On the importance of “average monthly active recipients” (AMAR) in the recent case law of the General Court

By Marie-José Garot, IE University

This post analyses how the General Court has clarified the concept of “average monthly active recipients” through three recent cases (Zalando, Meta, and TikTok). The interpretation given by the Court should oblige the European Commission to set a single methodology for calculating the AMAR, whatever the purpose it serves (for transparency obligations, the designation of VLOPs and VLOSEs, or the calculation of the supervisory fee).

Introduction

The DSA mentions the “average monthly active recipients” (AMAR) for three different purposes. It appears first as a transparency obligation of platforms and search engines, since they must publish it every six months “on a publicly available section of their online interface “(Article 24.2). Second, it serves as a threshold for the designation process of Very Large Online Platform or Very Large Online Search Engine (VLOP/SEs) as stated in Article 33.1. Platforms and search engines with 45 million or more AMAR in the European Union must be designated by the European Commission as a VLOP or VLOSE. Finally, the supervisory fee that VLOPs and VLOSEs must pay to the European Commission is calculated in proportion to their AMAR (Article 43.5. b)). Yet the DSA does not provide much information on how it should be calculated, except in Recital 77 . Furthermore, while under Article 33.3, the Commission may adopt delegated acts laying down the methodology for designation purpose, it has not done so yet, although it released some guidelines in January 2023.

Determining the AMAR raises important questions, including what methodology must be used for its calculation, what constitutes an “active recipient”, and who is responsible for its calculation. Save for a few exceptions , these questions have been largely overlooked by scholarship. Yet, given the importance of being designated as a VLOP or a VLOSE (especially, in regard to the due diligence obligations imposed on them, which the General Court has qualified as an interference with the freedom to conduct a business protected by Article 16 CFREU) these issues deserve to be addressed.

VLOP/SE designations and challenges thus far

In the case of the designation of VLOP/SEs, the rationale behind the 45 million AMAR threshold is that quantity matters. The EU legislature considered that platforms with such a large number of users have the capacity to create or amplify systemic risks affecting society, including risks relating to fundamental rights, public discourse, consumer protection, and public security. Within this framework, the risks that a platform or service poses to society are determined by the number of its recipients. Meeting the AMAR threshold therefore leads to designation as a VLOP or VLOSE and, consequently, to the application of the DSA’s additional due diligence obligations.

Article 33.4, which lays down the different steps of the designation process, provides that the data used by the Commission to designate a VLOP or a VLOSE may come from three different sources: from the platforms and search engines themselves (pursuant to their transparency obligations); at the request of the national Digital Services Coordinators or the Commission; and from “any other available information to the Commission”.

These designation decisions have already given rise to several legal challenges. At the time of this writing, 24 VLOP and 2 VLOSE designations by the Commission remain in force. Six VLOPs have challenged their respective designation decision before the General Court:

Zalando (Case T-348/23 Zalando v Commission ),
Amazon (Case T-367/23 Amazon EU v Commission),
Pornhub (Case T-138/24 Aylo Freesites LTD v Commission, pending),
XNXX (Case T-486/24 NKL Associates v Commission, removed from the Register on 12 February 2026 ),
XVideos (Case T-139/24 WebGroup Czech Republic v Commission, removed from the Register on 12 February 2026),
Stripchat (Case T-134/24 Technius v Commission, pending; on 27 May 2025, the Commission adopted a decision terminating Stripchat’s designation).

The AMAR at the centre of Zalando, Meta, and TikTok cases

The Court has already ruled in the case of Zalando and Amazon , in September and November 2025 respectively, and upheld the designation decisions taken by the Commission. The two rulings provide interesting insights into the designation process and into some of the most prominent legal questions related to it — the AMAR question in particular is pivotal in the Zalando ruling.

Zalando challenged the appropriateness of the AMAR threshold as a means of effectively identifying platforms that pose systemic risks to society, arguing that a qualitative criterion would have been more appropriate. In addition, it contended that the vagueness of the concept of AMAR as well as of the criteria for defining an “active recipient” lead to a breach of the principle of legal certainty and the principle of equality, as each platform may have used its own methodology for calculating such a number, leading some platforms to overestimate their AMAR while others might underestimate it. Furthermore, Zalando argued that the Commission has relied on different data and on different methodologies for the designation procedure and for the calculation of the supervisory fee (paragraph 109).

