Overview
Overview
In the earlier installments of our antitrust and artificial intelligence series, we looked at AI and anticompetitive agreements and then at the unlawful unilateral use of AI. We have outlined some of the challenges faced by antitrust enforcers in determining whether AI has illegally facilitated anticompetitive agreements in areas such as market allocation, customer division, or information exchanges. The use of AI by businesses for example, for modeling customer preferences, market density, or vertical management of sales channels reveals risks which antitrust enforcers on both sides of the Atlantic have recognized and which resulted in the issuance of a joint warning in July 2024 of potential collusion among rivals 'steering market outcomes.1
In this installment, we will look at various responses by enforcers to the unilateral deployment of AI tools by companies having a strong market presence. Digital markets are fast moving, AI tools are rapidly adaptive, and nuanced targeting can be opaque and hard to detect, but nonetheless accretive to market shares and harmful to rivals and to consumer welfare. Antitrust enforcers in the US, UK, and EU are well aware of these challenges, although they have different enforcement tools and are bound by different tests and thresholds for dominance/monopolization.
A market share of 50% in the EU is presumptively dominant, but potentially insufficient for a completed monopolization offense in the US. On the other hand, US law recognizes an offense for attempted monopolization, referring to anticompetitive actions taken by a person or business with the specific intent to monopolize and a dangerous probability of achieving monopoly power. This concept is often used in antitrust law to describe efforts that may not have succeeded yet but pose a significant and current threat to competition. Enforcers around the world are aware that a company’s market presence may fall short of dominance/monopolization, but its use of AI tools may be designed to reinforce and strengthen its market position to a tipping point into dominance.
1. Challenges in Defining Markets in the Digital Economy
The growth of the digital economy brings new challenges for defining the relevant market, given the unique characteristics of digital markets and business models. Traditional market definition relies on formalistic standards, which can be difficult to apply and test when available information is limited. In the age of AI, digital platforms with deep access to diversified data (e.g., tracking usage trends, from customers, rival downstream providers and third party apps), are likely to hold significant market power, giving them a superior bargaining position over business partners and enabling potentially unfair trading terms and practices (to enhance their own offerings and increase costs to rivals). We considered this in our previous post looking at, for example, the Amazon Marketplace, Google Android, and Google Search cases. However, proving monopolization under Section 2 of the Sherman Act or dominance under Article 102 TFEU is challenging. The multi-sided nature of platforms and their innovative business models complicate market definition and assessment of lasting market power. In addition, AI tools may deploy algorithms that work autonomously, may micro-target customers, and work at such speed and scale, that it becomes almost practically impossible for enforcers to effectively analyze and find dominance in a timely manner.
To address this issue, the EU adopted the Digital Markets Act (DMA) as an ex-ante instrument, complementing traditional competition law enforcement. The DMA moves away from traditional dominance-based assessments and instead relies on structural, quantitative thresholds to identify gatekeepers. Under Article 3, an undertaking qualifies as a gatekeeper if it meets high thresholds related to turnover, market capitalization, and user numbers. The DMA's quantitative thresholds remain high, however, so many platforms with significant market power may not qualify as "gatekeepers," potentially leaving some concerns unresolved, particularly fast – growing firms or diversified conglomerates.
Some enforcers favor introducing a prohibition on the abuse of a form of relative dominance, a concept already recognized at a national level in certain EU Member States, to address platforms’ superior bargaining positions.2 This would tend to harmonize the EU approach with the US case law regarding attempts to monopolize.
2. Network Effects and Market Power in Practice
In digital markets, factors like network effects and data control can significantly reinforce a firm’s dominance.3 Network effects refer to a phenomenon in which the value of a product or service grows as more people or entities use it. Network effects create a self-reinforcing cycle where a large user base attracts even more users, raising barriers to entry for competitors. Similarly, control over extensive data allows dominant firms to refine algorithms and personalize services, making it difficult for rivals to compete on equal terms. When such firms engage in exclusionary conduct such as self-preferencing or limiting interoperability, they can entrench their position and exclude competition, potentially leading to monopoly power, even before consumer harm is directly observable.
