New EU Dominance Guidelines Target Rebates, Self-Preferencing and Digital Platforms

Executive Summary

• What’s new: The European Commission’s revised guidelines on abuses of dominance mark a return to formalism. Retreating from the economics-driven “as-efficient-competitor” test, they introduce presumptions of harm, a sliding scale of enforcement and per se abuses. There is a clear focus on digital and data-driven markets and a codifying of recent enforcement practices and European courts’ case law.
• Why it matters: Companies should be aware of the shift in the evidential balance intended to make it easier for the EC to establish infringements through presumptions of harm. In particular, companies should consider whether their pricing and rebate practices fall within the EC’s categories of presumptively illegal conduct where the EC states it will generally not accept cost-based defences. In the digital sector, companies should consider reviewing practices potentially subject to stricter standards, such as distributing suites of software via digital platforms, restricting platform access and “self-preferencing.”
• What's next: The new guidelines have been controversial. The EC has signaled a more aggressive enforcement stance in pricing and digital conduct that pushes the envelope of Article 102 TFEU case law. Its future enforcement practice may be challenged in the EU courts alongside the existing Article 102 pipeline of appeals.

__________

The Article 102 TFEU Guidelines

Article 102 of the Treaty on the Functioning of the European Union (TFEU) case law has long been in flux. The perceived rigidity of 1980s and 1990s case law led the European Commission (EC) in 2009 to adopt more economics-driven “Guidance on the Commission’s Enforcement Priorities in Applying Article 82 of the EC Treaty [now Article 102 TFEU] to Abusive Exclusionary Conduct by Dominant Undertakings ” (2009 Guidance).

Since then, reverses in the EU courts, as well as a swathe of decisions and appeals on conduct in the technology sector, have led the EC to issue a restatement of Article 102 TFEU in the 3 September 2026 guidelines on exclusionary abuse of dominance (the Guidelines).

In part, the Guidelines mark a return to formalism. They aim to codify recent case law to create presumptions of harm, burden-shifting frameworks and a controversial per se category of abuses.

The EC rows back from its own economics-driven innovation, the “as-efficient competitor” (AEC) test underpinning prior EC guidance. This is the principle that the antitrust rules should not shield inefficient companies. Instead, antitrust should focus on conduct excluding rivals at least as efficient as the dominant company — not pricing, quality or innovative efforts that rivals can match.

While economically sound, appellate reverses have led the EC to downplay the AEC principle in its enforcement practice. The Guidelines codify that development. They limit the AEC principle to certain categories of conduct. They emphasize less clear-cut concepts of “departing from competition on the merits” and a discretion to prosecute abusive conduct. They also identify when the AEC test is not “relevant, ” based on the EC’s prior experience.

The Guidelines also signal the EC’s continued focus on digital and innovation-driven markets. They offer insights into more controversial alleged abuse fact-patterns in the digital sector, including “tying” of products through digital distribution channels that confer an advantage on the dominant company, “self-preferencing,” and denying or degrading access to digital platforms.

Dominance

The Guidelines confirm the case law that dominance may be presumed, subject to rebuttal, where the company has a market share exceeding 50%. The EC notes that dominance is generally unlikely below 40%, though it states there can be exceptions (¶24). The Guidelines helpfully recognize market shares can mean little in technology-driven, or data-intensive, markets: “In fast-growing markets with short innovation cycles, high market shares may be a less useful indicator of dominance because those shares may turn out to be ephemeral” (¶27).

The Guidelines also identify barriers to entry in digital markets, reflecting technological advances, including access to unique datasets and data accumulation as well as digital ecosystems across products and platforms. Where artificial intelligence (AI) is involved, the Guidelines state that access to datasets and processing capacity are potential barriers (¶31).

Collective Dominance

Dominance encompasses both single firm and “collective” dominance. Collective dominance is a relative rarity in the case law. It can arise where firms act collectively, either because there are structural or contractual links between them, or where firms recognize they are economically interdependent in a tacitly coordinating oligopoly.

The Guidelines recognizes that collective dominance through tacit oligopoly — an artifact of case law in the 1980s and 1990s — has been applied “rarely.” But warns that “the increasing use of algorithms may lead to more findings of collective dominance based on tacit coordination” (¶48).

