Last year, 22 of the 30 platforms subject to the obligations of the European Digital Markets Act were designated as “gatekeepers” by the Commission. Each one conditions access to hundreds of millions of users, sets interoperability rules at its discretion, and determines in practice who can sell what to whom. It is on this terrain that Jean Tirole, 2014 Nobel Prize winner in Economics, proposes today a simple and universal rule: “fair gatekeeping.” The idea is elegant. The problem is that three blocs are each writing their own version of this rule, and none of them has an interest in adopting the others’.

The Essentials

  • The European Digital Markets Act, which came into force in March 2024, designates 22 platforms as “gatekeepers” and imposes interoperability and fairness obligations without binding equivalents in the United States or China.
  • Jean Tirole, in a recent working paper from the Toulouse School of Economics, proposes a principle of “fair gatekeeping”: a dominant platform cannot charge third-party actors more than what real costs and network effects justify.
  • The United States has chosen the judicial route on a case-by-case basis (DOJ cases against Google, FTC against Meta); China has developed a national regulation of platforms that predetermines no international harmonization.
  • Fragmentation into three incompatible regimes creates a risk of lasting lock-in: the interoperability standards set in the next fifteen years will predetermine access to global markets well beyond the digital sector.
  • The real leverage of indebted states facing platforms that finance their bond markets remains an open question, even when regulatory will is present.

What Tirole Puts on the Table

Tirole’s reasoning starts from a simple observation: a dominant platform is not a business like any other. It is not a natural monopoly in the classical sense either. It is an intermediary whose value rests on the number and diversity of actors participating in it, and whose power stems precisely from this central position. Applying to these actors the utility regulation of the last century or leaving them to self-regulate in the name of Schumpeterian competition are two symmetric errors.

What Tirole proposes, in his working paper published by the Toulouse School of Economics, is called “fair gatekeeping.” The principle is as follows: a platform that controls access to a market can collect remuneration for this service, provided that this remuneration reflects its real costs, the value it creates for both sides of the market, and the network effects it generates — and nothing beyond. Below this threshold, the rule does not apply. Above it, it requires either a reduction in fees or opening to interoperability.

It is an economic rule, not a political one. It does not say that platforms are bad or that they should be broken up. It says that their market power, where it exists, must be framed by a logic of reference pricing built on objective data. The analogy with the regulation of electricity or telecommunications networks is not coincidental: in both cases, the question was not about nationalizing the network, but about defining at what price third parties could access it.

This conceptual framework arrives at the right time. And at the wrong time.

The DMA: First Global Standard by Default

The European Union has taken a lead. The Digital Markets Act, adopted in 2022 and fully entering into force in spring 2024, is the first binding regulation in the world to impose ex ante obligations on dominant platforms. It does not wait for abuse to be proven to act — it defines in advance the prohibited behaviors for anyone exceeding the designation thresholds (45 million active monthly users in the EU, 10,000 professional user companies, market capitalization of 75 billion euros).

The obligations are precise: access to self-generated data, messaging interoperability, prohibition of self-preferencing, data portability. Apple, Alphabet, Meta, Microsoft, Amazon, and ByteDance were designated. The first infringement procedures were opened in 2024, notably against Apple on the App Store and Alphabet on Google Search.

Europe has thus created, in fact, a global standard by default. Companies operating on its territory — most of them American — must comply with it, whether they wish to or not. This is what legal experts call the Brussels Effect: when a market of 450 million consumers imposes its rules, global companies often adopt them beyond the territory where they apply, because managing two systems costs more than implementing one.

But the Brussels Effect has a limit. It works when the targeted companies need to access the European market and do not have equivalent countervailing power. It works less well when the United States actively defends its companies, and even less when China develops its own ecosystem, structurally impermeable to Western standards.

Two Models That Will Not Converge on Their Own

In the United States, the choice has been different. The American antitrust tradition is procedural and retrospective: one proves abuse, litigates, and renders judgment. The Department of Justice obtained in 2024 a first historic decision against Google on search and the advertising market. The FTC has initiated proceedings against Meta over the acquisition of Instagram and WhatsApp. These judicial victories are real. But they concern past conduct, they take years, and they do not create a general rule applicable ex ante to all platforms.

The Trump administration, returning to power in January 2025, has made explicit what was implicit in the previous ones: American platforms are national strategic assets. Regulation that constrains them too quickly or too heavily in commercial negotiations with Europe is seen as an advantage conceded to China. This logic is not absurd — it is actually coherent with a vision of geopolitical competition. It simply makes transatlantic convergence toward a common standard much more difficult.

China, for its part, has followed its own trajectory. It has developed aggressive national regulation of platforms — Alibaba, Didi, Meituan were heavily sanctioned between 2020 and 2023 — but in the name of internal political objectives: breaking up the concentration of private power, protecting data sovereignty, preventing the emergence of autonomous economic countervailing forces. This is not market regulation in the sense of Tirole. It is state regulation, whose criteria are not exportable and whose logic is not interoperable with that of the DMA or American antitrust.

We thus find ourselves with three regimes each applying to their own space, coexisting without speaking to each other, and whose fragmentation is already producing concrete effects on market access. Mexican nearshoring illustrates what it costs to lack solid institutions to stabilize a comparative advantage: digital platforms risk following the same trajectory on a global scale.

