Artificial intelligence and antitrust law as a key issue for companies in 2025

On 11 February 2025, European Commission President Ursula announced the InvestAI initiative at the Artificial Intelligence Action Summit in Paris. As part of this initiative, EUR 200 billion will be mobilised for investments in artificial intelligence (“AI”), including a new European fund of EUR 20 billion for AI gigafactories. This large AI infrastructure is supposed to enable open, collaborative development of the most complex AI models and turn Europe into an AI continent, thereby increasing the EU’s global competitiveness on AI markets.

The impact of AI on competition and the markets have also come into an ever-increasing focus of competition authorities all over the world. In looking at the likely key issues in antitrust law in 2025, companies should thus pay close attention to what this increased focus on AI could mean for their business. This article provides a short overview on the way that various competition authorities have started to scrutinize potential restrictions of competition connected to AI and the most critical concerns for companies in practice.

I. Increased scrutiny and cooperation by competition authorities worldwide

Numerous competition authorities around the world have outlined their views on the impact of AI on competition and indicated that they are highly alert and interested in closer cooperation of enforcement. This is demonstrated by the following recent and non-exhaustive examples:

In its Competition Policy Brief of September 2024, the European Commission (“Commission”) as the EU’s competition authority emphasized that it is actively monitoring potential anticompetitive practices connected to AI and is looking for greater coordination between competition authorities in enforcing antitrust law on AI-related markets. The Commission's particular leadership on AI and its willingness to seek cooperation with other competition authorities can also be seen in its previous joint statement on AI competition issues with the UK's Competition and Markets Authority ("CMA"), the US Department of Justice ("DOJ") and the US Federal Trade Commission ("FTC"). The joint statement sets out the authorities’ view of the competition risks related to AI and the principles that they will apply in assessing these risks. The competition authorities also jointly stressed the need for regulatory cooperation to ensure fair AI markets.

In addition, the G7 competition authorities (from Germany, France, the United Kingdom, Italy, Japan, Canada and the USA) published a joint Digital Competition Communiqué in October 2024. In view of the rapid developments in the AI landscape, the authorities formulated guiding principles to ensure fair competition and innovation. This includes measures to prevent established companies from exploiting their market power or suppressing innovation and to facilitate access to essential AI resources.

Further, in December 2024 the Korean Fair Trade Commission (“KFTC”) also published a comprehensive policy report on potential competition concerns resulting from AI based on a market survey, indicating that it will continue to scrutinize such concerns in-depth.

II. What are the most critical concerns?

1. (Unintended) Collusion through the use of AI tools

A long-standing concern of competition authorities is that algorithms and AI can facilitate or even trigger impermissible collusion between independent market participants. For instance, antitrust law generally prohibits the coordination of prices as well as the exchange of so-called “commercially sensitive information” (such as information on prices, costs, customers) between competitors. The use of AI could result in violations of these prohibitions:

• Price fixing – AI tools trained to automatically optimize pricing through analysing market conditions and competitors' behaviour in real time are gaining popularity. In situations where multiple competitors use AI tools for that purpose, a violation of the price fixing prohibition is possible where these AI tools are capable of coordinating with each other and setting prices based on this coordination. Similarly, a form of coordination could also occur where competitors merely use the same AI tool, which determines their pricing on the basis of the same information and the same underlying AI model (thus leading to a uniform pricing level). Even where this coordination is unintended, companies could in principle be found liable under applicable case law if the coordination could at least have been reasonably foreseen.

• Exchange of commercially sensitive information – AI models are trained on extensive data sets. It is thus possible that some AI tools are trained and make decisions based on data that would be considered as commercially sensitive and may not be exchanged between competitors under European antitrust law. Oftentimes, however, it is not evident for the users on what data AI tools are trained and whether data from the ongoing use of the same tools by competitors could be used to further train the model. Nonetheless, insofar as companies make use of AI tools trained on commercially sensitive data and if the commercially sensitive nature of the training data is known or at least reasonably foreseeable, liability for users under antitrust law is also conceivable in this situation.

In order to minimize antitrust law risks from the use of AI tools, companies are therefore well-advised to take proactive compliance measures, including know-your-supplier processes, contractual drafting to minimize liability as well as conducting regular audits of their usage of AI tools.