Secondary Markets

Companies selling spare parts, technical support or consumables for their own products will often have high shares of their own branded “aftermarket.” Allegations of restrictive aftermarket policies are a common antitrust fact pattern.¹ The Guidelines confirm dominance cannot be presumed in aftermarket cases.

Instead, the Guidelines require a fact-specific assessment of whether the primary market is sufficiently competitive that it constrains anticompetitive conduct in the secondary market. For example, customers may take a “total cost of ownership” approach to procurement, causing sellers to lose customers if they set uncompetitive aftermarket prices for spares or consumables.

The Guidelines confirm the EFIM criteria that primary market competition will discipline secondary market power where: (i) customers are able to make an informed choice based on life cycle pricing, (ii) customers are likely to do so, (iii) a sufficient number would shift their primary-market purchasing in response to exploitative after-market conduct (e.g., significant price increase), and (iv) they would do so within a reasonable time frame (¶29).²

That said, failure to meet the conditions does not automatically establish dominance in the after-market. Rather, dominance must be assessed on the basis of other relevant factors, including the market position of competitors, barriers to entry and expansion, and countervailing buyer power.

Of note, the Guidelines do not address common enforcement fact patterns in digital ecosystems. The owner of a digital ecosystem will tend to have a high share of new applications or functionality within its ecosystem. But this can similarly be constrained by platform-to-platform competition.³

The Test for Exclusionary Conduct

The Guidelines adopts a two-limbed overarching test for distortions of effective competition or exclusionary abuses, namely conduct that:

1. Departs from competition on the merits.
2. Is capable of having exclusionary effects (¶13).

Both limbs are much debated. The concept of competition on the merits, while firmly rooted in EU case law, has little economic meaning. It may be best understood in the context of pricing abuses, where a firm acts against its economic self-interest (for example, sacrificing profit to predate a rival). It can also make sense for more obviously illegitimate business conduct — for example, misleading disparagement, misusing legal processes or breaching other laws to distort competition (¶72).

The test’s utility for conduct such as exclusivity, self-preferencing or refusal to deal is marginal. Those practices may denote healthy competition — or not, depending on the facts. Implicitly acknowledging these challenges, the Guidelines state the EC will not seek to separately prove nonmerit-based competition where the EU courts’ case law has already defined the test for abuse or explained why there is no merit-based competition at play.

The former includes predatory pricing, margin squeezing, exclusive dealing, tying and refusal to supply. The latter includes platform access restrictions, loyalty rebates and self-preferencing.

Conduct “capable of having exclusionary effects” refers to conduct that excludes or weakens competition. It need not be the sole cause of exclusionary effects, and the likelihood of effects occurring as a result may be sufficient. However, those effects must be more than hypothetical (¶99).

Proof of actual harm is not required:

• For pricing practices (such as predatory pricing, margin squeeze and conditional rebates), the focus is typically on the AEC test, i.e., whether the conduct could exclude a hypothetical competitor that is as equally efficient as the dominant firm (¶¶88-90).
• For nonpricing conduct, the AEC test is often not required, and the EC can rely on a broader range of evidence, including qualitative factors and actual market developments. The Guidelines find support in the recent Google Android judgment that qualitative factors can be particularly relevant in digital markets, where features like network effects and access to data may allegedly make the emergence of an equally efficient rival practically impossible (¶¶91-96).

The Guidelines advocate a sliding scale approach to enforcement. It states that “[t]he 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.” It proposes certain per se abuses (for example, paying to destock rivals), presumptively illegal conduct (such as exclusive dealing) where evidence of capability to exclude is unnecessary and conduct is holistically assessed for its exclusionary potential (such as self-preferencing) (¶54).

The sliding scale approach is a novelty not supported by the EU courts’ case law.

As for the factors considered in determining whether conduct has the capability to exclude rivals, the Guidelines list the following:

• Market position.
• Entry and growth barriers.
• The position of rivals.
• Scope of the conduct.
• Evidence of exclusionary intent — for which internal documents recording an undertaking’s intentions and expected effects remain critical — or observed effects (from market developments).

The Guidelines helpfully explain that while there is no de minimis threshold for abuse, “the higher the share of total sales in the relevant market affected by the conduct and the longer the duration of the conduct, the greater is the capability of the conduct to produce exclusionary effects.”

Narrower conduct may be illegal if it targets customer segments or rivals that are important for competition (¶106(d)-(e)). Conversely, the Guidelines explain that neither intent evidence nor actual effects are required to show abusive conduct.

