The European Commission’s Article 102 TFEU Guidelines: A Critical Assessment

Introduction

On 3 September 2026, the European Commission (Commission) published its Guidelines on the application of Article 102 TFEU to exclusionary conduct by dominant undertakings (Guidelines). The Guidelines replace the 2008 Guidance on enforcement priorities (Guidance), which is withdrawn 30 days after publication of the new text. They arrive after a 2024 draft that was widely criticised for relying too heavily on form, presumptions, and conduct categories at the expense of case-specific analysis.

The Guidelines are structured around four pillars: (i) the assessment of dominance, (ii) general principles for determining whether conduct distorts effective competition, (iii) specific analytical frameworks for particular types of conduct, and (iv) objective justifications and efficiencies. While focused on exclusionary abuses, the principles on dominance and objective justifications also apply to exploitative conduct.

Overview of Key Revisions 

The final Guidelines are clearer and, in several respects, more balanced than the much-debated August 2024 draft. The Guidelines now set out with more clarity when the Commission will assess separately whether conduct departs from competition on the merits and whether it is capable of producing exclusionary effects, as well as set out when a single assessment suffices (Guidelines, paras. 63–65). The Guidelines also introduce the concept of a theory of harm, requiring the Commission to identify the mechanism by which the conduct is said to restrict competition, albeit only for explanatory purposes rather than as a formal analytical step. Further, the Guidelines reduce the range of conduct subject to presumptions and now explain the proposed rationale for each remaining presumption by reference to the nature and established propensity of the conduct. In addition, the AEC principle, which the draft appeared to marginalise, has a clearer and more prominent role: It applies principally to predatory pricing, margin squeeze, and conditional rebates, and a finding that conduct can exclude a hypothetical equally efficient competitor may establish both departure from competition on the merits and exclusionary capability. The final text consolidates naked restrictions and conduct harmful by its very nature in a single section, narrowing the category to practices with no plausible commercial rationale other than eliminating rivals. The Guidelines also substantially expand the treatment of objective justification. The objective justification chapter expands from five paragraphs to 44, covering objective necessity, efficiencies, sustainability benefits, resilience objectives, and out-of-market efficiencies. They also offer useful guidance on network effects, data advantages, aftermarkets, and multifaceted conduct.

Those improvements are welcome: The final Guidelines are easier to apply and more attentive to economic context than their predecessor. The final text is not, however, immune from criticism. The sections that follow examine the points of contention.

Legal Standing and Codification

The Guidelines are soft law. They are a Commission communication, not legislation, and do not themselves create binding legal obligations. Article 102 remains unchanged, and its meaning is ultimately for the EU courts, not the Commission, to determine. The Guidelines therefore neither create new forms of illegality nor narrow the existing prohibition.

They nonetheless serve a broader function than the earlier Guidance, which addressed the Commission’s enforcement priorities. The new text states how the Commission considers Article 102 should be applied. Paragraph 8 describes the Guidelines as outlining the current state of the case law and indicates that the Commission will state its own position where the case law is silent or open to interpretation. That is a more ambitious — and more contestable — exercise. The Commission has broad discretion to choose the cases it pursues, but it cannot determine the content of Article 102. In our view, parts of the Guidelines read the case law selectively, particularly where they suggest that an as-efficient-competitor analysis may be dispensed with. Further, the Guidelines repeatedly rely on General Court judgments that are currently under appeal (e.g., Qualcomm (Predation), Case T-671/19) and, in several instances, draw a contrario inferences from the Court’s wording. The Guidelines do, however, constrain the Commission. Having adopted them, it may not depart from them in an individual case without giving reasons. That self-imposed constraint has practical value, and an unexplained departure to an undertaking’s detriment, while entirely possible (see in this regard also our recent analysis on the judgment in Booking Holdings v Commission (eTraveli)), may provide an independent ground of challenge. It does not bind the EU courts, which did not treat the earlier Guidance as the governing legal test, despite having incorporated references to it in the form of obiter dicta (see e.g., Unilever (Case C-680/20 (para. 51))). The Guidelines may limit the Commission’s conduct, but they cannot extend Article 102 beyond the scope recognised by the Courts.

Codification is useful where it makes settled case law more accessible and clarifies the Commission’s analytical sequence. Greater operational certainty does not, however, ensure legal certainty. Where the case law is unsettled or unclear, the Guidelines may crystallise the Commission’s preferred interpretation before a court has endorsed it. This is especially important for the presumptions discussed below. A Commission communication cannot alter the authority’s legal burden of proving an infringement, and it remains untested whether the case law supports every presumption in the Guidelines, particularly where rebuttal is said to be possible only in exceptional circumstances. The presumptions reliably indicate how the Commission will conduct a case but not necessarily what the Courts will accept.

