When Machines Decide the Price: Algorithmic Pricing Emerges as Enforcement Priority for EU & UK Antitrust Regulators

Antitrust laws across the European Union and United Kingdom prohibit agreements or concerted practices that restrict competition. Under long-established EU and UK competition law, parties to such agreements may face severe penalties, including fines of up to 10% of global turnover. In addition, individuals involved in the infringement can be subject to personal fines, director disqualification, and, in some jurisdictions, imprisonment. 

Algorithmic pricing is relevant in this context because interoperable or shared software can potentially serve as a conduit for anticompetitive price coordination. 

EU and UK antitrust enforcement in this area is no longer merely a theoretical concern. The European Commission (EC) and the UK Competition and Markets Authority (CMA) have both identified algorithmic pricing – including tools used by competitors to generate or benchmark prices – as a priority enforcement area. In July 2025, a senior EC official confirmed multiple ongoing investigations into suspected algorithmic collusion. In September 2025, the CMA’s chief executive described the practice as an “area of focus and concern”, noting the agency is “watching and learning” from US cases and exploring links with generative artificial intelligence. 

What is Algorithmic Pricing 

Algorithmic pricing refers to the use of software to set prices for goods and services. Algorithms can adjust prices at high speed and with great precision, allowing firms to respond to market changes in ways that manual pricing cannot. While algorithmic pricing can improve efficiency and help firms stay competitive, it raises important legal questions related to situations, for example, where multiple firms that may be deemed competitors pool certain data or use certain software in common. As the technology continues to develop, future questions may also arise involving potential scenarios in which algorithms learn to coordinate with one another or produce pricing outcomes that resemble collusion, even without explicit communication between different firms. 

Historical Precursor: The EU Eturas Case as a Template for Algorithmic Collusion Analysis 

Although algorithmic pricing is a relatively new practice, the Court of Justice of the European Union’s (ECJ’s) 2016 ruling in Eturas (Case C-74/14) provides an example for how competition law risk can materialise when competitors use the same software platform.  

In Eturas, multiple travel agencies in Lithuania used a common online booking system called Eturas. The platform operator introduced a technical restriction on the travel agencies’ ability to offer discounts to consumers and sent a message notifying the travel agencies of this change. The ECJ held that this would constitute a concerted practice under Article 101 TFEU if it were proven that the travel agencies were aware of that message.  

As antitrust regulators confront algorithmic pricing cases, Eturas shows that shared pricing software can create a concerted practice where a “traditional” agreement between firms is absent. 

US Private Enforcement as a Proxy for Future EU/UK Risks 

While EU and UK regulators have raised algorithmic pricing as an area of enforcement focus during 2025, US private enforcement over the past few years provides a preview of how antitrust risk may materialise. Common patterns in recent US cases include plaintiffs’ allegations of the following: 

• Competitors relying on the same pricing or revenue-management application 

• The application ingesting disaggregated, current, or forward-looking non-public, competitively sensitive information (CSI) 

• The application generating price recommendations or competitor (price) benchmarks 

• Competitors adopting those outputs, leading to reduced price competition 

To date, notable US litigation examples include: 

• The RealPage Antitrust Litigations – The Department of Justice, state attorneys general and class plaintiffs have alleged in various lawsuits that RealPage’s software collected confidential, competitively sensitive data (such as future occupancy, lease terms, and rental rates) from competing landlords and used that information to recommend price increases. These recommendations were allegedly routinely accepted by landlords. While courts are considering proposed settlements in some of the cases, many of the claims remain live. Of note, in the RealPage case pending in the Middle District of Tennessee, the court suggested that the novelty of alleged agreements involving computerised algorithmic pricing rendered per se treatment inappropriate.   

• Duffy v. Yardi Systems – This class-action suit alleges that Yardi’s software facilitated similar coordination for Yardi’s RENTmaximizer software. Notably, the court in the Western District of Washington disagreed with the RealPage court in Tennessee, stating that an agreement accomplished by algorithm should be subject to per se treatment. 

• Hotel Litigations (Gibson v. Cendyn and Cornish v. Caesars) – Two separate class actions were brought against Cendyn and large hotels alleging that Cendyn’s software pooled competitor information and the hotel defendants followed pricing recommendations from the software that increased hotel room rates. Both of these cases were dismissed for their failure to plead sufficient allegations. The Ninth Circuit affirmed the dismissal and the appeal in the Third Circuit is currently pending. 