Guidance on Specific Types of Conduct

The Guidelines seek to set down a consolidated summary of the EU courts’ analytical framework for specific abuses, while noting this is not a closed list and abuse may be determined based on first principles.

Predatory Pricing (¶¶108-122)

Predatory pricing arises where the dominant company prices drop below the appropriate measure of cost. Pricing below average variable costs (AVC) is presumed illegal but can also be unlawful below average total costs (ATC) where there is an eliminatory plan.

An eliminatory plan is generally established from internal document evidence. However, the Guidelines are prepared to infer an eliminatory plan from the circumstances, based on “the duration, the continuity and the scale of the below-cost sales, as well as the targeted nature of the below cost pricing and the importance of the market (segment) in which it takes place” (¶114, fn 234).

Pricing that covers marginal costs, even if below total costs, is typically pro-competitive and will commonly target more price-sensitive or contestable customers. Absent intent evidence, it is difficult to see how legitimate keenness in pricing can meaningfully be distinguished from an eliminatory plan.

The Guidelines seek to address cost measures in innovation-driven and digital business models, incorporating considerations that will increase the measure of cost, and consequently lower the bar for predation:

• Average avoidable costs (AAC) may be the appropriate measure in predation cases, for example, where the dominant company invested in new capacity or additional marketing to be able to implement the predatory scheme.
• Long run average incremental costs (LRAIC). This covers not just marginal output but also fixed costs of production, such as research and development (R&D). This can be appropriate, states the EC, where the predatory conduct covers multiple R&D cycles and, in an innovation lead market, companies have to make back the fixed costs of R&D to remain viable (¶115, fn 241).
• Double-sided markets. In doubled-sided platform markets, the consumer-facing product may be free, while the platform generates revenue from business-facing advertising or supplier fees. The EC may look at revenues and costs on both sides of the platform in those circumstances (¶120, fn 254).
• Opportunity costs. A further novelty in the Guidelines is the warning that the EC may bring opportunity costs into account. When loss-making sales could have been far more profitably sold elsewhere, the EC states it may add the “cost” of that lost opportunity to the costs side of the ledger (¶120, fn 253).

Margin Squeeze (¶¶123-135)

A margin squeeze arises where a dominant supplier’s wholesale prices prevent downstream retail rivals from competing. This occurs where, if it were required to pay the same wholesale prices, the downstream retail operations of the dominant company would not cover their LRAIC costs.

The Guidelines set out the EU courts’ recent case law on margin squeeze, in particular that it can be abuse regardless of whether there are alternatives to the supplier’s input, that the supplier is not dominant at the retail level or that the rival’s retail operations may, in fact, have a lower cost base than the dominant company (¶133).

Conditional Rebates (¶¶136-151)

Trading rebates, discounts or other advantages for customer volume commitments are a common, pro-competitive practice. As the Guidelines note, “they may stimulate demand and benefit consumers.”

This has historically been one of the most complex areas of Article 102 counselling. Older case law tended to rule formalistically that any volume-linked rebate not justified by cost-efficiencies was illegal.⁴ Similarly, rebates based on customers buying all or most of their needs, whether described as exclusivity or simply judged based on their historic/project volumes, were also prohibited.⁵

The 2009 Guidance sought to bring rebates under an economically coherent standard, the logic being that case law rigidity should not discourage price competition. It asked whether the conditions attached to the rebate effectively meant smaller rivals could not profitably win business.

This was the genesis of the AEC test, which gained judicial approval inter alia in the long-running Intel case.⁶ The European Court of Justice (ECJ) — the highest court in the European Union — ultimately ruled the EC could not conclude that Intel’s rebates were unlawful without addressing Intel’s AEC submissions that they did not foreclose AMD, the plaintiff.⁷

But AEC is a highly fact-sensitive appraisal. In Intel, conflicting evidence on how much demand rivals could win from Intel (the “contestable demand”) — materially different figures derived from both observed-switching and Intel and customer estimates — meant the EC had not carried its burden. The EC suffered a similar reverse in Android, with the General Court faulting the EC’s AEC test on Google’s handset incentives.⁸

The Guidelines adopt a more formalistic approach. In general, conditional rebates are considered holistically. The test will consider inter alia, market coverage, the extent of “must have” demand from the dominant company, the size of the rebates, whether it applies “retroactively” or only on incremental sales, whether it is linked to standardized or individualized volume targets, whether the rules of the scheme are arbitrary or clear, the duration of the scheme and whether there is an eliminatory plan (¶142).

