The neo-Brandeisian movement—emphasizing market structure and fairness over consumer welfare—has struggled to gain traction in several jurisdictions. In Europe, by contrast, lawmakers have codified its core premises. The Digital Markets Act (DMA) operationalizes this approach by prioritizing ex ante structural interventions intended to reshape rivalry in digital markets.
The regime’s most significant feature lies not in its high-level goals, but in its enforcement tools. Through “specification proceedings” under Article 8(2), the European Commission has begun to move from policing competitive constraints to directing system design. The Commission’s recent actions involving Google and the Android ecosystem illustrate that transition most clearly.
Brussels Writes the Code
On Jan. 27, the European Commission opened two specification proceedings concerning Google’s Android operating system, focusing on system-level features used by Google’s AI services, including the generative AI model Gemini. The Commission stated that the proceedings would “assist” Google in complying with its obligations under Articles 6(7) and 6(11) of the DMA.
Those provisions impose demanding duties on designated core-platform services, including Google’s. Article 6(7) requires a gatekeeper that controls an operating system to provide third-party developers “free and effective interoperability” with the same OS, hardware, and software features available to the gatekeeper and its affiliates. Article 6(11) requires designated search engines to grant rival search engines and other business users access to aggregated ranking, query, click, and view data generated on the platform on fair, reasonable, and non-discriminatory (FRAND) terms.
Article 8(2) supplies the enforcement mechanism. It empowers the Commission to adopt an implementing act specifying the precise measures a gatekeeper must take to comply with Articles 6 and 7. The DMA’s primary rules describe outcomes. Article 8(2) proceedings dictate methods. By initiating specification, the Commission signals that a firm’s existing compliance efforts are insufficient and that the regulator will prescribe the technical steps needed to bridge the gap between statutory obligations and operational software.
The Commission portrays this process as assistance—administrative guidance rather than compulsion. The description obscures the practical effect. Specification is not merely interpretive; it is prescriptive. By detailing concrete technical measures, the Commission moves from articulating competitive constraints to directing implementation. The regulator no longer states what market structure should achieve. It dictates how a product must function.
That shift raises an institutional question: when does ex ante competition regulation become de facto system design, and is the European Commission institutionally equipped, technically competent, or legitimately empowered to serve as a central technology architect for the digital economy?
This is not the Commission’s first use of specification proceedings. In 2024, it initiated similar proceedings against Apple and prescribed measures under Article 6(7) for iOS. Apple responded immediately, challenging the decision before the European Court of Justice and seeking, inter alia, a declaration that Article 6(7) was inapplicable and “that Article 1 of the Contested Decision is unlawful insofar as it imposes and specifies obligations under Article 6(7) DMA on Apple in relation to iOS.”
Google has also criticized the growing regulatory intrusion. As Clare Kelly, Google’s senior competition counsel for Europe, the Middle East and Africa, stated publicly:
Android is open by design, and we’re already licensing Search data to competitors under the DMA. However, we are concerned that further rules, often driven by competitor grievances rather than consumer interests, will compromise user privacy, security, and innovation.
You Can’t Regulate Latency Away
The Commission appears to assume that mandating equal access to OS features will mechanically produce competitive neutrality. In deeply integrated systems like Android, that assumption does not hold.
Technical parity is difficult to achieve. Equal access to application programming interfaces (APIs) does not ensure equal performance. Performance depends on system-level optimization, prioritization, and integration choices that third parties cannot easily replicate through external access alone. An operating system, for example, may grant its native AI assistant prioritized, direct access to a device’s neural processing unit (NPU) for instantaneous responses, while third-party models must queue for the same resources through generic software layers, producing perceptible latency.
Proponents of the DMA may treat that gap as justification for stricter mandates. The logic reverses cause and effect. The software layers that introduce latency are not arbitrary barriers; they are core abstractions that maintain stability, security, and battery management. Granting third parties bare-metal access equivalent to first-party components is often technically infeasible without compromising the device’s integrity.
More fundamentally, total parity weakens incentives to innovate. Integrated firms invest in complex hardware-software optimization to produce differentiated products. If regulation requires every latency-reducing improvement to be shared immediately and perfectly with rivals, the incentive to develop those improvements declines. The likely outcome is not faster competing apps, but fewer latency-reducing features built at all.
A further risk follows. If a firm cannot safely extend identical access to third parties, regulators may pressure it to withhold features from its own services. The platform would then be engineered for the lowest common denominator of external compatibility. Competition policy would function less as a spur to innovation than as a mechanism for forced leveling.
The institutional implications are equally significant. Once the Commission specifies which OS features must be opened, on what terms, and with what performance characteristics, it assumes a role resembling a central planner—or at minimum, a permanent arbiter of operating-system tradeoffs.
That role carries real costs. OS-level regulatory errors do not self-correct easily. Rigid specifications, once embedded in compliance systems, may ossify architectures, impede security updates, or degrade battery life and user experience. In a rapidly evolving AI environment, regulatory lag risks locking yesterday’s technical assumptions into tomorrow’s infrastructure.
Moreover, what qualifies as “effective interoperability” or FRAND terms today may become technically obsolete within a year. Maintaining relevance would require constant revision of technical specifications, drawing the regulator ever deeper into day-to-day product governance and engineering decisions and placing a continuing regulatory shadow over the firm’s development process.
One Regulator to Rule Them All
The specification process also carries constitutional and administrative implications. Under the DMA’s enforcement model, the European Commission defines the abstract obligations through designation, interprets them through guidance, specifies their technical content through Article 8(2) proceedings, and then enforces compliance, including the imposition of substantial fines.
Judicial review remains available, but it moves slowly. Final decisions can take years, and courts typically defer to the Commission on complex technical and economic questions. In practice, regulatory authority concentrates in a single agency, constrained only imperfectly by delayed and nonexpert ex post oversight.
The arrangement raises questions about institutional balance, procedural fairness, and the long-term legitimacy of the DMA’s enforcement structure. The regime grants an administrative body extensive prescriptive authority without correspondingly strong ex ante checks.
When Regulation Becomes the Disease
Classical philosophy offers a cautionary lesson. In “The Republic,” Plato criticizes the physician Herodicus for confusing cure with maintenance. A treatment that restores a patient to ordinary functioning succeeds. A regimen that demands constant supervision signals failure.
The Commission’s turn to technical specification under the DMA presents a similar risk. If interoperability obligations require continuous oversight, perpetual negotiation, and ever finer calibration, competition policy may preserve only the appearance of contestability while displacing its core purpose: fostering self-sustaining, dynamic rivalry.
The challenge for the Commission is to ensure that intervention remains genuinely curative or preventive, rather than valetudinarian. The aim should be to restore robust, decentralized competitive processes, not to convert digital markets into permanent objects of administrative care.
