Key Takeaways

Contents

Introduction. 2

Convergence and Divergence Between U.S. and EU Antitrust 4

Reconvergence: The Brussels Model and U.S. Antitrust Populists 6

An Inconvenient Fact for the Brussels Model 7

The Overlooked Error Costs of “Big is Bad” Antitrust 9

The End of the “End of History” in Antitrust 12

Endnotes 14

Introduction

Over the past several decades, antitrust regimes have proliferated around the world. In the 1980s, fewer than 30 countries had antitrust or competition laws, whereas today, over 130 countries have them.0F[1] While the United States pioneered antitrust law with passage of the Sherman Act in 1890, the wave of antitrust globalization over the past half-century has tended to rely on European competition law as the normative model. By contrast, the U.S. antitrust model, as it has developed through judicial decisions and regulatory guidelines, faces significant challenges to its global intellectual leadership. Domestically, that model is also challenged by U.S. policymakers who endorse the revival of a populist approach that is skeptical of large firms on principle and seek to deploy antitrust for a variety of redistributive, structuralist, and other goals that go beyond, and may conflict with, the preservation of competitive markets.

While EU courts and regulators have sometimes adopted certain methodological or substantive elements of U.S. antitrust, the EU and U.S. approaches ultimately reflect different understandings of the role of competition law in market-driven economies. Whereas EU competition law has tended to reflect an “ordoliberal” tradition that seeks to preserve decentralized market structures and “fair” competition, U.S. antitrust law since at least the late 1970s has reflected a commitment to market efficiency and consumer welfare by targeting practices that impede the free play of competitive forces.

In digital markets in particular, EU competition enforcement arguably reflects a structuralist bias against large firms and high market shares as a matter of principle, accompanied in some cases by calls for “digital sovereignty” that raise geopolitical implications.1F[2] Under EU competition law, regulators at the EU and national levels have principally targeted large U.S. “Big Tech” firms, resulting in over 60 antitrust cases as of year-end 2025 and massive fines in just the past three years alone: $4.38 billion in 2025, $6.7 billion in 2024, and $2.03 billion in 2023.2F[3]

These interventionist tendencies culminated in the enactment in 2022 of the Digital Markets Act (DMA), which essentially adopts per se liability rules for certain practices undertaken by the largest digital platforms—all but two of which are based in the United States.3F[4] Implementation of the DMA has put in place a quasi-administrative process involving repeated interactions between regulators and platforms, appeals of regulatory determinations, and periodic submissions of compliance reports, compounded by an overlapping grid of EU-level and national-level enforcement authorities.4F[5] The DMA has already yielded significant results: in 2025, EU regulators assessed fines under the statute of approximately $570 million and $200 million against Apple and Meta, respectively.5F[6]

The DMA’s ex ante enforcement approach, implemented through categorical prohibitions of certain practices when undertaken by the largest digital platforms, contrasts with the continued ex post and evidence-based enforcement approach that characterizes U.S. antitrust jurisprudence and agency policy. Starting with the Supreme Court’s landmark 1977 decision in Continental T.V. Inc. v. GTE Sylvania Inc., the federal judiciary has consistently rejected antitrust suits that cannot show evidence of harm to competitive markets, outside cases of explicit collusion.6F[7] Vigorous debates have taken place during this period within an analytical framework anchored in the objective of preserving competitive markets through fact-intensive assessments of contested business practices.

The EU’s new regulatory approach toward antitrust enforcement has elicited admirers and imitators around the world through what has become known as the “Brussels effect.” In the view of many policymakers and scholars globally, the EU model has offered a superior alternative to the conventional U.S. approach, whose prudential insistence on robust evidence of harm to competition (and associated emphasis on avoiding false-positive error costs) purportedly facilitated the rise of tech platforms in the digital economy. To avoid this predicament going forward, EU regulators and like-minded commentators have argued that competition law should be enforced preemptively against dominant platforms.7F[8]

The global transformation of antitrust law into a tool of social policy has had a boomerang effect. Some U.S. policymakers and commentators now seek to refashion antitrust based on elements of the EU competition-law model and its culmination in the quasi-regulatory regime implemented through the DMA.