The Guidelines delineate a number of subcategories of rebates depending on conditions applicable. Importantly, the Guidelines apply a stricter standard to rebate schemes linked to purchases of all or most of a customer’s needs (the exclusive dealing test) compared to those that are based on minimum volumes (the conditional rebates test) (¶139).

Mechanism ¶	Example	Cost Threshold
Incremental ¶147	10% rebate on units purchased after $100,000 threshold met (only applicable to units >$100,000 threshold)	Whether rebated price is below dominant company’s AAC or below LRAIC with an eliminatory plan (¶147)
Retroactive ¶148	5% rebate on all units if $100,000 threshold met (applicable to all units <$100,000 if threshold met)	Whether rebated price is below dominant company’s AAC (or below LRAIC with an eliminatory plan) in respect of effective price over the contestable demand. For example, if a contestable demand is 10% of the customer’s needs then the effective price is determined by dividing the 10% contestable share by the 5% rebate, giving an effective price of a 50% discount.
Multiproduct ¶150	10% rebate on units X, Y and Z if customer buys combined $100,000 target. X, Y and Z are available as stand-alones.	Whether applying aggregate value of rebate to each stand-alone product is below dominant company’s AAC (or below LRAIC with an eliminatory plan)9
All or most (>75%) of customer demand ¶152-155	(i) 5% rebate if customer buys 80% of needs (ii) 5% rebate if customer buys $100,000 (if this represents >75% of customer needs)	Presumed illegal unless company shows AEC defence as “retroactive rebates” Guidelines state AEC defence is not available for rebates linked to a specific percentage of demand (example (i)), but only where it is linked to minimum volume (example (ii))

 

The Guidelines warn the EC will scrutinize AEC submissions carefully. Evidential challenges, in particular as to “contestable demand” — such as those that caused reversals in Intel and Android — will be construed against the dominant company (¶149, fn 294).

Applying different, and stricter, standards to certain rebate schemes is questionable as a matter of both case law and economics. In Intel, to the contrary, the ECJ faults the EC for asserting there is a legal distinction between types of loyalty rebates, holding there is no per se category of illegal rebates.¹⁰ Rather, they must be judged on their facts and, if raised during the administrative proceedings, any AEC defence put forward by the dominant undertaking.¹¹

Additionally, as a matter of economics, there is no difference in the assessment of potentially exclusionary effects based on which categorization applies.¹² Moreover, as a practical matter, rebate schemes tend to be on a spectrum. It may not be clear to the supplier that the target volume in fact represents all or most of the customer’s demand, and the volume will change year to year. Rebate schemes are also typically based on many graduated steps (at least one of which may come close to the customer’s needs), not “one shot” targets.¹³

The a priori exclusion of an AEC analysis for rebates linked to a high proportion of customer demand contradicts the ECJ’s Intel judgment that applied the AEC test specifically to rebate schemes said to be based on 80% or more of customer demand.¹⁴

One consequence of this categorization is that it tilts the evidential balance against volume-linked rebates. The AEC test is evidence-heavy, requiring a close assessment of costs, contestable share and customer demand. Accordingly, companies may well err on the side of declining to grant otherwise pro-competitive rebates.

Exclusive Dealing (¶¶151-161)

Exclusive dealing can be abusive where it forecloses rivals’ access to the market. Here, the Guidelines apply a presumption of harm. The Guidelines define exclusivity broadly to include de jure and de facto exclusivity both via contractual obligations and pricing (¶¶151-154).

Requirements of more than 75% of a customer’s needs are also considered exclusive dealing, whether expressed in percentage terms or as volumes that coincide with that share of the customer’s business.

More recent ECJ precedent takes a more nuanced approach to exclusive dealing, recognizing there can be circumstances in which it does not harm competition.¹⁵ The Guidelines confirm that exclusivity does not restrict competition where:

• It has a limited impact in terms of market coverage or duration (Google Adsense);
• It is simply hypothetical, because the customer has no alternative choices (Qualcomm); or
• Rivals may be able to compete for the customer’s entire demand (¶157).

