Washington rarely gets to see the results of a policy before enacting it. On ex-ante digital regulation, it can. The European Union’s Digital Markets Act (DMA) has been in force against designated “gatekeepers” since March 2024, and two and a half years of implementation have produced a measurable record. The costs are high and were badly underestimated. The benefits are small, niche, and mostly captured by other large intermediaries rather than consumers. And the European consumers the law purportedly was written to help largely do not know it exists and do not like what it has done to their online services.
That record should weigh heavily on the American Innovation and Choice Online Act (AICOA, S. 4746), which Congress first introduced in 2021 after the EU proposed the DMA in 2020, and which was reintroduced by Senators Grassley and Klobuchar on June 10, 2026. AICOA takes the DMA’s architecture, aims it at the same companies for the same conduct, and in several respects makes it harsher.
The DMA’s Costs Are Roughly 100 Times What Was Promised
The European Commission’s 2020 impact assessment projected DMA compliance costs of about €1.41 million per platform per year. As the CCIA Research Center documented in April 2026, the actual figures run two orders of magnitude higher. Meta told the Commission’s own compliance workshop that it had spent roughly €5 billion on DMA compliance, deploying 11,000 employees and about 600,000 engineering hours. Google assigned roughly 3,000 employees for two years to a single article of the regulation. A CCIA Research Center study estimated direct DMA compliance costs at roughly $200 million per year per U.S. gatekeeper, about $1 billion per year across the five, within a broader $2.2 billion annual compliance burden from EU digital rules and up to $97.6 billion annually in combined costs, fines exposure, and revenue losses.
Compliance spending is the visible cost. The higher costs are the products Europeans no longer get. Apple has delayed iPhone Mirroring, AirPods Live Translation, and on-device Maps features in the EU because Apple says it has not found a way to meet the DMA’s interoperability mandate for those features without exposing users’ data to third parties. Google says regulatory burdens and uncertainty are delaying its newest AI features in Europe by up to a year after they launch elsewhere. Downstream, a Copenhagen Business School and LMU Munich study estimated that DMA-driven restrictions could cost European businesses that rely on gatekeeper services up to €114 billion per year in lost revenue.
The DMA’s Benefits Are Small and Don’t Reach Consumers
The DMA’s headline promises were alternative app stores, lower app prices, and more competition in search and browsers. Two years in, alternative app marketplaces have attracted fewer than 3% of EU iPhone users, and one of the six that launched, Setapp Mobile, shut down in February 2026. Apple cut its most common EU commission from 30% to 20%, but an Analysis Group study of 41 million App Store transactions found that developers kept prices unchanged or raised them more than 90% of the time, and that over 86% of the savings went to developers outside the EU. The commission reduction was a transfer to developers, most of them non-European, with no measurable consumer benefit.
Browser and search choice screens are the DMA’s most successful features, and even there the aggregate effect is small: Google still holds roughly 89% of European search, and the DMA’s messaging interoperability mandate connects WhatsApp to two little-known third-party services. Meanwhile, the anti-self-preferencing rules produced a textbook case of misdirected benefit. Google’s DMA changes to European search results cut clicks to hotel websites through Google Hotel Ads by about 30% and direct bookings from that channel by as much as 36% in affected markets relative to unaffected ones, according to hospitality analytics firm Mirai’s analysis of 3,450 hotels. Mirai identified large online travel agencies as the most likely destination for the lost clicks, pushing hotels to pay commissions to intermediaries for customers they previously reached directly.
Europeans Neither Recognize Nor Value the DMA
European consumers have not enjoyed the impacts of the DMA on their digital experiences. A Nextrade Group survey of 5,000 consumers across 20 EU Member States found that two-thirds of Europeans now need more clicks or more complex search terms to find what they want. Among frequent searchers, 61% report spending up to 50% more time per search than before the DMA. 42% of frequent travelers say flight and hotel searches have gotten worse, and 35% say Google Maps integration has deteriorated. 59% of respondents said they would pay to get the pre-DMA experience back, averaging about $300 per person per year to restore two lost features. Applied to roughly 420 million European digital-service users, that implies a consumer welfare loss of more than $70 billion per year.
An Ipsos survey of 3,500 consumers in seven Central and Eastern European countries, published by ECIPE, found that in some countries up to 85% of respondents had never heard of the DMA or knew little about it. A 39% plurality said routine online tasks now take more steps, and the authors found no evidence of lower prices, better privacy, or more contestable markets. When a regulation’s intended beneficiaries cannot identify it, and those who notice its effects want them reversed, the benefit side of the ledger is effectively empty.
