The European Commission’s €2.95 billion decision against Google arrives amid heightened EU-U.S. trade tensions, the Trump administration’s recent criticism of EU tech regulations, internal disagreements within the Commission, and heated commentary on both sides of the Atlantic.
But beyond geopolitics, the decision also raises important legal and policy questions. In particular, the Commission’s apparent request for a structural divestiture as a remedy involves tradeoffs worthy of consideration.
The Decision: What’s New?
The Commission essentially contends that Google’s presence in multiple parts of the advertising-technology (or “adtech”) ecosystem is an inherent “conflict of interest” that creates incentives for Google to “self-preference” its own businesses to the detriment of competitors. While the full version of the decision is not yet available, one feature stands out from the press release: For the first time, the Commission has signaled that an effective remedy to an abuse of dominance might require structural remedies.
The decision appears to oblige Google to “implement measures to cease its inherent conflicts of interest along the adtech supply chain.” In practice, this request would leave Google with little choice but to divest part of its adtech business. That is, indeed, what the Commission appears to envisage.
This is remarkable. Structural remedies have to date been associated with merger control, where the Commission routinely uses divestitures to preserve rivalry. In unilateral-conduct cases, Regulation 1/2003 allows for structural remedies only where behavioral remedies would be either insufficient or more burdensome. Even then, the Commission has been cautious. Many landmark abuse cases relied on behavioral remedies, with structural separation never seriously pursued. The recent Google decision therefore marks a potentially unprecedented turn.
The paradox is that the Commission has advanced this new step while relying on an enforcement technique that was already contested in earlier cases—namely what one might call black-box enforcement. In Google (Shopping), the Commission found an abuse, imposed an obligation of equal treatment, and left Google to design the mechanics. In Microsoft, the obligation was to disclose interoperability information and submit licensing terms for approval. In each of these, the Commission did not prescribe the remedy but required the firm to propose one itself—under pain of periodic fines.
On its face, the black-box approach to remedy design is not necessarily problematic. The General Court stated in Microsoft that undertakings to which an infringement decision is addressed must “take all the measures necessary to comply with that provision, even in the absence of specific measures prescribed by the Commission.” With that said, with its recent decision, the Commission seems to apply the same model even when the implied remedy is structural. The combination of the black-box approach with an unprecedented remedy design therefore raises some questions.
The Law and Policy of Black-Box Enforcement
One question about the recent black-box-enforcement approach is whether it implies that the Commission does not understand the abuse in sufficient detail to design the cure. It is one thing to identify exclusionary effects; it is another to know what specific measures would restore competition. When the Commission stops at high-level obligations, it risks acting without a full grasp of the technical or commercial mechanics. The message appears to be: “We know this is wrong, now you fix it.”
To be sure, humility in the face of complex markets is to be applauded. Platform design is hard. Technology and competition evolve too quickly for regulators to micromanage. A principles-based order avoids the rapid obsolescence of competition enforcement. Moreover, the firm under scrutiny has superior knowledge of its systems, and likely is therefore best placed to design compliant measures. In this light, not specifying the means, but ensuring the ends, seems defensible.
With that said, the black-box-enforcement model may hand too much leverage to the Commission, amounting to “regulatory arm twisting.” Under Article 7, the undertaking faces a duty of result. If its proposals are rejected, it risks massive periodic penalty payments.
This places the firm in a bind. It must come forward with a “satisfactory” solution, but the Commission has veto power. In practice, this shifts the balance of bargaining power decisively toward the regulator, particularly where fines already run into billions. While there are legal safeguards, such as the judicial control of the proportionality of remedies, the policy tradeoffs are likely to be significant.
For example, while a principles-based obligation may grant flexibility in dynamic markets, it also risks vagueness and protracted litigation (Google has already indicated its intention to appeal the Commission’s decision). Allowing firms to propose their own fixes can produce innovative solutions. But when the implicit expectation is a structural divestiture, the line between regulatory oversight and coercion becomes blurry.
Some of these tradeoffs were already present in Google (Shopping), Microsoft, and other cases. As noted, what is new in Google Adtech is that the Commission appears to be using the same model to press for a remedy that is, in substance, structural.
By leaving it to Google to “implement measures” that would resolve conflicts of interest, while signaling that divestiture may be the only effective way, the Commission is nudging Google toward divestiture while avoiding the legal and political costs of ordering it outright. This looks like structural remedies by stealth, with the Commission wielding fines to compel voluntary divestiture without formally ordering it. One may recall the much–criticized series of commitments decisions a couple of decades ago, when the Commission “moved fast and broke things” by ordering firms to effectively divest parts of their business infrastructure under pain of formal proceedings and lengthy litigation.
There is also a constitutional angle. EU competition law has long balanced intervention against principles of legal certainty and proportionality. Structural remedies under merger control are accepted because the legal basis is conducive, and the context is preventive. Under Article 102 TFEU, the legal basis is narrower and tied to proportionality. If black-box enforcement is used to press firms into divestiture, the courts may have to decide whether this is still within the spirit of Regulation 1/2003, or whether it amounts to coercive regulation without a sufficient legal anchor. The Google Adtech case may thus be remembered not only for its substantive findings on adtech, but also for what it reveals about the evolution of the Commission’s remedial practice.
Conclusion
Is the black-box-remedy model acceptable as a matter of enforcement policy? In some cases, the answer will be yes: the Commission is simply modernizing its approach and aligning with the Digital Markets Act, which explicitly requires gatekeepers to propose compliance programs subject to approval.
In others, the answer will be no: using black-box enforcement to drive structural remedies risks undermining legal certainty and upsetting the balance of power between regulators and undertakings.
Either way, the stakes are high. Structural remedies are no longer confined to mergers. If the EU courts uphold the Commission’s approach, the latter may be emboldened to cement the black-box approach to remedy design as part of its enforcement policy, with consequences for EU competition enforcement that extend well beyond Google.