The decisional practice has seen a number of cases where the EC has accepted exclusivity. Typically, these are “all or nothing” type markets — business energy supplies, soft drinks for restaurants, sports collectibles or supplies agreements — where customers are unlikely to choose multiple suppliers, but which offer rivals a fair opportunity to compete for the business over time.¹⁶

The Guidelines accept that AEC evidence may show that exclusive dealing does not foreclose "as-efficient" rivals. However, they signal the EC’s enforcement intent by heavily circumscribing the circumstances in which AEC evidence will be probative, noting that “in the Commission’s experience, [the AEC] test is generally not relevant for a range of exclusive dealing scenarios.”

These include:

• Exclusivity obligations or rebates conditioned on the purchase of a minimum share of the customer’s needs, rather than a minimum volume (¶158, fn 312).
• Exclusivity across a fragmented customer base deprives rivals of access to a sufficient number of customers to gain the necessary economies of scale to compete (¶158, fn 312).
• Where an “as-efficient-competitor" is not the appropriate benchmark, because regulatory barriers to entry limit the contestability of the dominant company (for example, rivals cannot access the customer base of a regulated utility) or, in the digital sector, “where features such as innovation, access to data, multi-sidedness, user behaviour or network effects play a decisive role” (¶¶94-95,158).

Even where the presumption has been rebutted, the Guidelines state that the EC will still conduct a holistic assessment of the dominant company’s position, the scope of the conduct, the mechanics of the exclusive dealing arrangement and its potential exclusionary effects (¶161).

The three-part burden shifting — presumption, rebuttal, holistic assessment — is not one supported by the case law and signals heightened scrutiny of exclusive dealing. In particular, as noted above, the exclusion of AEC rebuttal evidence for rebate schemes linked to a high percentage of customer demand is a stricter legal standard than the Intel case law requires.

Tying and Bundling (¶¶162-178)

Tying or bundling may be an abuse where the dominant company leverages its power in a dominant tying product to exclude rivals from competing for the tied product. Tying requires:

• Distinct products.
• A dominant tying product.
• A prevention of stand-alone access to the tying product (“coercion”).
• Potential exclusionary effects.

The Guidelines recognize tying can be pro-competitive, leading to “efficient product integration, lower production costs and lower transaction costs, as well as increased convenience and quality improvements for consumers.” But they take the opportunity to consolidate their decisional practice and the EU courts’ case law, and update it to cover digital cases and the app economy.

The Guidelines codify the EC’s digital markets decisional practice on when tying and exclusionary effects occur. A “tie” may be alleged where a product is distributed with other apps through preinstallation in a digital platform. Customers are “coerced,” the Guidelines state, even where the tied product is free, can be uninstalled and/or rivals’ alternatives can be acquired without cost (¶¶174, 177).

The EC will consider whether platform bundling “confers a significant competitive advantage … rather than the quality of the tied product, where that advantage is unlikely to be offset by competitor” (¶178(a)).

The latter may be the case where the distribution penetration advantage cannot be offset by rivals via other channels. Similarly, even where a product bundle is made available for free and can be uninstalled, the lack of rivals observed over time may suggestion exclusionary effects. This may be inferred, according to the Guidelines, from the broad common customer base for the products (typically a given for a preinstalled bundle of apps), market power in the tied product, barriers such as “networks effects … in digital markets,” and consumer inertia or bias (¶178).

It may be questioned whether poorly defined criteria such as “competitive advantage,” “consumer inertia” or “status quo bias” should be factors in determining unlawful tying. They materially broaden the category of tying abuse to encompass common, and legitimate, fact patterns in digital markets arising from first mover advantages or switching inertia.

An ex post prohibition is ill-suited to regulating these market features. Indeed, in cases involving status quo bias, the EU courts took into account the existence of broader exclusive arrangements in the sector as reinforcing exclusionary factors.¹⁷

Rather, if there is market failure inter alia because of these features of a market, an ex ante regulatory regime — for example, consumer protection or gatekeeper laws such as the DMA — would generally be the more appropriate legal mechanism.

Access Restrictions (¶¶179-182) and Refusal to Deal (¶¶183-189)

Historically, EU law set a high bar for finding cessation or refusal of supply to be an abuse. The case law recognized that forcing the dominant company to share the fruits of its investments engaged constitutional issues of property rights and freedom of contract, and could discourage investment and innovation.

