Progress at the Expense of Accountability: Why the European Commission’s Approach to DMA Compromise Falls Short

A few days ago, Apple announced that it would be introducing changes to its apps in the European Union, “following close collaboration with the European Commission (… to) resolve Apple’s disagreements with the Commission over business and alternative distribution”. The announcement can only be interpreted as the compromise that the DMA’s enforcer could extract from the gatekeeper to settle its ongoing non-compliance procedure that touches upon the exact same concerns. Apple’s statement comes months after Meta (and the European Commission) agreed to a resolution to the impending challenges posed by the gatekeeper’s pay or consent subscription model.

There is an undeniable pattern in both developments that sets a dangerous precedent. From what we can learn from the brief announcements issued by Meta and Apple (which will transform the economic reality of hundreds of business users), the European Commission is ready to negotiate behind closed doors to settle how the DMA should be interpreted, without factoring in the input of third parties. The precedent impacts the DMA’s enforcement via two clear aspects: accountability and third-party redress.

The European Commission compromises on the DMA’s enforcement: the pay or consent example

In December 2025, the European Commission explicitly recognised that Meta would be allowed to “offer users in the EU an alternative choice of Facebook and Instagram services that would show them less personalised ads” and sold it in its press release as a win because this was “the first time that such a choice is offered on Meta’s social networks”. Whilst that might be true, when one reads through the European Commission’s decision fining Meta for its breach of Article 5(2) DMA, it is not particularly clear that the subscription model aligns with the spirit and letter of the provision (for comment of the decision, see here).

In that decision, the Commission considered that the subscription model did not comply with one of two cumulative conditions required by the consent mechanism underlying the exemption of Article 5(2) DMA. In other words, Meta’s subscription model failed to provide the end user with specific choice (i.e., the first of the two cumulative conditions) due to the fact that the configuration of the two options did not provide the end user with the possibility to access the services by choosing a less personalised but equivalent alternative to its non-ads services (para 271 of the decision). That finding was “sufficient” to conclude that the subscription model did not comply with Article 5(2) DMA (para 176), but the European Commission provided a brief outline of the reasons why the subscription model did not abide by the second condition relating to the granting of consent by the end user according to the Articles 4(11) and 7 GDPR standards. First, the criterion of imbalance of power between Meta acting as the data controller vis-à-vis the end user (aka data subject) makes it impossible to determine that consent can be granted freely (paras 180-188). Second, end users cannot refuse consent to the combination of their personal data, without suffering detriment (paras 189-200). To some extent, this argument was tied to the finding of the detriment caused to the end user via the binary choice through non-equivalent possibilities. Third, in the absence of any further free of charge alternative option without behavioural advertising, the subscription model, by virtue of its binary nature, was incapable of removing, reducing or mitigating the detriment that may arise for non-consent users (paras 201 and 202).

Let’s fast-forward then to the moment when the European Commission announced its progress when it came to Meta’s non-compliance with Article 5(2) DMA. The provision of an option with less personalised services addresses reasons two and three. It does not reverse the situation relating to the imbalance of power that still exists between Meta and its data subjects (that the European Commission presented as a reason to annul the validity of effective consent in the sense of the GDPR, although I do argue that this reasoning only leads to a circular outcome with no solution in sight). For argument’s sake, let’s say that there is a way in which consent can be freely given under Article 5(2) DMA in the context of a subscription model like Meta’s. And that is precisely what the European Commission’s argument is in its statement (or I hope that it implicitly is).

However, there is no way in which we (read scholars, stakeholders and third parties to the DMA more generally) learn how the Commission overcame that legal hurdle to arrive to this particular outcome. In the absence of a decision, no one can question whether the European Commission’s decision is correct, nor can any third party sufficiently impacted by such a decision contest the interpretation before the Courts (in a couple of weeks the General Court will decide in Case T-357/24 whether that is the case). In sum, the enforcer signals a clear predilection for behind-closed-doors agreements with gatekeepers to steer clear of accountability, both judicially and socially. This could have been anticipated (and was, by Cseres and de Korte) from the DMA’s lack of any interest to integrate third-party participation into its enforcement framework. That does not make it any less worrisome and dangerous. EU institutions are built on the premise of legitimacy, and no legitimacy can be derived without checks and balances.