How EU and UK Regulators Are Tackling Algorithmic Pricing: Legal Framework and Theories 

European regulators are expected to treat the existing competition laws as broad enough to deal with algorithmic pricing concerns with respect to competition law. Under Article 101 TFEU in the European Union and Chapter I CA98 in the United Kingdom, regulators can pursue conduct that constitutes coordination between competitors, regardless of whether a traditional “meeting of minds” is evidenced.  

Regulators are expected to rely on a range of established legal theories, which they may adapt to the algorithmic context: 

Concerted practices 

Regulators can try to characterise algorithmically facilitated coordination as a concerted practice where competitors achieve a form of alignment through their parallel use of a pricing tool, even if there is no direct agreement or communication, as is evident from Eturas. This theory is likely to be central to regulators’ inquiries where firms adopt similar pricing outputs because they use the same pricing software. 

Hub-and-spoke liability 

Algorithms can potentially act as a functional “hub” in some circumstances in which an intermediary channels CSI from one competitor to another. Regulators may use hub-and-spoke principles to argue that a pricing software provider facilitates indirect horizontal collusion, particularly if the tool aggregates or benchmarks competitor data in ways that predictably align prices. 

CSI input risks 

Providing CSI into a shared platform may itself be potentially problematic. EU case law recognises that unilateral disclosure of sensitive information, where the disclosing firm can reasonably expect that rivals will use it to guide competitive conduct, can form part of a concerted practice. Transposed to algorithmic tools, regulators may argue that a firm entering non-public, current, or forward-looking data into a third-party algorithm should have known that the tool would share this information with other users, depending on the circumstances.  

Software provider liability 

Software providers themselves may face scrutiny. EU law permits enforcement against facilitators that knowingly contribute to coordination between competitors. A software provider could therefore be potentially liable if it is found to knowingly design tools in ways that encourage aligned pricing, or advertises the tool as capable of producing higher or more stable industry-wide pricing. Regulators may also consider whether the software provider failed to implement adequate firewalling, aggregation, or other safeguards. 

While these theories come from traditional antitrust doctrine, their application to algorithmic pricing is still evolving. The key open question is whether the existing regulatory toolkit can meaningfully capture the unique features of modern AI-driven pricing systems – especially their opacity, speed, and potential ability to internalise large amounts of data without human involvement. EU and UK enforcement activity in the coming years will likely test the boundaries of these legal concepts and determine how far they can stretch to meet the challenges posed by algorithmic coordination. 

Mitigating Risk 

While any business will have to analyse how the relevant laws apply to their own particular circumstances and activities, companies using or contemplating adopting algorithmic pricing tools in the European Union and United Kingdom should consider whether it is appropriate to take proactive steps to manage risk, including the following: 

• Conduct robust due diligence. Assess third-party pricing tools to understand precisely how they operate: What data do they ingest? Do they incorporate competitor CSI? What output do they create? Clear explanations from software providers regarding data sources, algorithm logic, and any mechanisms designed to prevent the transmission of CSI are beneficial to understand. 

• Exercise legal caution before adoption. Before implementing an algorithmic pricing tool that may also be used by other businesses that could be deemed direct competitors, firms should consider seeking legal advice. 

• Manage CSI inputs carefully. Firms should exercise due caution before providing non-public, current, or forward-looking CSI (such as pricing, capacity, volumes, or customer information) into third-party applications. If there is a specific business need to supply non-public information to an algorithmic tool, assess any appropriate safeguards. These may include data aggregation, time lags, minimum market-width thresholds, or other design features that reduce the risk of revealing granular competitively sensitive information. Appropriate thresholds may vary depending on industry characteristics, market structure, and geographic scope. 

• Document governance and procompetitive intent. As appropriate, maintain clear internal documentation that records the intended use cases of pricing tools, the categories of data provided, exclusions applied, relevant model settings, and any internal or software provider-side safeguards. Documented evidence of procompetitive benefits (such as improved demand forecasting, lower prices, or increased output) may also be valuable for explaining how the tool may benefit competition. 

• Train personnel and embed compliance. Consider appropriate training for employees involved in pricing and data operations on the legal risks associated with algorithmic pricing tools, including the implications of sharing CSI or relying on outputs derived from data pooled across competitors. Training should emphasise that pricing decisions must ultimately remain independent and uncoordinated. 

Taken together, these measures may help firms reduce exposure under EU and UK antitrust rules while enabling them to harness the legitimate efficiencies offered by modern pricing technologies.