The test thus required that the denied input:

• Be indispensable.
• Eliminated all effective competition.
• The recipient was not just seeking to replicate the dominant company’s products, but to create new or innovative ones.¹⁸

More recently, the EU courts have limited the scope of “pure” refusal to deal cases, applying a lower bar where access is only available on unacceptable terms or refused despite the open design architecture of a platform. The Guidelines codify the movement in the case law and signal areas of enforcement focus.

Access restrictions can be unlawful, the Guidelines state, regardless of whether the input is indispensable, simply where the restriction has an exclusionary effect on competitors (¶¶180-181). This can be the case where interrupting supplies restricts downstream rivalry or, in the digital economy, where access is restricted to open digital platforms designed for third parties. The same is true if the dominant digital platform provider delays or sets unfair terms for access (¶¶182).

Conversely, the stricter criteria of “indispensable input” and “eliminating all effective competition,” the EC states, applies only for refusal to deal, where there’s no preexisting relationship (¶¶185-187). The still stricter “new products” criterion is reserved only for refusal to license intellectual property (IP) cases, in the Guidelines’ view (¶189).

Conduct ¶	Example	Test
Access restriction ¶¶180-181	Ceasing raw material supplies Refusing or degrading access to open architected platform Unfair or restrictive access terms	Refusal has the capability to exclude competition
Refusal to deal ¶¶185-187	Refusal to supply (no prior relationship) Refusal of network or platform access (not open design)	(i) Indispensable input (ii) Eliminate all effective competition
Refusal to license ¶189	Access to software, interface information or other IP	(i) Indispensable input (ii) Eliminate all effective competition (iii) New or improved product (not copying dominant company)

 

The differential standards for similar conduct is open to debate. The ECJ’s Bronner formulation was generally considered to have created a unified refusal to deal standard, albeit Bronner has been since distinguished in digital cases.

There are likely to be many edge cases in which access denial cuts across many of these categories. Given the importance of maintaining innovation incentives, particularly in fast-moving markets, these standards will likely be litigated.

Self-Preferencing (¶¶190-196)

Again signalling its future enforcement priorities, the Guidelines create a new species of “self-preferencing” abuse typically applicable to dominant digital platforms or ecosystems (¶¶190-192). Self-preferencing may occur in preferential display, consumer traffic or digital auctioning favouring the dominant platform/ecosystem owner (¶194).

Self-preferencing will likely involve unlawful exclusionary effects, the Guidelines state, where the platform is an important source of business, the self-preferencing is not merits-based and the outcome will likely influence user behaviour to favour the dominant platform owner regardless of the merits of its products (¶196).

Differential treatment will generally be inferred where the platform is designed or holds itself to take a neutral position between its business users, or neutrality is considered the market standard (¶196, fn 404-405).

The breadth of the Guidelines’ self-preferencing category is wider than the narrower fact patterns in the case law. Platforms adding an attractive new functionality is a common, pro-competitive business occurrence. The ECJ recognises this when it states that “it cannot be considered that … a dominant undertaking which treats its own products or services more favourably than it treats those of its competitors is engaging in conduct which departs from competition on the merits.”¹⁹ Rather, plus factors would need to be present to demonstrate that conduct is exclusionary.

The implications of the Guidelines commentary is also likely to be contentious, particularly when set alongside the expansive guidance on access restrictions. The two novel abuse fact patterns entail both:

• Allowing rivals access to a platform.
• Requiring treatment just as beneficial as the platform owners’ own products.

In combination, the two rules would seem to create material disincentives to engage in pro-competitive investment in innovation and platform improvements.

Automatically Unlawful Conduct (¶¶198-201)

The Guidelines state that certain conduct may be so clearly anticompetitive that no analysis of its potential to restrict competition is required; these are the so-called “naked restrictions” described in the 2024 draft guidelines.

This may be the case where the dominant company pays to have its rivals destocked, dismantles the infrastructure on which rivals depend to compete or uses its quasi-regulatory powers — for example a professional association that also acts as rule maker for the sector — to prevent competition. The EC supposes that challenges to such a finding could succeed only “very exceptionally.”

Objective Justifications

Where the EC demonstrates that the conduct distorts effective competition, the undertaking can show that the conduct is objectively justified. This would be the case where the conduct is either objectively necessary to achieve a legitimate aim and proportionate to that aim, or the conduct gives rise to efficiencies that outweigh alleged harm without removing sources of actual or potential competition (¶220).