Nor do the Guidelines bind national competition authorities or national courts. The European Competition Network welcomed them while noting that national authorities remain free to apply Article 102 in accordance with the case law. National authorities and claimants are nevertheless likely to invoke the framework in investigations and damages actions. Uptake may therefore be rapid but uneven, with some authorities adopting the more contested passages and others taking a different approach.

The Two-Prong Test for Exclusionary Abuses and the Circular Burden of Proof

The general test for distortion of effective competition has two prongs (Guidelines, para. 59). First, does the conduct depart from competition on the merits? Second, is the conduct capable of producing exclusionary effects by materially hindering rivals’ ability or incentive to constrain the dominant undertaking? Effects must be more than hypothetical, but the Commission need not prove actual exclusion, direct consumer harm, profitability, or sole causation. It must, however, explain how the conduct increases the likelihood of the alleged effect, assessed on the facts known at the time of implementation.

A separate assessment of both prongs may be unnecessary in three situations: when a specific analytical framework applies, when the conduct can exclude a hypothetical equally efficient competitor, or when the conduct is harmful by its very nature. Classification is therefore decisive and carries the risk that a conduct label or price-cost result substitutes for a complete explanation of why the practice is abusive in its market context.

Classification is not the only source of difficulty. A further tension arises from the relationship between the concept of distortion of effective competition, objective justification, and departure from competition on the merits.

Paragraph 13 of the Guidelines lays out a three-step infringement test: 

(i) dominant position
(ii) conduct that distorts effective competition
(iii) objective justification

Treating objective justification as a separate step implies that distortion is established before justification is considered — yet the Guidelines’ own definition of distortion of effective competition (fn 7) folds objective justification into the concept, effectively double-counting it.

Competition on the merits adds a further layer. Under the two-prong test for distortion (Guidelines, para. 59), the first question is whether the conduct departs from competition on the merits. Departure from competition on the merits is therefore part of establishing distortion, meaning step (ii) of the paragraph 13 test. Yet paragraph 6 of the Guidelines defines departure from competition on the merits by reference to conduct that hinders competition “through recourse to methods different from those governing normal competition” and then adds that “such behaviour, if not objectively justified, is referred to as an ‘exclusionary abuse.’” Objective justification therefore not only reappears inside the very concept it is supposed to follow, but the reader is also left unclear on the boundaries between distortion of effective competition and competition on the merits.

The Guidelines never clearly locate competition on the merits within the three-step infringement test, and two readings are possible. If distortion already incorporates the absence of justification (para. 6 and fn 7), then departure from competition on the merits and the absence of justification are assessed together at step (ii), and step (iii) becomes redundant. If distortion does not incorporate objective justification (paragraph 13’s sequential logic), then it is unclear whether departure from competition on the merits is assessed at step (ii) alongside exclusionary capability or partly at step (iii) alongside objective justification — since what counts as “normal” competition may itself depend on whether the conduct serves a legitimate aim.

This ambiguity matters also practically because it affects when the burden shifts. The Commission bears the burden of proving distortion. The undertaking bears the burden of establishing objective justification. If departure from competition on the merits sits entirely within step (ii), the Commission must prove it. If elements of that assessment bleed into step (iii), the undertaking may effectively have to disprove departure from the merits as part of its justification case.

Conduct-Specific Analytical Frameworks

The final Guidelines replace the draft’s broad two-category structure with eight analytical frameworks. Predatory pricing, margin squeeze, exclusive dealing, tying and bundling, and refusal to supply each have a dedicated framework. Conditional rebates unrelated to exclusivity, self-preferencing, and access restrictions are assessed under the general principles with conduct-specific guidance. Section 4.10 separately addresses conduct harmful by its very nature, including practices with no plausible commercial rationale other than eliminating rivals.

The shorter and more differentiated taxonomy is an improvement, but its legal implications may warrant closer scrutiny. Once the conditions of a dedicated framework are met, the Commission may treat the conduct as distorting competition without repeating the general two-prong assessment. The category assigned to a practice can therefore determine the proof required and the safeguards available. The same concern applies to harmful-by-nature conduct: Linking that category to object-type abuse may be justified for conduct that is plainly exclusionary, but the category should remain narrow and should not become a route for avoiding analysis of plausible commercial explanations.

Paragraph 58: From Sliding Scale to Burden Shift

Paragraph 58 links the evidence the Commission must produce to a type of conduct’s general propensity to harm competition. It states that “the more a given conduct is considered generally likely to distort effective competition, the less case-specific evidence is required.” As a practical matter, that proposition is unobjectionable. Conduct with a well-established capacity to foreclose should not have to be analysed from first principles in every case.

