Two’s Company, Three’s a Crowd: The General Court Rejects Opera’s Appeal on the Commission’s Non-Designation of Microsoft Edge (Case T-357/24)

Last week, the General Court (GC) issued its ruling relating to the European Commission’s (EC) decision not to designate Microsoft’s web browser Edge. This was the first ruling where the GC adjudicated in an action addressed by a third party to the Digital Markets Act (DMA) and not a gatekeeper (as it has already done in Cases T-1077/23, ByteDance Ltd v European Commission, T-1078/23, Meta Platforms, Inc. v European Commission and T-1079/23, T-1080/23 and T-214/24, Apple and Apple Distribution International v Commission, see the corresponding comments here, here and here). This time, it was Opera, a competing web browser to Microsoft’s Edge, that decided to appeal a decision adopted by the DMA’s enforcer. However, it did not seek to annul the designation decision, but to annul a non-designation decision that the EC had taken following the market investigation procedure under Article 17(3) DMA.

The GC’s ruling builds on the corpus of evidence that a potential gatekeeper can bring to the table when arguing against its designation under the DMA as well as in setting out the role of third parties within the ex ante regulatory framework. For me, the GC’s judgement indicates that third parties cannot be simply obliviated from the DMA space, even if they are not given much space to intervene in its administrative proceedings.

Admissibility, Again: The General Court Adds Some Nuance

In its ruling before the summer relating to Meta’s designation decision, the GC did not shy away from considering admissibility exceptions in detail in light of the DMA provisions. In that particular case, it assessed whether Meta still held an interest in pursuing the appeal of the designation as a core platform service (CPS) that the EC had decided to de-designate before the GC had an opportunity to pronounce itself. The GC responded positively. In this instance, the GC also ruled favourably in favour of the applicant, despite the EC’s narrow reading of Article 263 TFEU.

The EC contended that the non-designation decision was not of individual concern to Opera, because it lacked active participation in the procedure leading to the adoption of the decision and was not impacted by it (para 26). It set out several examples of how its lack of active participation manifested, such as the fact that it replied in an incomplete manner to its request for information on Edge’s classification as an important gateway for business users to reach end users (para 27) and that it was very much less involved than other competitors of Edge, which proactively submitted additional observations (para 29). The enforcer even points out that Opera “could have taken the initiative to engage in that investigation” when the Commission published its press release announcing the opening of the market investigation (para 29). To the extent that it did none of those things, the EC assumes away the lack of active participation in the administrative designation proceedings. Similarly, the enforcer also establishes that nothing indicates that Opera was in a different position as opposed to other alternative web browsers when it made the decision not to designate Edge, which would justify the GC’s ruling in favour of the action’s inadmissibility (para 30).

The GC interpreted both lines of reasoning with quite an open mind. First, it considered the effects that Opera suffered on its market position as a consequence of the EC’s decision not to designate Edge. Even though the case law requires that the applicant must have suffered a substantial effect on its market position, the GC considered that it was enough that it could produce a prima facie finding that the lack of designation led to a substantial adverse effect on that position (para 37). By doing so, the GC lowered the threshold quite substantially.

From that point onwards, it argued how Opera was impacted in its market position in the form of enduring “less favourable developments than would have been the case had the undertaking which was not considered to be a gatekeeper been designated as such” (para 38). The GC laser focused on the impacts that Edge’s designation would have caused, i.e., the obligations that would have kicked in in the following 6 months to the designation. Article 6(3) DMA was the clearest candidate for the job. Opera would have been directly impacted by the designation because Microsoft would have been forced to display the choice screen of web browsers mandated under this particular provision (para 45).

From the evidence before the GC, it was, therefore, particularly clear that Opera would have been included in the choice screen, given that in 2023 and 2024 it was the fifth most significant browser on personal computers and the sixth on all devices in terms of share of total number of web page views (para 42). The fact that Opera could shrink in the future (and therefore not merit sufficient standing to be shown in the choice screen) and that it had already been included in Alphabet’s Chrome and Apple’s Safari did not trump the fulfilment of these criteria (paras 47 and 48).

Regarding the second tenet required by Article 263 TFEU, the GC acknowledged the limited role that third parties play in the DMA. As opposed to the EC’s reasoning that such limitations should not limit the GC’s findings on admissibility, the GC took into account the “specific features of the regulatory framework (…) and in particular the fact that (it …) does not provide for the participation of third parties in the administrative procedure, unless the Commission decides to send them a request for information under Article 21 of the DMA or to interview them on the basis of Article 22 DMA” (para 59).