The dominant company has the burden of proof and can invoke technical reasons, such as the need to maintain or improve product performance, as an objective necessity. However, when it comes to interoperability, simply pointing to the difficulty of developing a technical solution will not suffice unless granting access would genuinely compromise the product’s integrity or security, or would be technically impossible.²⁰

Claimed efficiencies must be objective, concrete and verifiable, substantiated where possible by contemporaneous documents, financial data, expert studies or economic models, and quantified as precisely as reasonably possible (¶¶223-225). The Guidelines also recognise efficiencies arising in a related market, where the consumers harmed by the conduct and the consumers benefiting from the efficiencies substantially overlap (¶¶236-238).

However, the greater the conduct’s potential to harm competition, the less likely an efficiencies defence is to be successful.

The Guidelines signal that sustainability-driven conduct by dominant companies can, in the right circumstances, be defended on efficiency grounds — an important aspect for businesses looking to align competition compliance with their environmental, social and governance (ESG) strategies.

The Guidelines recognise that sustainability benefits can qualify as qualitative efficiencies, e.g., where a dominant company’s conduct enables the use of fewer raw materials, cleaner production or distribution methods, more recyclable products, more resilient infrastructure, reduced supply chain disruption, or faster time-to-market for sustainable goods.

In addition, sustainability gains can translate into direct consumer benefits in the form of lower costs — for instance, where more sustainable products can be produced or distributed more cheaply. The availability of cheaper sustainable alternatives also creates a positive externality if it encourages consumers to switch away from less sustainable (e.g., polluting) products.

The Guidelines offer guidance on what the EC will accept as a business justification, whether grounded in objective necessity or genuine efficiencies. It remains to be seen whether this opens the door to a less sceptical review of justifications. To date, the EC has almost always rejected even established benefits as insufficient or judged that the benefits could have been obtained through a less restrictive business model.²¹

Given that the Guidelines limit economic analysis (for example, the AEC test) at its first-stage analysis of potential restriction, greater openness to economic justifications at the objective justification stage would be the logical corollary. Particularly in digital platforms and ecosystems, the benefits to consumers of new business models are well attested. Any restriction in the rules that have enabled them to grow and function properly should be weighed against their benefits.

Looking Ahead

Businesses may want to monitor how the EU courts respond to several key elements of the Guidelines. The EC’s proposed allocation of the burden of proof across different categories of conduct, and the legal tests it sets out for specific practices, may not survive judicial challenge. The presumption that certain types of conduct are inherently capable of producing exclusionary effects is likely to be a particular flashpoint in future litigation.

Similarly, the EC’s reliance on selected case law to frame the applicable legal standards remains open to debate. Until the courts have ruled, there is significant uncertainty around the precise boundaries of these new enforcement standards, and dominant undertakings should factor that uncertainty into their compliance strategies.

_______________

1 Press release of the European Commission IP/97/868 of 10 October 1997: The European Commission accepts an undertaking from Digital concerning its supply and pricing practices in the field of computer maintenance services; COMP/39.692 - IBM Maintenance Services, Decision of the European Commission of 13 December 2011; AT.40823, SAP ER Aftermarket Support Services (Case AT.40823), Commission Decision of 9 July 2026.

2 Case C-56/12 P EFIM v Commission EU:C:2013:575 [12].

3 U.S. Epic v. Apple, ND Cal pp. 46-47 (10 September 2021) (Consumers choose between mobile ecoystems.); Compare, Case AT.40437 – Apple – App Store Practices (music streaming), Commission Decision of 4 March 2024, paras. 415-505 (Competition for mobile devices does not constrain music streaming aftermarkets).

4 Case C-95/04 P British Airways plc v Commission ECLI:EU:C:2007:166 [86]-[87]; Case 322/81 NV Nederlandsche Banden Industrie Michelin v Commission ECLI:EU:C:1983:313.

5 Case 85/76 Hoffmann-La Roche & Co. AG v Commission of the European Communities ECLI:EU:C:1979:36; Case C‑549/10 P. Tomra Systems ASA and Others v Commission ECLI:EU:C:2012:221; Case T-851/14 Slovak Telekom a.s. v Commission ECLI:EU:T:2018:929.

