in the particular circumstances. Pending such clarification, the EC considers that such a challenge could succeed only “very exceptionally”. The EC will, however, adapt its practice if and when the Courts prescribe the same evidentiary approach as for restrictions by object under Article 101 TFEU.[36] Companies may raise objective necessity and efficiency arguments, but the EC considers that they are “very unlikely” to succeed.[37]  

VI. Defenses: More Guidance, but Burden and Standard Maintained

The Guidelines provide more detailed guidance on business justifications, making clear that a dominant company bears the burden of substantiating such justifications. The defenses recognized in the 2008 Guidance and the Draft Guidelines (objective necessity and efficiencies) are maintained.[38] The efficiency defense retains four cumulative conditions: that the distortion of competition caused by a dominant firm’s conduct is counterbalanced by efficiencies, offsets harm to competition and consumers, is necessary, and preserves effective competition.[39] The Guidelines provide more detailed guidance on qualifying objectives, evidence, consumer benefits, and less restrictive alternatives.[40]

Objective necessity. The 2008 Guidance addressed objective necessity briefly and focused principally on external health and safety factors.[41] The Draft Guidelines broadened the examples to legitimate commercial, technical, and public-interest considerations.[42] The Guidelines retain this approach and add protecting the integrity or security of a dominant firm’s products and the disposal of excess or obsolete stock via below-cost pricing.[43] They also introduce guidance on interoperability: technical difficulty is not in itself sufficient to justify a refusal to ensure interoperability with an input controlled by the dominant firm; refusal will be regarded as objectively necessary only if interoperability would compromise product integrity or security or the required solution is technically impossible. A failure to respond to an interoperability request may be treated as evidence that a refusal to interoperate is not objectively justified.[44]

The defense remains narrow: vague claims are insufficient, an undertaking may not enforce other companies’ legal compliance, and the conduct must be proportionate with no less restrictive means available.[45]

Efficiencies. The Guidelines expand the 2008 Guidance’s brief references to technical improvements and cost reductions and the Draft Guidelines’ four-part efficiency defense test.[46] They recognize static and dynamic efficiencies, including R&D, innovation, infrastructure, investment recoupment, relationship-specific investments, hold-up and free-rider concerns, capital-market imperfections, new distribution methods, sustainability, and resilience.[47]

The Guidelines also take a step toward recognizing out-of-market efficiencies, as they permit consideration of efficiencies in related markets where the harmed and benefiting consumers substantially overlap.[48] The Guidelines refine the Draft Guidelines’ treatment of harmful conduct through a “sliding scale”: the greater the potential for harm, the less likely the defense will succeed.[49]

Evidence. The burden remains on a dominant undertaking.[50] The Guidelines require objective, concrete, and verifiable efficiencies supported by evidence such as contemporaneous documents, financial records, historical examples, expert studies, and fact-specific economic models.[51] Efficiencies and consumer benefits should be quantified where reasonably possible; projected benefits must be timely, substantiated, and discounted; and pass-on must at least compensate affected consumers.[52]

In bringing an efficiency defense, a dominant firm must explain why realistic, less restrictive alternatives would be significantly less efficient.[53] The conduct must not remove all or most remaining sources of actual or potential competition.[54]

The expanded framework makes contemporaneous documentation more important. Documents supporting business decisions should record the commercial objective, causal mechanism, expected consumer benefit, alternatives considered, investment risk, recoupment period, and remaining sources of competition.

VII. Conclusion

The Guidelines codify, clarify, and expand the EU Courts’ and EC’s interpretation and application of Article 102 TFEU. In so doing, they serve as an operational manual that should allow the EC to bring cases more easily and enable dominant companies to self-assess. In providing greater predictability, a clearer analytical framework, and more detailed guidance on business justifications, the Guidelines should allow dominant companies to predict with greater certainty whether specific types of conduct may be unlawful in the EU and to develop contemporaneous evidence justifying their conduct. 

---

[1] See Cleary Antitrust Watch, Commission Revamps its Article 102 Guidance and Will Adopt new Guidelines in 2025, March 30, 2023, available at https://www.clearyantitrustwatch.com/2023/03/commission-revamps-its-article-102-guidance-and-will-adopt-new-guidelines-in-2025/. 

[2] Guidelines, para. 24.

[3] Guidelines, para. 19.

[4] Guidelines, paras. 31-34.

[5] Guidelines, para. 30.

[6] Guidelines, para. 39.

[7] Guidelines, para. 48 and footnote 105.

[8] Draft Guidelines, paras. 45-46.

[9] Guidelines, paras. 59-62.

[10] Guidelines, paras. 62-65.

[11] Guidelines, Sections 4.2 to 4.8.

[12] Draft Guidelines, para. 60(a)-(c).

[13] Guidelines, para. 58.

[14] Guidelines, paras. 155-160 and 197-200.

[15] Guidelines, paras. 80-83.

[16] Guidelines, para. 82.

[17] Guidelines, paras. 83 and 99.

[18] Guidelines, paras. 57 and 102.

[19] 2008 Guidance, paras. 23-27.

[20] Guidelines, para. 85.

[21] Guidelines, para. 90.

[22] Guidelines, paras. 64 and 89.

[23] Guidelines, para. 94.

[24] Guidelines, para. 93.

[25] Guidelines, para. 191.

[26] Guidelines, para. 195.

[27] Guidelines, para. 192.

[28] Guidelines, para. 194.

[29] Guidelines, para. 196.

[30] Guidelines, para. 196(a).

[31] Guidelines, paras. 31, 34, and 195.

[32] Draft Guidelines, para. 60(c); Guidelines, paras.197–201.

[33] Guidelines, para. 198(a)-(d).

[34] Guidelines, para. 197.

[35] Guidelines, para. 199.

[36] Guidelines, para. 200.

[37] Guidelines, para. 201.

[38] 2008 Guidance, para. 28; Draft Guidelines, para. 167; Guidelines, para. 202.

[39] 2008 Guidance, para. 30; Draft Guidelines, para. 169; Guidelines, para. 220.

[40] 2008 Guidance, paras. 28-31; Draft Guidelines, paras. 167-171; Guidelines, paras. 202-245.

[41] 2008 Guidance, para. 29.

[42] Draft Guidelines, para. 168.

[43] Guidelines, paras. 208-210.

[44] Guidelines, para. 209.

[45] Guidelines, paras. 205 and 210-211.

[46] 2008 Guidance, para. 30; Draft Guidelines, para. 169; Guidelines, paras. 212-245.

[47] Guidelines, paras. 213-218.

[48] Guidelines, paras. 236-238.

[49] Draft Guidelines, para. 170; Guidelines, para. 222.

[50] 2008 Guidance, para. 31; Draft Guidelines, para. 171; Guidelines, para. 204.

[51] Guidelines, paras. 223-225.

[52] Guidelines, paras. 226-231.

[53] Guidelines, para. 239.

[54] Guidelines, paras. 243-245.