In that context, the GC does not believe that Opera can be sanctioned for answering a request for information which was based on Article 21(2) (aka it was a simple request for information, the DMA’s words, not mine) and not Article 21(3) DMA (para 60), which requires wider efforts from the undertakings who are addressed these types of requests. The EC chose Opera as “one of a limited number of Microsoft’s competitors” to respond to its requests for information and it was “the only formal opportunity” to hear from alternative web browser providers (paras 56 and 61). This is enough to consider that Opera’s participation in the market investigation, “albeit limited” formed an additional element of the body of evidence that supported the GC’s rejection of the plea of inadmissibility (paras 62-64).

The GC does acknowledge the limitations that third parties face when trying to intervene in the administrative proceedings steered by the EC vis-à-vis the gatekeepers. A natural consequence of those limitations is that third parties cannot be expected to deliver spontaneous information to the EC (para 61), so that they can potentially open the door for future appeals that they may wish to pose in case they disagree with the EC’s final position. This finding regarding the plea of inadmissibility is particularly relevant, insofar as it provides more leeway to third parties to scrutinise the EC’s actions when it comes to DMA enforcement, which adds to the level of accountability that it is subject to (aside from the possibility of gatekeepers to appeal its decisions).

The DMA is addressed to create opportunities for third parties (aka business users, if one uses the regulation’s terminology), so it follows naturally that third parties must be granted access to the courts to hold the enforcer accountable on how it is applying its provisions. This was not particularly clear from the previous DMA case law.

However, it still poses an alternative world-of-things that does not open the door for third parties entirely. In this case, the EC chose that Opera was a qualified actor from whom it sought its opinions to grasp the web browser picture better. What would happen in those cases where the EC decides not to contact a particular undertaking that wishes to appeal an EC DMA decision? Following the GC’s ruling, the action would be found inadmissible, to the extent that it had no access to the administrative proceeding because the enforcer chose that its views were not relevant enough to include in the scope of its investigation. Put in other words, the DMA squeezes third parties out of the regulation’s enforcement, whereas the GC’s finding can unintentionally make third parties more irrelevant within the subsequent regulatory iterations.

Asking the EC to address requests for information comprehensively to all market participants sounds like a high ask to make, too, particularly given that the DMA is aimed to deliver results quickly. An ad infinitum list of market participants to contact will not deliver the regulatory objectives. The alternative is, therefore, to tweak the DMA’s administrative procedure from the inside. It is a given that the EC has expanded the provision’s scope through subsequent implementing acts (which do not have that nature, in my opinion, but that’s a discussion you’ll find here), so it could expand the instances where third parties can ‘spontaneously’ participate in the proceedings, without requiring that they approach the EC when it publishes a press releases noting an enforcement development.

The Merits of the Case: Ecosystems, Gatekeeper Control and the Nature of Arguments in Designation

After the GC’s rejection of the plea of inadmissibility, the Court rejected Opera’s plea in seeking to annul Edge’s lack of designation which was based in contesting: i) the low scale of usage of Edge in the web browser CPS category; ii) Microsoft’s lack of control over Edge’s relevant architecture due to its dependence on the browser engine Blink; and iii) the finding that Microsoft’s ecosystem did not sufficiently contribute to making Edge an important gateway. The merits of the appeal revolved around the EC’s decision not to classify Edge as an important gateway for business users to reach end users, in the sense of the second requirement of Article 3(1) and corresponding to the second threshold under Article 3(2) DMA.

Back in 2023, the EC decided to accept the rebuttal of the presumption to the quantitative thresholds submitted by Microsoft on its web browser Edge, online search engine Bing and online advertising service Microsoft Advertising. In relation to Edge, Microsoft put forward two main arguments before the EC. First, that it had a relatively small scale considering the overall scale of activities within the web browser CPS category and in comparison to other web browsers offered in the Union. Second, that it cannot influence the operations of business users to its advantage because business users do not optimise their websites specifically for Edge. Microsoft decided to stop offering its own browser engine and it started to rely on Google’s open source Blink (paras 7 and 8 of the Decision opening the market investigation). As a response, the EC opened a market investigation under Article 17(3) DMA and already indicated that it would be checking the role that Microsoft’s CPS Windows PC OS played and the possible impact that the gatekeeper exercises control over its platform ecosystem, including Edge, Bing and Microsoft Advertising (para 16).

Once the EC finished its market investigation, three main sets of arguments confirmed that Edge did not satisfy the legal requirement under Article 3(1) DMA. First, its relatively small scale, which accounted for 5,8% of the overall activity of the web browsers across all devices (para 44 of the Decision closing the market investigation). In the EC’s own words, that intensity of usage indicated that it was not an important gateway (para 55).