6 Case C-240/22 P Commission v Intel Corporation Inc EU:C:2024:915 [181]; see, also, Case C-680/20 Unilever Italia Mkt.Operations Srl v Autorità Garante della Concorrenza e del Mercato EU:C:2023:33 [54]; Case C-377/20 Servizio Elettrico Nazionale SpA v Autorità Garante della Concorrenza e del Mercato EU:C:2022:379 [80]-[82]; Case C-738/22 P Google and Alphabet v Commission ECLI:EU:C:2026:533 [270, 272].

7 Case C-240/22 P Commission v Intel Corporation Inc EU:C:2024:915 [181].

8 Case T-604/18 Google LLC and Alphabet Inc v Commission EU:T:2022:541 [798]-[799].

9 The Guidelines assume bundled rebates are offered across products in fixed quantities (e.g., 10% rebate if you buy one of each of X and Y). That is rarely the case in practice. Rebate schemes tend to be based on a dollar value of products from multiple categories, with no obligation on the customer to purchase from a specific category. In that fact pattern, the better approach is to use the AEC test for retroactive rebates, since any potentially leveraging effect of a bundled rebate without fixed bundled quantities can be quantified in the same way. Typically, the contestable share will tend to be commensurately smaller for a rival that sells only one of the products in the bundle, and so the value of any rebate would have to be lower (or given away gradually over smaller steps in the rebate scheme) to pass the AEC threshold.

10 Case C-240/22 P European Commission v Intel Corporation Inc. ECLI:EU:C:2024:915 [136-139, 144-146, 180-181].

11 Ibid. [144]

12 “The Velux Case: An In-Depth Look at Rebates and More,” Svend Albaek and Adina Claici, Commission Competition Policy newsletter 2 (2009), pp45-47.

13 The Guidelines state that to be exclusive dealing, the dominant supplier must intend to capture all or most of the customer’s demand (¶154). But this seems to be a distinction with little economic meaning. A supplier will naturally want to secure as much business as possible.

14 Case C-240/22 P Commission v Intel Corporation Inc. ECLI:EU:C:2024:915 [181].

15 Case C-680/20 Unilever Italia Mkt Operations v Autorità Garante della Concorrenza e del Mercato EU:C:2023:33 [51]; Case T-334/19 Google LLC and Alphabet Inc v European Commission EU:T:2024:634, [384].

16 Case COMP/A.39.116/B2 Coca-Cola Commission Decision 2005/670/EC [2005] OJ L253/21; Case COMP/B-1/37.966 Distrigaz Commission Decision [2008] OJ C9/8; T-699/14 Topps Europe v Commission ECLI:EU:T:2017:2; Case COMP/E-2/38.316 Vega SpA/CIK-FIA Commission Decision of 19 March 2004 [78-79]; Case T-515/18 Fakro v Commission ECLI:EU:T:2020:620 [186-187] (Exclusivity did not prevent rivals accessing inputs from other suppliers).

17 See “Google Android, the Final Chapter: Court of Justice Clarifies the Standard for Exclusionary Abuse in Digital Markets,” Skadden, 8 July 2026.

18 C-7/97 Oscar Bronner GmbH & Co.KG v Mediaprint ECLI:EU:C:1998:569 [41]; C-418/01 IMS Health GmbH & Co. OHG v NDC Health GmbH & Co. KG ECLI:EU:C:2004:257 [37-38].

19 C-48/22 P Google and Alphabet v Commission (Google Shopping), ECLI:EU:C:2024:726, [186-187].

20 Case C‑233/23 Alphabet and Others v AGCM EU:C:2025:110 [73-81].

21 Case T-612/17 Google v Commission (Shopping) ECLI:EU:T:2021:763 [266, 284-293]; ECJ judgment, C-48/22P Google LLC and Alphabet Inc. v European Commission ECLI:EU:C:2024:726 [187, 223-230] (Benefits of the search box considered insufficient to outweigh restraint); Case C-333/21 European Superleague Company, SL v Fédération internationale de football association (FIFA) and Union of European Football Associations (UEFA) ECLI:EU:C:2023:1011 [209] (Prior approval of third-party competitions could only be objectively justified “if it is demonstrated, through convincing arguments and evidence, that all of the conditions required for those purposes are satisfied.”)

This memorandum is provided by Skadden, Arps, Slate, Meagher & Flom LLP and its affiliates for educational and informational purposes only and is not intended and should not be construed as legal advice. This memorandum is considered advertising under applicable state laws.