The DMA Shows the Limits of Ex-Ante Regulation Generally
The DMA is the strongest test case ex-ante digital regulation is ever likely to get. It was drafted over years, enforced by a Commission with two decades of experience litigating against these same firms, and backed by fines of up to 10% of global turnover. If the model was to work anywhere, it would have been in Brussels.
It failed for structural reasons that travel across borders. Ex-ante rules prohibit categories of conduct rather than proven harms. Most conduct in those categories, such as integrating a map into a search result, setting a default, or using platform data to improve a product, is precisely what makes integrated digital services useful. Banning the category degrades the product for every user in order to reach the small subset of conduct that might be harmful, and even then the “benefit” tends to flow to the next-largest intermediary rather than to consumers or small businesses. Compliance costs scale with the ambiguity of the rules, and the rules are often ambiguous because they are often written before the conduct exists, and tend to stay static even as the conduct evolves in response to innovation and competition. Europe’s experience demonstrates the model, and AICOA proposes to import it.
AICOA Would Be the DMA with an American Accent
The coalition letter opposing S. 4746 put it plainly: the AICOA bill emulates the DMA. A law firm comparison of the 2023 and 2026 texts describes the new bill as a slightly modified version of its predecessor, with “covered platforms” relabeled “systemically important platforms” and the same prohibitions on self-preferencing, tying, data restrictions, user lock-in, and retaliation.
It targets the same companies. AICOA’s thresholds, $175 billion in average annual revenue and monthly active users equal to at least 34% of Americans over age 12, reach Alphabet, Amazon, Apple, Meta, and likely Microsoft, five of the DMA’s seven gatekeepers. It targets the same conduct. Every prohibition in Section 3(a) has a DMA counterpart: self-preferencing and ranking (Article 6(5)), use of business-user data (Article 6(2)), interoperability with platform features (Article 6(7)), tying (Articles 5(7) and 5(8)), defaults and lock-in (Article 6(3)), data portability (Articles 6(9) and 6(10)), and anti-retaliation (Article 5(6)).
In several respects, AICOA is harsher than the law it copies:
- It defines “materially harms competition” as any actual or reasonable risk of lessening competition “that is more than a de minimis amount,” a threshold that virtually any integrated product feature could trip.
- Its safety, privacy, and legal-compliance affirmative defense requires clear and convincing evidence, contemporaneous business records created at or before the time of the conduct, and proof that the conduct was not a pretext; a separate no-harm defense covers most of the bill’s prohibitions but not its core self-preferencing ban, and the 2023 version’s defense for protecting a platform’s core functionality has been deleted.
- The DMA has no minimum fine; AICOA sets a floor of 1% of a company’s total U.S. revenue for the period of the violation, alongside a 10% ceiling, and authorizes courts to order forfeiture of executive compensation after a pattern-or-practice finding.
- AICOA authorizes 120-day temporary injunctions before any final finding of liability, directs courts to endeavor to reach final judgment within a year, and lets every state attorney general sue as parens patriae, replacing centralized Commission enforcement in Brussels with federal, state, and territorial enforcers in the United States.
The Enormous Tally of AICOA’s Expected Costs
Using the most conservative estimate of DMA compliance costs, about $1 billion per year across 420 million EU users, the implied per-user cost is $2.38. Across 332 million American internet users, that is roughly $790 million per year in direct compliance costs before accounting for AICOA’s tighter liability standard, minimum penalties, and multi-enforcer litigation exposure, each of which raises the cost of a compliance posture. If American consumers valued the features they would lose as Europeans say they do, 59% of users willing to pay about $300 per year, the implied annual consumer welfare loss would be on the order of $58 billion. Against those costs, the DMA’s record offers little in the way of benefits: sub-3% alternative app store adoption, unchanged app prices for consumers, and a search market that barely moved.
Congress does not have to guess how ex-ante platform regulation performs. Europe has already run the experiment, published the survey data, and counted the costs. AICOA proposes the same design, aimed at the same firms, with a minimum penalty the DMA lacks, fewer defenses than its own 2023 version, and exposure to dozens of public enforcers. The predictable result is the European one, at American scale, and at Americans’ expense.