Opera contested that finding based on two different reasons. On one side, it challenged the decision by stating that the EC calculated Microsoft Edge’s market share with reference to the share of the total number of web pages, even though the DMA considers that market positions and non-quantitative criteria cannot be factored by the enforcer in the designation process (para 74 of the ruling). The GC considered that the regulation only bars the EC from considering justifications on economic grounds seeking to enter into market definition or to demonstrate efficiencies (para 80). Comparing the scale of usage of the CPSs did not fall into any of those categories (paras 87 and 88). Thus, the GC rejected those allegations. Later on in the ruling, the GC upholds the argument that the EC is not automatically forced to review all reasons sustaining qualitative designation under Article 3(8) to meet its evidentiary burden for the market investigation under Article 17(3) DMA (para 245). On the other side, it put forward a number of reasons that could have sustained Edge’s designation, such as the high number of end users of Edge (para 99), its pre-installation on computers equipped with Windows PC OS (paras 111 and 112) and its rapid increase in Europe from 2015 to 2024 (para 114). The GC considered that none of those facts were sufficiently strong to outweigh the relevance of Edge’s low scale of usage compared with other web browsers (paras 99 and 103).

Second, Edge’s links to other Microsoft services in Windows PC OS notably Bing and Microsoft Advertising, where it underperforms, and its reliance on Chromium Blink’s engine (paras 45 and 46 of the Decision closing the market investigation). The EC concluded that the Microsoft platform ecosystem does not sufficiently contribute to Edge being an important gateway, due to the fact that its ability to use Edge as a lever to drive usage of other Microsoft products was limited and insufficient (para 68). Based on those facts, the EC did not designate Edge. However, it noted that the results of the market investigation showed that there may be a potential for Edge becoming a more important web browser for business users to reach end users in view of Microsoft’s launch and integration of AI-based services and futures (para 69).

Building on the ambiguity of the non-designation decision, Opera drilled down on the questioning of the validity of these arguments by arguing that the use of Blink as a browser engine could not meaningfully contribute to rebutting the presumption that Edge was an important gateway (para 121 of the GC’s ruling). The EC and Microsoft maintained that “the main reason for using Blink was that business users did not ensure the operability of their websites with its browser engine and that software providers did not create extensions for that browser engine” (para 134). Some websites did not support Microsoft’s proprietary browser engine, which resulted in their suboptimal performance on Edge, so that end users switched to its competitors to enjoy a better browsing experience. Therefore, Microsoft abandoned its own browser engine in favour of Chromium’s open-source Blink.

Following the same reasoning that the EC brought up when it considered the designation of Alphabet’s email service Gmail (Alphabet’s designation decision, paras 144 and 145) and Microsoft’s own Outlook.com (Microsoft’s designation decision, para 129), the EC defended that the gatekeeper’s lack of control over the browser engine entailed that it could not influence the operations of a substantial part of business users to its advantage (paras 149 and 157 of the GC’s ruling). Opera contested the finding by arguing that: i) the DMA does not point a strong correlation between a gatekeeper’s browser engine and the catering of its web browser to reach to such a conclusion (para 146); and Microsoft did have control over Edge’s relevant architecture to the extent that Blink is an open-source code product, which Microsoft could develop, modify or adapt (para 154).

The GC highlighted that Recital 43 of the DMA sets out the relevance of the browser engine in relation to the powering of web browsers as the “key browser functionality (rendering its) speed, reliability and compatibility of web pages with the web browser” (para 158). Having said that, the GC then fleshes out that Microsoft had no incentive to apply modifications to Blink because it would give rise to the same suboptimal performance of Edge that led to the replacement of its own browser engine in the first place (para 161). The only relevant modification that Microsoft did incorporate into Blink was that of allowing websites to seek end users’ consent regarding privacy, such as blocking third-party cookies (para 163). However, the GC believes that Opera did not sufficiently explain how the use of this software “gave the possibility of influencing to its advantage the operations of a substantial part of business users” (para 164). Even if Opera would have presented enough justification, it is unlikely that the reasoning would have produced any operative changes to the GC’s ruling, given that, at the end of the day, it is Alphabet, as Chromium’s administrator that makes the decision to review and approve those decisions. Thus, it is crystal clear to the EC that Microsoft lacks sufficient autonomy over its relevant architecture to establish that it holds substantial control over its business users (para 165).

