September 2, 2025
Written Testimony for the House Judiciary Committee Hearing on “Europe’s Threat to American Speech and Innovation”
Chairman Jordan, Ranking Member Raskin, and Members of the Committee:
Thank you for the opportunity to submit this written testimony. My bottom line is simple: the European penchant for burdensome ex ante rules is disastrous for free speech and for innovation alike. This hearing focuses on how the ex ante regulation in the EU’s Digital Services Act and UK Online Safety Act trample free speech. Yet this prescriptive regulatory approach also tramples innovation when it micromanages competition and product design by digital platforms, as the UK’s Digital Markets, Competition and Consumers Act (DMCC) and the EU’s Digital Markets Act (DMA) do. The AI Act is the EU’s latest comprehensive regulatory scheme and applies this ex ante regulatory approach to the rapidly developing area of artificial intelligence.
The United States has historically taken a different approach to technology regulation. We aggressively prohibit government regulation of speech and enforce antitrust and many forms of consumer protection law ex post and case-by-case. The U.S. approach is strong, dynamic, and successful. The European approach is failing. It is accompanied by weak productivity growth, smaller risk capital markets, and a shrinking role in frontier AI.
Yet some in the U.S. wish to import the European approach. Legislation like the American Innovation and Choice Online Act borrows heavily from the DMA, even though it “departs in some respects from accepted principles of [U.S.] competition law and in so doing risks causing unpredicted and unintended consequences.”
Progressive state legislatures are also actively attempting to import European-style, ex ante AI regulations. Many of them explicitly borrow concepts and requirements from the EU AI Act. Colorado’s controversial (and currently delayed) AI Act, for example, copies key structures from the EU AI Act. California, Illinois, New York, and Connecticut have all pursued similar bills.
Importing the EU model now would hurt American innovation at the very moment we are racing to lead the world in AI.
The DMA exemplifies the European ex ante approach in competition policy
The DMA designates “gatekeepers” based on size and reach, then imposes a rulebook of “do’s and don’ts” in advance of any proven harm. Obligations include bans on self‑preferencing, anti‑steering rules that force platforms to promote outside offers, interoperability mandates, and limits on combining data across services. Noncompliance can bring fines of up to 10% of worldwide turnover, rising to 20% for repeat findings, with the possibility of structural remedies after a finding of systematic noncompliance.
This is classic ex ante regulation. As the Commission itself notes,
“The DMA is one of the first regulatory tools to comprehensively regulate the gatekeeper power of the largest digital companies. The DMA complements, but does not change EU competition rules, which continue to apply fully.”
These restrictions apply without considering compliance and innovation costs or any offsetting benefits of the prohibited behavior. In the DMA, the EU writes rules first and then the only question is whether the rules are being followed.
The DMA is enforced centrally by the European Commission (“EC”). The rules entered into force on November 1, 2022, and became applicable on February 5, 2023. In the past two and a half years, the EC has already opened noncompliance cases and issued the first fines. The Commission designated six gatekeepers in 2023—Alphabet, Amazon, Apple, ByteDance, Meta, and Microsoft—and later added Apple’s iPadOS and Booking.com. In April 2025, the Commission issued the first DMA penalties against Apple and Meta. That is fast, top‑down rulemaking by design.
The United States is different. The U.S. formalized modern antitrust law, and we have always taken a case-by-case, ex post approach that applies general principles (as detailed in the antitrust statutes) to specific factual findings. Those factual findings usually include detailed economic analysis as judges seek to ensure that interventions leave consumers better off. While the federal antitrust agencies have summarized some of these principles in various guideline documents, those are non‑binding guidance for how the agencies analyze cases. The decisions still turn on facts and law in specific matters. This ex post, effects‑based enforcement does not predetermine the legality of most conduct.
This difference matters because ex ante legislation locks in today’s guesses about tomorrow’s markets. Even EU institutions and analysts describe the DMA explicitly as an up‑front rulebook intended to move beyond slower case-by-case antitrust. Such rules shift the burden from the government having to prove harm to companies having to prove compliance. They distract enforcers from what actually improves market competition and consumer outcomes. And they incentivize caution over experimentation.
Europe’s ex ante approach suppresses innovation
This ex ante approach has harmed European entrepreneurs and innovators. Despite having about 30% more people than the U.S., Europe lags significantly on key measures of technological innovation and economic dynamism. One analysis concludes that, compared to the U.S., the EU has:
One-fifth the venture capital investment;
One-seventh the value of unicorns (private companies valued at more than $1 billion); and
Only 14—compared to 241 in the U.S.—large companies founded in the past 50 years.
The European Commission’s own 2025 Single Market and Competitiveness Report puts EU labor productivity at 77.8% of U.S. levels. It warns that the EU’s attractiveness as a place to build and scale firms has declined:
“Since 2008, a third of so-called ‘unicorn companies’ decided to relocate abroad. Only 4 out of the 50 largest tech companies are based in the EU and none of the EU’s most valued companies have been created from the ground up in the last 50 years, signaling a lack of market dynamism, insufficient innovation climate and high barriers to market entry and scale-up.”
No surprise, then, that while U.S. 2024 venture investment was at least $209 billion, European venture investment was roughly a quarter of that, at $45 billion.
While no single policy explains Europe’s innovation gap, these results are consistent with an ex ante system that prioritizes precaution and procedure over scale and speed. Economist, former European Central Bank president, and former Italian Prime Minister Mario Draghi released a bombshell competitiveness review of the EU that highlights the ex ante approach as core weaknesses:
“Regulatory barriers to scaling up are particularly onerous in the tech sector, especially for young companies… [T]he EU now has around 100 tech-focused laws and over 270 regulators active in digital networks across all Member States. Many EU laws take a precautionary approach, dictating specific business practices ex ante to avert potential risks ex post. … The net effect of this burden of regulation is that only larger companies—which are often non-EU based—have the financial capacity and incentive to bear the costs of complying. Young innovative tech companies may choose not to operate in the EU at all.”