The Real Leverage of Indebted States

There is an angle that Tirole’s theoretical framework, precisely because it is economic, does not address head-on: the balance of power between the regulating states and the platforms they claim to regulate.

The work of François Ecalle on French and European public finances raises an uncomfortable question. Eurozone states display public debt exceeding 100% of GDP on average — France exceeds 113%, Italy 137%. The bond markets on which these states refinance their debt are partly animated by the same large financial institutions that are themselves clients of the cloud infrastructure and data services of the platforms. The link is not mechanical. But the dependency is real.

More directly: Apple, Alphabet, and Microsoft hold combined treasuries exceeding 400 billion dollars — more than the foreign exchange reserves of most European states. When the European Commission fines Apple 13 billion euros for undue tax advantages in Ireland, the procedure takes eight years and ends partially before the Court of Justice of the European Union. When it opens an infringement procedure under the DMA, the investigation period is twelve to eighteen months, the sanction is capped at 10% of worldwide turnover, and the targeted companies have legal resources without comparison to those of the regulatory authorities.

This imbalance of means is known. What aggravates it is structural: the platforms have time on their side. Each year of delay in implementing a rule is an additional year to entrench network effects, expand the user base, make platform switching more costly for everyone. Tirole’s rule may be economically correct. But one still needs to have the institutional capacity to enforce it quickly, at scale, and without being exhausted by litigation.

Interoperability as the Frontier of the Century

Let us step back. What is at stake in the debate over “fair gatekeeping” is not only a question of access fees to App Stores or messaging interoperability. It is the question of who will control the infrastructure for access to global markets in the next fifteen to twenty years.

Interoperability standards are to the digital economy what container standards were to maritime trade in the 1960s. When the world agreed on the dimensions of Maersk containers, it simultaneously decided on the geography of the ports that would dominate global trade for the following fifty years. The standards for interfaces between digital platforms, between AI systems, between payment protocols, between digital identities — all of this is freezing. Not because we decided it. Because no one agreed on anything else.

The stakes go beyond the digital sector. Stablecoins are dollarizing the emerging world precisely because the United States did not regulate its emission protocol in a way to make it neutral: it allowed private actors to create a standard that extends, de facto, the sphere of the dollar without international agreement. The same mechanism is at work in platforms: a non-negotiated technical standard is an undeliberate geopolitical act.

This is where Tirole’s proposal recovers its strategic dimension. A “fair gatekeeping” rule built on objective economic criteria — real costs, measured network effects, calculable reference prices — could serve as a basis for negotiation between blocs, just as WTO agreements on tariffs provided a common language for the trade in goods. The principle is not that each country abandons its regulatory sovereignty. It is that there exists a common floor below which fragmentation is considered a distortion of competition, just as tariff dumping is.

This convergence is difficult. It is not impossible. The Basel Accords on bank capital ratios took fifteen years to negotiate and are today applied in more than one hundred twenty jurisdictions. IFRS accounting standards, adopted by more than one hundred forty countries, have imposed a common financial language without eliminating national variations. The precedent exists.

The digital economy, however, has a characteristic that finance and accounting do not have: its network effects render the dominant standard self-reinforcing at a speed that international negotiations cannot keep pace with. The risk is not that the blocs never find an agreement. It is that they find it too late — after the architectures of market access are already frozen, and negotiation concerns only the margins.

What Is at Stake Now

The European Commission scheduled its first review of the DMA for 2026. It will evaluate whether the obligations imposed have actually reduced the power of gatekeepers, or whether the platforms have found clever workarounds to satisfy the letter while circumventing the spirit. The first results on messaging interoperability — WhatsApp is technically required to open its protocol to third-party applications since March 2024 — are mixed: the opening exists, but the technical friction remains high enough to discourage migration.

This is the empirical test of Tirole’s framework. Not whether the rule is theoretically correct — it is — but whether it is institutionally implementable under the real conditions of digital regulation: states with asymmetric means, companies with superior legal and technical resources, network effects that move faster than procedures.

The way science itself adapts to the acceleration of AI systems offers an unexpected parallel: the challenge is not to produce the right rules, but to build the institutions capable of evolving them at the same pace as the systems they regulate.

The question is not whether a common standard is economically desirable — Tirole demonstrates this convincingly. The question is whether the actors who would have the means to negotiate it still have an interest in doing so, and whether the actors who have an interest in defending it will have sufficiently early on the institutions capable of making it credible.

Sources

  1. Jean Tirole, “Fair Gatekeeping in Digital Ecosystems,” Toulouse School of Economics Working Paper — https://www.tse-fr.eu/publications/fair-gatekeeping-digital-ecosystems
  2. European Commission, Regulation (EU) 2022/1925 on Digital Markets (Digital Markets Act) — https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32022R1925
  3. European Commission, list of gatekeepers designated under the DMA — https://digital-markets-act.ec.europa.eu/gatekeepers_en
  4. Department of Justice, United States v. Google LLC (search and advertising case) — according to DOJ, first-instance decision rendered in 2024
  5. FTC, proceeding against Meta Platforms Inc. — according to the Federal Trade Commission
  6. Eurostat, gross public debt of Eurozone member states, 2024 — according to Eurostat
  7. François Ecalle, analyses on French public finances — Fipeco, https://www.fipeco.fr