Aside from the factual component of those findings, the GC also drew a particularly interesting distinction between the arguments that gatekeepers can put forward to rebut the quantitative presumption under Article 3(2) DMA and those that are “additional factors” that support such a finding (para 170). When one reads the GC’s reasoning based on the two main bases for the appeal, the Court clearly holds that Edge’s low scale of usage was the main argument sustaining its non-designation, whereas its lack of control over its browser engine operated as an additional factor which reduced “the importance of Edge’s high number of end users”. In other words, not all arguments set forth by the gatekeeper play the same role nor hold the same nature for a CPSs designation (or non-designation, for that case). Edge’s reliance on Blink as a browser engine held, therefore, less weight than the low scale of usage of the non-designated CPS vis-à-vis Alphabet’s and Apple’s. In turn, if the applicant’s argument would have succeeded in demonstrating that it does control the relevant architecture of its browser engine, it would not have overturned the outcome of the EC’s finding. 

As opposed to what the GC highlights throughout the ruling that the gatekeeper must set out arguments discrediting the application of the quantitative presumption under Article 3(2) DMA, when a third party seeks the annulment of a non-designation decision, it must discredit the EC’s main argument (i.e., the size of Edge in relation to other web browsers) in making that particular decision. In this sense, the GC went even further and classified the EC’s reliance on the lack-of-an-ecosystem-effect as a “relevant factor” for assessing whether the CPS constitutes an important gateway (para 188). Hierarchically, we would theoretically place the EC’s key finding (low scale of usage vis-à-vis other designated CPSs), the additional elements that contribute to that finding and the relevant factors that also serve as a means to perform that same assessment. Thus, the EC’s consideration of the impacts of Microsoft’s ecosystem on Edge results in an afterthought. Mirroring my previous argument, should Opera have succeeded in proving that Microsoft’s ecosystem did act as a lever for Edge, then that finding would still not have been potent enough to annul the EC’s decision.

Procedural Guarantees Under Article 17(3) DMA

Within this hierarchy of arguments, reasons and key findings, Opera also appealed a somewhat obscured part of Edge’s non-designation, which is the procedural safeguards that the EC must apply when conducting a market investigation under Article 17(3) DMA.

When the EC intervened before the Court of Justice on ByteDance’s appeal brought against the General Court’s initial judgment (Case C-627/24 P, see comment of the hearing here), it basically recognised that the rebuttal of the quantitative presumptions operated procedurally as a sounding board. That is to say, the EC responds to the arguments set forth by the gatekeeper seeking to contest its designation. In the hearing before the CJEU, when asked by AG Biondi on the procedural safeguards that applied under the market investigation under Article 17(3) DMA, the EC’s representatives recognised that the DMA’s enforcer is not only forced to assess the rebuttal arguments, but also to collect additional evidence. In any case, according to the EC, the burden on the undertaking remains high and does not shift back to the EC when it agrees to conduct a market investigation.

The GC’s ruling contributes to clarifying how bound the EC is with respect to the general principle of good administration when conducting a market investigation under Article 17(3) DMA and when adopting its closing decision (para 222 of the GC’s ruling). According to the GC, the EC cannot simply “confine itself to assessing the arguments and evidence adduced by the undertaking seeking to rebut the presumptions laid down in Article 3(2) DMA (…), but must, as a matter of good administration, play its part, using the means available to it, in ascertaining the relevant facts and circumstances in order to carry out an appropriate assessment concerning the rebuttal of those presumptions” (para 224). In this sense, the EC must take into account the information submitted by the undertaking, the participants in the market investigation, “but also other information which appears relevant for assessing whether that undertaking has demonstrated that it does not satisfy the requirements laid down in Article 3(1) DMA” (para 227). The GC places a caveat to this widespread obligation by highlighting that the EC cannot be expected to seek on its own initiative and in the absence of any evidence to that effect, all the information which might be connected with the case before it (para 228). In my own view, this caveat makes it practically impossible for any potential annulment of the EC’s findings when it pursues a market investigation under Article 17(3) DMA to succeed (au contraire to Case T-1078/23 where the EC was sanctioned for not considering the circumstances in which the designated CPS operated).

With those limitations in mind, the GC considered that the EC went far and beyond when conducting its market investigation under Article 17(3) because it considered Edge’s integration into Microsoft’s ecosystem and whether Edge could benefit from advantages linked to the use of AI (paras 237 and 238). Although the former analysis is clearly apparent from its Decision closing the market investigation, the latter was not included in the EC’s reasoning, so it is not completely clear how the EC’s reasoning on this front managed to overcome the evidentiary burden that the principle of good administration requires of it. Despite the EC’s few indications (in its hearing before the CJEU) and the GC’s clarifications in this ruling, the terms that third parties (and gatekeepers) can hold the EC to account for its market investigation under Article 17(3) DMA are still ambiguous, with the EC only needing to fulfil a general obligation to collect additional evidence, but not any evidence connected to the case.