Empirical work on the EU’s longest-standing ex ante technology regulation, the General Data Protection Regulation (GDPR), shows meaningful negative effects on early‑stage innovation. Studies find fewer venture financings for EU tech ventures after GDPR took effect, and large exit and entry effects in mobile apps, with consumer surplus losses. While privacy and safety are important goals, these results demonstrate how sweeping, ex ante rules can reduce investment, entry, and usage in digital markets.
The DMA is a similar structural intervention, but framed as pro-competition. Likewise, the AI Act lays out a detailed regulatory framework for a nascent industry and has received pushback from EU member states and EU companies for its negative effects on innovation. The risk to dynamism is real.
The DMA differs in one way from other European ex ante regulation like GDPR and the AI Act: it is clearly targeted at U.S. firms. Economic theory and common sense suggest that imposing regulation on U.S. companies could benefit their EU competitors. But theory and experience also predict that this would come at the expense of European consumers—see, for example, the delayed rollout of certain new products. And it harms overall competition that drives innovation. Hence the stagnation in the statistics detailed above.
More recently, there is strong evidence that the EU is falling behind in the latest area of tech innovation: artificial intelligence. For example:
U.S. private AI investment reached about $109 billion in 2024—nearly 12x China’s and 24x the UK’s (Europe’s largest).
In 2024, U.S.‑based institutions produced 40 notable AI models, compared to three for Europe.
As of May 2025, the United States hosted about three‑quarters of global GPU-cluster performance used for AI. The EU share was in the single digits.
These are outcome measures. They show where frontier AI innovation happens today. It is not in Europe.
Importing EU-style ex ante laws would be a mistake for the U.S., especially in AI
America leads in AI because we couple strong, principled antitrust and consumer protection enforcement with freedom to experiment. We punish proven harm. We do not pre‑clear product design. That balance has encouraged risk capital, talent inflows, and rapid iteration.
DMA-style regulation in the U.S. would threaten the Trump Administration’s goal of AI dominance.
First, it would slow product iteration with prescriptive design rules. The DMA’s anti‑steering and self‑preferencing bans, as well as interoperability mandates, reach deep into how platforms connect services, present choices, and monetize. Those choices are exactly where AI services are moving fastest. Conditioning innovation on a compliance checklist would raise legal risk for shipping new features and would shift engineering effort from user value to paperwork. Recent noncompliance actions and fines show how quickly design disputes become enforcement.
Second, it would chill the startup exit market. The DMA requires gatekeepers to notify the Commission of all intended mergers, even small ones, and invites added scrutiny. In practice, this reduces the speed and certainty of acquisitions that are a key part of the U.S. startup ecosystem. Slower, less certain exits mean fewer seed bets. That shows up later as fewer start-ups and fewer scale‑ups.
Third, it would misallocate scarce AI resources. The U.S. currently holds the majority of advanced AI compute. Our advantage depends on turning compute and talent into new products quickly. A pre‑emptive rulebook diverts managerial attention and engineering cycles away from model and product advances. In a global race, friction matters. Europe’s own reports warn about sluggish productivity and regulatory burdens. We should not copy them.
Fourth, it would put U.S. firms in a double bind. The DMA already targets mostly American firms. Importing that model here would layer U.S. ex ante rules on top of EU ex ante rules. The “Brussels Effect” itself is bad enough – no need to compound it with additional U.S. restrictions.
Policy recommendations
Keep the U.S. model. Preserve case‑by‑case antitrust and avoid EU‑style ex ante digital codes. Enforcement agencies, state Attorneys General, and private parties can and do bring cases when facts and law support them. Congress should not replace that system with a rulebook that presumes harm.
Preempt EU-style state AI laws. Progressive legislatures in states like California, New York, and Illinois are copying Europe with ex ante laws that have extraterritorial effects on all U.S. innovators. Congress needs to set the direction and clear the way on this national priority.
Expand the inputs to innovation. Increase U.S. leadership where it matters: compute, energy, talent, and capital formation. Faster permitting for data centers, reliable baseload power, flexible high‑skilled immigration, and deep private capital markets will deliver more for consumers than conduct codes. The data on AI investment, model production, and compute share show where we win and what to reinforce.
Use a scalpel, not a sledgehammer. When there is specific conduct that truly harms competition or consumers, use existing tools. If Congress legislates, target narrow problems with measurable consumer harm and clear benefits. Do not adopt an open‑ended list of global design mandates.
Encourage American scale‑ups and exits. Ease the path for startups to go public or be acquired by non‑dominant buyers. Do not create broad acquisition presumptions against “big buyers.” Doing so would block the only realistic exit for many founders and stall what is often the fastest way to scale innovation to benefit millions.
Conclusion
Innovators win when law enforcers punish proven harm and otherwise get out of the way. European regulators currently have a different philosophy. They write the laws first and hope innovation happens anyway. That hope has proven unjustified.
Still, Europe is free to make that choice for its own companies and citizens. The United States should push back when Europe’s choices harm our citizens and our companies. And we certainly should not import that “regulate first” mindset to our shores.
The American approach has fostered the world’s deepest investment markets, most capable technology companies, and most advanced AI labs. We should double down on what works. We should not import the EU’s precautionary approach at the very moment when speed, scale, and experimentation will decide who stays at the frontier of technological innovation.