Two other rulings must be read in conjunction with the Zalando case: Meta (case T55/24) and TikTok (case T-58/24) challenged the legality of their supervisory fees, including the Commission’s choice of legal instrument. In these cases, the Commission relied on estimates provided by third-party companies and specified the methodology in the individual implementing decisions setting the fees. Both services challenged the Commission’s use of that common methodology and argued that, if such a methodology were adopted, it should have been laid down in a delegated act rather than in the implementing decisions.

While the General Court confirmed that the Commission could rely on estimates provided by third-party companies, it annulled the fee decisions because the methodology for calculating AMAR should have been laid down in a delegated act. It nevertheless maintained the effects of the decisions until they are replaced, and gave the Commission one year from the date of the judgement to adopt a proper delegated act.

One AMAR, one methodology

The three rulings give an opportunity to the General Court to interpret the DSA and to remedy some of its lacunae.

In the Zalando judgment, the General Court held first that the AMAR is an appropriate criterion to designate VLOP/SEs, considering the risks marketplaces such as Zalando may pose to society. Nevertheless, the General Court seemed to acknowledge that some platforms may pose greater risks than others (in paragraph 138, the General Court compares social media to marketplaces). The openness of the concept allows it to cover different types of platforms, irrespective of the specific risks they pose to society.

Second, the General Court endorsed the definition of an “active recipient” as “the person who requests that platform to host information or the person who is exposed to that information” (Article 3 (p) DSA). Following Recital 77, it underlined that no specific interaction is required (such as clicking further, commenting, sharing or buying). The General Court explained “that a provider of an online platform cannot exclude, for the purposes of calculating the AMAR, recipients of the service who, although they have been exposed to the information hosted by that platform, have not registered with it, have not entered into a transaction on it, have remained inactive for a certain period of time, have refused to use cookies or have accessed the platform from an application.” (paragraph 95).

The Court also confirmed that it is the responsibility of platforms not to double-count the same recipient (although they are prevented from profiling or tracking their users) and not to consider bots (“where possible” as states Recital 77). The Court noted that the minimum duration of engagement can range from 3 to 45 seconds depending “on the nature of the service and the way recipients of the service interact with it.” The Court also accepted that platforms may exclude accidental visits that did not result in actual exposure to hosted information.

The Court adopted therefore a very broad and adaptive definition of the term “active recipient”. Again, the openness of the different criteria is justified by the necessity to tackle different platforms and services, a kind of “one size fits all”. What seems important to the Court, interpreting Recital 77, is that the recipient directly accesses the service or “engages” with it (paragraphs 60 and 61). The Court is aware that the broad definition of “active recipient” may lead some platforms or search engines to overestimate their AMAR. However, it considered that overestimation does not pose a problem with regard to designation, which is based on a threshold rather than on a specific number. By contrast, underestimating the AMAR would be contrary to the objectives of the DSA (paragraph 98).

Finally, in the Meta and TikTok rulings, the General Court recognised an obligation for the Commission to adopt a single methodology for calculating the AMAR (whatever the purpose is) through a delegated act. Although Article 43 DSA does not expressly require it, the Court considered that the Commission must adopt such an act, relying on a teleological interpretation of the DSA. It gave it one year from the date of the judgment to do so. The two rulings definitively put the AMAR at the centre of the regulation of digital platforms, as it stated that “The AMAR is both an essential element of the methodology for determining the supervisory fee and a concept which must be understood uniformly and consistently throughout the DSA” (paragraph 48 in Meta and paragraph 50 in TikTok).

These rulings reveal an irony pointed out by Martin Husovec : In Article 25.3 of its original DSA proposal, the Commission proposed that it be required to adopt a delegated act laying down the methodology for calculating AMAR for designation purposes. This became optional in the final version of the DSA: under Article 33(3), the Commission “may” adopt such a delegated act. Through its insistence on a uniform methodology applicable across the DSA, however, the General Court has effectively turned the “may” of Article 33(3) back into a “shall”.

Conclusion

Although imperfect, the decisions of the General Court can help provide clarity and coherence for calculating the AMAR. However, the story has not yet come to an end, as Zalando has brought an appeal before the Court of Justice and the Commission has done the same against the Meta and TikTok rulings. Given the central role of AMAR in triggering the due diligence obligations for VLOP/SEs, the judgments of the Court of Justice will be important both for the Commission, by clarifying the scope of its powers under Article 33 and 43 DSA, and for VLOP/SEs, by enhancing legal certainty.

This research is partially funded by PID2023-149184OB-C43 (“The Strategic Responsibility of the Digital State”) granted by MCIU /AEI/10.13039/501100011033 / FEDER, UE