Taking it a Step Further: Apple Goes All in

A few months ago, the Commission sold the idea in its press release surrounding Meta’s ‘renewed’ subscription model that it could negotiate DMA outcomes without any accountability and that that was an acceptable course of action. Apple’s announcement goes one step further by touching upon one of the most contested provisions within the DMA, the alternative distribution mandate under Article 6(4) DMA, aside from providing the resolution for the Commission’s recent non-compliance procedure relating to the steering mandate under Article 6(4) DMA (see comment on the decision here and on the evolution of the compliance solution here). 

Concerning the latter, the announcement confirms that “the new terms (…) eliminate the initial acquisition fee and store services fee”. Both these fees represented the commission that Apple applied to each app developer when an end user was steered onto a website to complete a transaction with it. There is nothing fundamentally wrong about the approach, since Apple had not complied with the provision since the Commission issued its non-compliance decision back in April 2025 (regardless of the fact that the enforcer did not impose any periodic penalties due to this reason). Article 5(4) DMA compels gatekeepers to enable steering free of charge with a few exceptions. The initial acquisition fee and the store services fee surpassed those limitations by far.

Notwithstanding, Apple also operated a fundamental reworking of its compliance strategy in alternative app distribution with the European Commission’s acquiescence. Under its previous approach to compliance, Apple had basically incorporated the changes that the DMA compelled it to make into its business model by fencing its application per developer. Each app developer had to choose whether it wished to stay with the old business terms that would still charge the commission per transaction (and not enjoy any of the advantages created by the DMA) or opt in to the new business terms which mandate additional limitations on app distribution and payment processing.

Apple created the problem and now it is providing a solution by unifying its business terms “to a single set of terms, under which Apple charges a commission on the sale of digital goods and services”. As of August 18, there is no iOS EU world where the DMA provisions are not applicable and readily available for any developer to access.

As always, however, the devil is in the details. For those developers that wish to perform distribution via alternative means, Apple will no longer charge the flat rate of 0,50€ per first annual install (Core Technology Fee). Instead, the Core Technology Fee will be replaced by the Core Technology Commission, “a simple 5 percent commission on digital transactions in apps distributed outside the App Store”. Practically, the change was happening in any case, since Apple announced that the Core Technology Commission would be applied a year ago in its compliance workshop (see comment here).

In addition, Apple is also adjusting some of the commission rates it will apply to its alternative app payments, alternatively distributed apps and across the App Store to, according to the gatekeeper, reflect “the many ways Apple creates value for developers’ apps, whether they use the App Store and/or Apple-In App Purchase”. See below the fees to be applied in the following scenarios:

Table 1. Fees to be charged now by Apple to app developers after August 2026 changes.

From the table above we can actually see that the changes introduced, either a) produce no change in the situation that Apple had already presented in its 2025 compliance workshop; b) actually increase the fee that is being charged on the app developers that agree to distribute their apps on the App Store and process their payments through IAP (and adds that app developers must select their payment options for a minimum of 12 months without the possibility of altering that choice further); or c) recalibrates the distribution (and name) of the fees in breach of Article 5(4) DMA and refurbishes into a new limitation on steering. Fundamentally, the European Commission’s compromise, which has been commended by one of its spokespersons, falls short of any idea of compliance under Articles 5(4) and 6(4), as opposed to Apple’s belief that it would “resolve (its) disagreements with the Commission over business terms and alternative distribution”.

On substance, Apple seems to provide the impression that it is loosening its control over alternative app distribution because it will not apply its Notarization process to alternative app stores available on iOS. Mimicking Google’s preoccupation with the risks posed by sideloading (as demonstrated in its own 2025 compliance workshop, see comment here), Apple contends that sideloading “means a bad actor distributing via the web can operate for a long time, harming users, before anyone catches it”. Thus, the gatekeeper will continue to require sideloaded apps to adhere to its baseline review that it controls, despite the fact that Apple’s own engineers liken the App Store security (and screening for apps) to a butter knife in a gunfight.

Having said that, I hope that Apple’s “close collaboration” with the European Commission has not come to a close and this last announcement will not be conducive to the enforcer’s closing of its non-compliance procedure against the gatekeeper. Such a move would legitimise the precise conduct that the DMA wants to do without, with the added component of the Commission’s lack of any accountability by doing so.

My brief response to this would be: let’s open some of the debate for discussion via a public consultation or a transparent debate where all parties involved are, at the very least, heard on what fairness means for them under Article 6(4) DMA. The regulation cannot be effective (if it, indeed, wants to produce outcomes in terms of contestability) in any other way.