2. Consolidation of AI markets by large digital companies

Another significant concern of competition authorities worldwide is that a few large digital companies with access to important resources for AI may consolidate AI markets and thereby limit competition:

• Consolidation through partnerships – In AI markets, access to large, high-quality data sets, computing power and specialized personnel are all key resources. Major digital companies like Google, Amazon, and Microsoft possess significant financial power, enabling them to control vast amounts of data and IT services and to hire specialized personnel. Without access to these essential resources, new market entrants could struggle to establish themselves in AI markets. For this reason, many small, innovative companies have been entering into partnerships with large digital companies in order to access these essential resources.
  
  This has led the competition authorities to review whether merger control rules are applicable to such partnerships. A prominent case is the partnership between Microsoft and OpenAI, which has been investigated by the Commission, the FCO, and the CMA under their respective merger control powers. Further, the cooperation between Alphabet (holding company of Google) and Anthropic as well as Amazon's parallel partnership with Anthropic were also recently subject to merger control reviews.

• Consolidation through acquisitions – Competition concerns also arise when established companies buy smaller players to remove potential competitors (also referred to as “killer acquisitions”). This was recently being assessed in the context of the acquisition of AI startup Run:ai Labs Ltd by U.S. chipmaker NVIDIA Corporation. Ultimately, however, the Commission (acting after accepting a referral by the Italian competition authority) found that the acquisition did not raise competition concerns and approved the transaction.
  
  Another example of a possible consolidation through acquisition is the Microsoft/Inflection case reviewed by the Commission, the FCO, the CMA and other competition authorities. In this case, Microsoft made use of the so-called practice of “acqui-hire”, i.e. transactions primarily aimed at acquiring qualified and specialised personnel (see our Noerr Insight for more details on this case).

• Consolidation through exclusivity agreements – In its Competition Policy Brief, the Commission further explained that it is investigating whether agreements between Google and device manufacturers (such as Samsung) to preinstall Google’s small AI model, Gemini Nano, on mobile devices could raise antitrust law concerns. Such exclusivity agreements could make it difficult for developers of similar AI models to compete for users. That is because users of those mobile devices are likely to stick with the pre-installed AI model rather than installing a competing model. Such conduct by the major digital companies could be investigated by the competition authorities under the theory that it constitutes an abuse of a market dominant position.

All market participants should take note of this focus by the competition authorities on different forms of possible consolidation of AI markets. Companies that intend to enter into partnerships, exclusivity agreements or agreements for acquisitions are well-advised to assess in advance whether merger control or other antitrust rules apply. Companies affected by agreements concluded by competitors on AI markets on the other hand should consider whether to notify the competition authorities of their competitive concerns.

III. Outlook

These recent interventions and statements show that many competition authorities are heavily focused on competitive issues caused by AI. It is therefore likely that the use of AI tools and a possible consolidation of markets related to AI will continue to be heavily scrutinised and be a key antitrust law issue in 2025. This also concerns the question of whether the Digital Markets Act (“DMA”) should be used to address potential issues, specifically the conduct by those major digital companies that may have significant market power on all levels of the AI supply stack – an issue that Andreas Mundt, President of the FCO, has repeatedly pointed to (see his most recent comments here). In the context of the Artificial Intelligence Action Summit in Paris, for instance, amendments to the DMA adding new AI-specific rules were (once again) discussed.

At the same time, competition authorities may also find themselves under pressure not to hinder innovation on AI markets. This is particularly true of the competition authorities in the EU, which has been concerned with not losing even more of its global competitiveness, as evidenced by Mario Draghi’s report on “The Future of European Competitiveness” (see in more detail our Noerr Insight on the report). In this regard, competition authorities will have to strike a delicate balance, in particular in cases involving European AI companies. Another challenge will certainly be the speed with which AI develops, requiring competition authorities to keep up with the rapid developments in the markets.

In any event, similar to how AI has already brought about vast changes, it may also heavily shape the practical application of antitrust law in 2025. All companies making use of AI tools and all stakeholders in markets related to AI should keep a close eye on how competition authorities continue to scrutinize these markets and what this means for their own business.