Short answer
These practices are not automatically unlawful. Article 102 TFEU requires dominance in a properly defined market, possible effects on trade between Member States, and conduct outside competition on the merits that is capable of excluding rivals or obstructing entry. The legal template and evidence depend on the practice.
A genuine refusal over self-developed infrastructure usually faces the demanding Bronner indispensability test. Tying examines distinct products, customer choice, foreclosure and justification. Exclusivity and rebates require evidence of exclusionary capacity; an as-efficient-competitor test can help but is not always mandatory. Predatory pricing turns on costs and, in the intermediate cost range, an exclusionary plan. Self-preferencing requires contextual assessment.
Classification is especially important for access disputes: Bronner does not automatically govern platforms designed for third parties or discriminatory access conditions. Sector rules such as the DMA may impose additional duties without replacing Article 102 analysis.
Each route therefore starts with careful classification of the conduct.
Article 102 protects the competitive process, not every competitor
The Treaty prohibits abuse of dominance, not dominance itself. Its examples include limiting markets or technical development, discriminatory conditions and contractual tying:
“Any abuse by one or more undertakings of a dominant position within the internal market or in a substantial part of it shall be prohibited as incompatible with the internal market in so far as it may affect trade between Member States.”
—
The assessment must define the market, establish dominance and identify the counterfactual and possible exclusion. Market power or control of an important input does not itself prove abuse.
The Court of Justice's 2024 judgment in Case C-48/22 P, Google and Alphabet v Commission (Google Shopping), states the general exclusionary test:
“it is necessary, as a rule, to demonstrate, through the use of methods other than those which are part of competition on the merits between undertakings, that that conduct has the actual or potential effect of restricting that competition by excluding equally efficient competing undertakings from the market or markets concerned, or by hindering their growth on those markets”
—
The Court requires the appropriate conduct-specific test and tangible evidence of at least potential exclusion. Foreclosure need not be complete, but a rival's difficulty alone does not prove harm.
When can self-preferencing be abusive?
Self-preferencing favours the dominant firm's downstream offering over rivals using its platform or gateway. Article 102 does not require every dominant firm to treat its own and rivals' products identically.
In Google Shopping, the abuse combined more favourable display of Google's comparison-shopping results with algorithmic demotion of competitors. The Court considered the importance and limited substitutability of search traffic, market characteristics and potential exclusion. It did not make every product-design advantage abusive.
For designated gatekeepers, Article 6(5) of Regulation (EU) 2022/1925, the Digital Markets Act (DMA), separately prohibits more favourable ranking, indexing and crawling of their own offerings. The DMA operates without prejudice to Article 102, so the two findings are not interchangeable.
Tying requires distinct products, coercion and foreclosure capability
Tying occurs where customers wanting one product must also accept another. Article 102(d) expressly addresses supplementary obligations unrelated by nature or commercial usage to the contract's subject, but abusive bundling is not confined to that example.
In Case T-201/04, Microsoft v Commission, the General Court asked whether there were distinct products, tying-market dominance, no choice to obtain the tying product alone, foreclosure and no objective justification. Separate demand and independent suppliers can show distinct products; technical integration or a zero separate price is not decisive.
In , decided on 2 July 2026, the Court dismissed Google's appeal and reaffirmed those criteria for app pre-installation ties. Foreclosure was assessed in the full economic and legal context.
Ask whether customers are compelled to take the tied item and whether the tie can weaken rivals' market access. Claimed integration, security or efficiency benefits require evidence and proportionality.
Refusing data or access faces a demanding threshold
A firm is generally free to choose its trading partners and use assets it developed for its own business. For a pure refusal to grant a rival access to such infrastructure, the Bronner conditions protect that freedom and the incentive to invest:
“not only [must] the refusal of the service … be likely to eliminate all competition … and … be incapable of being objectively justified, but also … the service in itself [must] be indispensable … inasmuch as there is no actual or potential substitute in existence”
—
Being cheaper or more convenient is insufficient. The applicant must address realistic technical, legal and economic alternatives at a comparable scale. For intellectual property, prevention of a new product may also matter.
Classification matters. Google Shopping confines Bronner to a genuine refusal that would force the owner to contract, not differentiated treatment or another independent practice. For data, identify what is requested, who created it, substitutes, interoperability routes and technical feasibility.
Where infrastructure was designed for third parties, the boundary changes. In , Bronner indispensability did not apply because Android Auto was not developed solely for Google's business. Refusing interoperability could be abusive even if the app worked elsewhere, subject to exclusionary capability and possible security or technical justification.
adds that Bronner can cover publicly developed infrastructure competitively transferred to an undertaking with full access autonomy. Regulatory access duties point the other way.
Exclusivity and rebates require evidence of exclusionary capacity
Exclusive purchasing can deny rivals enough demand to compete. Loyalty rebates can similarly penalise switching a contestable share by withdrawing discounts on a larger purchase block. Genuine cost-based quantity discounts differ.
In Case C-680/20, Unilever Italia, the Court rejected an irrebuttable shortcut for contractual exclusivity:
“a competition authority is required, in order to find an abuse of a dominant position, to establish, in the light of all the relevant circumstances … that those clauses are capable of restricting competition.”
—
Relevant circumstances include market position, demand coverage, duration, switching, entry barriers and likely effects. An as-efficient-competitor test compares the effective price a rival must match for contestable demand. It is optional, but an authority must assess its probative value if the dominant undertaking submits one. The precise test remains sensitive to design and evidence.
Predatory pricing uses cost thresholds, not low price alone
Low prices ordinarily benefit consumers. They become predatory where a dominant firm deliberately incurs losses or sacrifices profits through prices capable of excluding competitors, rather than competing on efficiency.
The Court's judgment in Case C-62/86, AKZO v Commission, established two principal cost bands. Prices below average variable cost are prima facie abusive because each sale fails to cover the costs attributable to producing it. Prices below average total cost but above average variable cost are abusive when fixed as part of a plan to eliminate a competitor. Proper cost allocation, the relevant period and evidence of strategy are therefore critical. A temporary promotion or meeting competition is not assessed merely by observing a low headline price.
In Case C-202/07 P, France Télécom v Commission, the Court confirmed that proving a realistic possibility of later recouping the losses is not a necessary precondition under EU law. Recoupment evidence may still help explain whether the conduct has an exclusionary rationale, but its absence does not automatically defeat the case.
A practical decision framework
For any of these practices, ask:
- What markets and evidence establish dominance at the material time?
- Is the conduct a refusal, discrimination, tie, exclusive term, rebate or below-cost price?
- Which rivals, entrants or adjacent markets can it affect, and how?
- What is the realistic counterfactual, coverage and duration?
- Does the legal template require indispensability, cost evidence or coercion?
- What specific evidence shows actual or potential exclusion?
- Is an evidenced necessity or efficiency proportionate?
- Does a sector rule such as the DMA apply separately?
Conclusion
The label “self-preferencing”, “tie”, “refusal”, “exclusive”, “rebate” or “cheap” is not decisive. Abuse arises when the conduct-specific test and Article 102 framework show, on specific evidence and in context, methods outside competition on the merits capable of excluding effective competition or blocking entry. Bronner sets a high threshold for compelled access; tying requires distinct products, coercion and foreclosure capability; exclusivity and rebates demand effects-based scrutiny; and predatory pricing turns on costs and, in the intermediate band, exclusionary strategy.
Sources
This article provides general legal information, not advice for a specific matter. Laws, facts and procedural deadlines may change the result. Follow the linked primary sources and obtain local advice before acting.