Network effects played a crucial role in reinforcing Google’s dominance in both the Google Search and Google Android cases in both the US and the EU. In Google Search (Shopping), the European Commission found that the more users relied on Google’s search engine, the more data it collected, allowing it to continuously improve the quality and relevance of results – attracting even more users in a self-reinforcing cycle. This data-driven network effect created high barriers to entry and made it nearly impossible for rival search engines or comparison-shopping services to compete effectively. Similarly, in Google Android, network effects operated across multiple sides of the mobile ecosystem: more users encouraged more app developers, which in turn attracted more users and device manufacturers. Combined with Google’s pre-installation and tying practices, these feedback loops locked users and manufacturers into Google’s ecosystem, entrenching its dominance. Together, these cases demonstrate how network effects in digital markets can consolidate power, limit effective competition, and push markets toward monopoly conditions.
Similarly, the August 2024 US district court ruling that Google had monopolized the general search market by its series of default search engine agreements was predicated on the network effects resulting from those agreements.4 This network effects concern, in turn, drove the September 2025 remedy decision ordering Google to make certain search index and user-interaction data available to rivals and potential rivals.5
3. Preventive Tools: Market-Wide Investigations and Ex-Ante Measures
In modern, fast-evolving markets – especially digital and data-driven sectors – competition problems can emerge early, through structural features rather than overtly illegal conduct. In such cases, the relevant enforcement authority may not yet be able to show that any one company abused its position – but it can already observe that the market is trending toward concentration and consumer harm. This is where sector inquiries (market-wide investigations) come in.
In fast-evolving digital markets, such investigations play a preventive role, identifying risks – like network effects, data-driven feedback loops, and gatekeeping power – that can lead to dominance or monopoly even before formal dominance is established. Insights from the EU’s Digital Markets Inquiry (2020-2023) directly informed the DMA. Similarly, in the UK, the Digital Advertising Market Study (2019–2020), found that strong network effects, extensive data advantages, and self-reinforcing feedback loops allowed Google and Meta to dominate and reduce effective competition, and led eventually to the Digital Markets, Competition and Consumers (DMCC) Act 2024. Pursuant to the DMCC Act, both Google and Apple have been confirmed as firms with Strategic Market Status (SMS); this initiative reflects the UK’s move toward an ex-ante regulatory approach similar to the European Commission’s, which is considered further in one of our latest blogs.
The EU, through the EU Artificial Intelligence Act (EU Regulation 2024/1689) aims to establish clear rules for the introduction, deployment, and use of certain AI systems, taking an ex-ante approach to supporting responsible innovation while maintaining high standards of public interest protection – including health, safety, fundamental rights, democracy, rule of law, and environmental protection (para. 8 AI Act). While not an antitrust instrument, the AI Act complements competition law by promoting a level playing field in AI markets.
Implications for the Future
The performance or quality of AI relates primarily to the breadth and diversity of its training data. Larger and more varied datasets enable AI to respond more accurately and produce more sophisticated outputs. Companies that already possess abundant user data can gain a head start in AI, and this advantage can create structural barriers for competitors who lack similar data resources, as exemplified by the Amazon6 and Google7 investigations and litigation. Thus, the largest tech companies derive competitive advantages from two interconnected forces: Data Accumulation and Network Effects. These advantages can become anticompetitive when they are used not merely to compete with their competitors, but to foreclose competitors from the market.
When companies with important and meaningful data sets enter adjacent markets, they may establish barriers for competitors who lack equivalent resources. Network effects magnify these risks. This area remains a center of attention both for opening investigations and bringing cases, as well as seeking appropriate ex-ante solutions.
1 Joint Statement on competition in generative AI foundation models and AI products, CMA, 23rd July 2024, EU Commission (see link) - CMA (see link) - US Federal Trade Commission (see link)
2 Section 19a of the German GWB law which allows for earlier more proactive intervention by the German Federal Cartel Office in the case of companies deemed to have 'paramount significance for competition across markets'. Similarly, according to Art. 4a of the Austrian KartG, abusive behavior of companies having "relative" market power (Relative Marktmacht) in relation to their suppliers or customers is also prohibited. Article 9 of the Italian Law No. 192/1998 (Law on Subcontracting in Production Activities) prohibits the abuse of economic dependence, i.e. when a business finds itself in a position to bring about excessive imbalances in the rights and obligations pertaining to its commercial relations with another business.
3 See Draft Guidelines on the Application of Article 102, para. 31.
4 United States v. Google LLC, 747 F.Supp.3d 1 (D.DC., Aug. 5, 2024)
5 United States et al. v. Google LLC, 2025 U.S. Dist. LEXIS 170459 (D.D.C., Sep. 2, 2025)
6 See Subcommittee on Antitrust, Commercial, and Administrative Law of the committee on the Judiciary of the house of representative Investigation of Competition in Digital Markets (2023) at 319