Paragraph 58 then goes further. It states that “the demonstration of certain factual elements may lead to the application of a presumption” that the conduct distorts competition. Such a presumption “shifts the evidentiary burden from the Commission onto the dominant undertaking.” The Guidelines therefore move from a sliding scale governing the amount of evidence required to a presumption that changes who must produce evidence.

Those are distinct questions. A type of conduct’s general likelihood of causing harm may affect the amount of case-specific evidence the Commission needs to establish exclusionary capability. It does not follow that the undertaking must disprove the abuse. EU case law recognises presumptions for particular practices, including exclusivity under Hoffmann-La Roche (C-85/76) and pricing below average variable cost under AKZO (C-62/86). It does not establish a general sliding scale under which the perceived likelihood of harm determines both the evidence the Commission must produce and whether the undertaking bears the burden of rebuttal.

Paragraph 58 is therefore open to criticism on two grounds. First, it presents the general sliding-scale principle as settled law rather than identifying it as the Commission’s own position. Second, it uses that principle to support a broader transfer of the evidentiary burden. The Commission may require less evidence where conduct has a recognised propensity to foreclose, but it must still prove the facts triggering any presumption and produce affirmative evidence of exclusionary capability. If a finding of abuse rests principally on the undertaking’s failure to rebut, a contested interpretation will have entered the enforcement framework before the EU courts have endorsed it.

Rebutting Presumptions

The legal burden described above must be distinguished from the evidentiary burden. Once the Commission proves the facts that trigger a presumption, the Guidelines place the onus on the dominant undertaking to produce rebuttal evidence. This does not transfer the ultimate burden of proof, but it may require the undertaking to challenge the Commission’s theory before the Commission has conducted a full assessment of exclusionary capability.

The required standard for rebuttal heightens this concern. In Intel (Case C-413/14 P) and Unilever (Case C-680/20), the Court held that, where an undertaking submits supporting evidence that its conduct could not restrict competition, the Commission must assess the conduct’s capability to do so in light of all relevant circumstances. Supported arguments are therefore sufficient to trigger the Commission’s duty to investigate. The Guidelines appear to set a higher threshold: The undertaking must produce evidence that is relevant and of “sufficient probative value” to rebut the presumption. The difference is material. Under the case law, the undertaking’s evidence activates the Commission’s obligation to assess capability. Under the Guidelines, the Commission may first decide whether the undertaking’s evidence is strong enough to rebut the presumption — and only then proceed to a substantive assessment.

If rebuttal succeeds, the question becomes whether the Commission must still conduct a full effects analysis. Paragraph 160 expressly requires an overall assessment of all relevant circumstances where a presumption applying to exclusive dealing is rebutted. Paragraph 58, however, does not impose the same requirement whenever a presumption applying to another form of conduct is rebutted. This creates uncertainty: For conduct other than exclusive dealing, it is unclear whether the Commission must undertake a comprehensive assessment of exclusionary capability or may simply conclude that the rebuttal evidence is insufficient and maintain its finding of abuse.

The Guidelines do not mention the presumption of innocence or the principle that doubt must benefit the undertaking. That silence cannot override either rule, but it leaves key safeguards absent from the text that authorities, courts, and undertakings will reach for first. A decision based mainly on an undertaking’s failure to rebut a presumption, without cogent affirmative evidence of exclusionary capability, would face substantial risk on appeal.

The As-Efficient Competitor Principle

Compared with the August 2024 draft, the final Guidelines demarginalize the AEC principle and give it a clearer yet deliberately confined role to pricing abuses. Price-cost analysis is principally relevant to predatory pricing, margin squeeze, and conditional rebates. A finding that conduct can exclude a hypothetical equally efficient competitor may establish both departure from competition on the merits and exclusionary capability. Conversely, evidence that an equally efficient rival could compete profitably remains relevant to the defence, but it is not a general safe harbour.

The limitation of the AEC principle to pricing conduct was expected given the EU court’s direction of travel in recent jurisprudence and is, in itself, broadly welcome. The AEC test can be difficult to apply and is often uninformative for nonpricing conduct or in markets where scale, network effects, and data advantages prevent rivals from matching the dominant undertaking’s cost structure. A rival that is less efficient or has not yet reached equal efficiency may still impose a genuine competitive constraint. The Guidelines therefore allow nonpricing cases to focus on other forms of exclusionary effect. The concern, however, is that the Guidelines go further than confining the AEC principle to pricing cases: They appear to treat it as dispensable even within pricing and rebate cases where the case law has recognised its relevance. Where an undertaking submits a properly supported AEC analysis, Intel requires the Commission to engage with it rather than dismiss it because the Commission did not consider the test necessary at the outset.

Overall, unlike the 2008 Guidance — which organised the analysis of anticompetitive conduct around consumer welfare, anticompetitive foreclosure, and the AEC test (particularly for pricing conduct) — the Guidelines reconstruct the analysis around the protection of effective competition and effective competitive constraints. The shift is perhaps most clearly illustrated by the Guidelines’ reliance on Generics (Case C-307/18), where the Court held that “it is not the place of a dominant undertaking to dictate how many viable competitors are to be allowed to compete with it.” Read in that light, the Guidelines appear to place greater emphasis on preserving the competitive process itself and ensuring that effective competitive constraints remain on the market rather than focusing exclusively on whether conduct can be shown to harm consumers through a particular economic test.

Objective Justification and Efficiencies 

The treatment of objective justification is one of the most substantial and welcome revisions. Paragraphs 202 to 245 explain that an undertaking may show either that its conduct is objectively necessary to achieve a legitimate aim or that it produces efficiencies satisfying four cumulative conditions: the efficiencies are verifiable, consumers receive benefits that counteract the likely harm, the conduct is indispensable to achieving them, and effective competition is not eliminated. The evidentiary burden remains demanding and grows harder to discharge as the potential exclusionary harm increases.

The expanded guidance also recognises a broader range of benefits. Sustainability benefits, including less-polluting production and greater recyclability, may qualify as qualitative efficiencies. Certain collective environmental or social benefits may count even where individual consumers do not separately value them, provided the required consumer overlap and balancing conditions are met. The Guidelines also address resilience objectives, such as reducing supply chain dependencies or preserving capacity relevant to security and defence, and give more detail on interoperability refusals and out-of-market efficiencies.

This fuller, more technical framework improves the undertaking’s ability to identify the evidence it must produce, but it does not make an efficiency-based defence any easier. Commercial assertions, as previously, will not suffice. The undertaking must show a causal link which “normally” should be “direct” between the practice and its legitimate aim to a measurable efficiency (paras. 228–229), explain why a less restrictive alternative would not achieve the same result, and substantiate the “fair share” of passed-on consumer benefits with contemporaneous evidence (para. 231). The evidentiary requirements are still exacting, and it remains to be seen whether any investigated undertaking will successfully rely on an efficiency defence to rebut a finding of abuse under Article 102 of TFEU in the future. To date, this has not occurred. The practical significance of the expanded guidance, therefore, lies primarily in increased transparency and predictability rather than in making success on an efficiency defence any easier to achieve.

Implications for Dominant Undertakings

The division of authority has practical consequences. Adherence to the Guidelines is not a safe harbour: It does not preclude a finding of abuse by the EU courts, a national competition authority, or a national court, nor does it prevent a damages claim. Departure from the Guidelines, even where supported by a sound legal argument, may increase the risk of investigation. Where the evidence supports the conclusion that conduct falls outside the Guidelines, that position remains worth advancing before the Commission and, if necessary, on appeal.

We therefore advise dominant undertakings to calibrate their compliance arrangements to the positions in the Guidelines while preserving arguments that those positions overstate the legal test. Where the Guidelines attach a presumption to a category of conduct, the evidence capable of rebutting it should be identified and retained from the outset rather than assembled after an investigation has begun.

What to Expect From the New Framework

• Legal certainty: The Guidelines improve process predictability, particularly where they consolidate settled case law into a more accessible analytical framework.
• Residual uncertainty: Where the case law is unsettled or silent, the Commission’s stated position may crystallise one interpretation before a court has endorsed it. 
• Enforcement: The framework is likely to accelerate case assessment, and the Commission can be expected to rely more heavily on presumptions and demand contemporaneous evidence earlier in proceedings.
• Case selection: Digital ecosystems, aftermarkets, and multifaceted conduct now have a clearer route into Article 102 cases, broadening the range of practices that may attract scrutiny.
• Court review: On appeal, challenges can be expected to centre on whether the Commission departed without sufficient justification from its own framework, whether the facts genuinely triggered the presumption, whether causation was established, and whether the Commission properly engaged with rebuttal evidence rather than treating a failed rebuttal as sufficient proof.
• Further risk: Because the Guidelines bind only the Commission, uneven uptake by national competition authorities and national courts may preserve enforcement fragmentation despite the existence of a common text.
• External constraint: The presumption of innocence and the Commission’s burden of proof remain controlling principles.