So far, however, attempts to mimic the DMA’s preemptive approach and per se prohibitions through legislation have failed in Congress, as did the Federal Trade Commission’s (FTC’s) effort to undertake rule-making to target practices deemed to be “unfair methods of competition.”8F[9] (The current FTC withdrew its objections to a lawsuit challenging its rule concerning noncompetes and appears to have abandoned the previous administration’s rule-making initiative.)9F[10] Still, while dismissing decades of antitrust jurisprudence and sometimes even rejecting the relevance of economic analysis, many voices in the United States now advocate using antitrust law to promote redistributive, structuralist, and other goals that stand outside the preservation of competitive markets, including reducing concentrations of political power, empowering labor, protecting small business, and promoting sustainability.10F[11]

Rhetoric in the Biden administration often adopted a populist approach to antitrust enforcement, while some enforcers in the Trump administration occasionally use similar language. Jonathan Kanter, head of the Antitrust Division under President Biden, argued that antitrust law is a vehicle for achieving certain redistributive goals and should be “enforce[d] … not merely for the institution of antitrust and competition law.”11F[12] Crossing political boundaries, both Lina Khan, the past chair of the FTC, and current FTC Commissioner Mark Meador have essentially suggested that economics should play no more than a supplemental role in antitrust enforcement and adjudication.12F[13] While Gail Slater, the recently departed head of the Antitrust Division, and Commissioner Meador have rejected the EU’s ex ante regulatory model, they have sometimes endorsed what appears to be a modified “big is bad” principle that supports taking action against dominant platforms and other large companies to mitigate concentrations of economic and political power.13F[14] Outside the retention of the revised merger guidelines, however, it is not clear to what extent this rhetoric has impacted agency action under the current administration.

Efforts to incorporate European concepts into U.S. antitrust law also exist at the state level. New York and California have been considering amendments to state antitrust statutes modeled on the abuse of dominance cause of action under EU competition law.14F[15] In December 2025, the California Law Revision Commission released a draft final recommendation in favor of a monopolization cause of action that departs significantly from Section 2 of the Sherman Act and which, in a limited form, has now been proposed as legislation.15F[16] The proposed changes identify “fair competition” as a legislative objective, bar consideration of efficiencies for other persons or in other markets, and relieve plaintiffs from defining a relevant market in certain circumstances.16F[17] The recommendation also includes a cause of action for a unilateral “restraint of trade,” which apparently lacks any requirement to show market power and could therefore be used by courts to reinstate long-abandoned per se prohibitions on vertical restraints. If enacted, the California proposal would arguably go beyond the EU abuse of dominance principle and would be among the most interventionist antitrust legislation in the world. Indeed, California has already enacted new laws banning “common pricing algorithms” as part of an agreement “to restrain trade” and instituting a premerger notification system for businesses based, or with significant sales, in California.17F[18]

This policy brief argues that refashioning U.S. antitrust based on the weaker probative burdens and categorical prohibitions of the EU competition-law model is likely to adversely impact the global competitiveness and innovative dynamism of the U.S. economy and the U.S. tech ecosystem in particular. The same is true for efforts by certain U.S. policymakers and like-minded commentators and advocates, at both the federal and state levels, to turn back the historical clock and revert to the “big is bad” intuitions that once drove U.S. antitrust jurisprudence. The same is also true for other countries that seek to excel in the global innovation economy. Contrary to the views of many regulators and some scholars, the EU’s structuralist approach, especially as set forth in the DMA, is unlikely to provide a sound model for antitrust law in a dynamic and innovation-driven market economy.

Convergence and Divergence Between U.S. and EU Antitrust

Since at least the late 1970s and early 1980s, U.S. antitrust law has been anchored in the consumer welfare standard as a matter of substance and economic analysis as a matter of methodology. This common framework has persisted across administrations and cuts across the Chicago, Harvard, and post-Chicago schools of antitrust thought.18F[19] Endorsed by the Supreme Court in a 1979 decision, adopted by the antitrust agencies in the 1982 Department of Justice Merger Guidelines, and refined in the 2010 Horizontal Merger Guidelines, the consumer welfare standard directs regulators and courts to assess any contested practice in terms of its actual or likely impact on competitive conditions, as measured by effects on price, output, quality, or innovation.19F[20],20F[21],21F[22]

This approach has at least four core features:

1. Outside cases of outright horizontal collusion, it generally deploys some form of a rule-of-reason-style analysis (ranging from “quick look” to more fully developed varieties) to assess net anticompetitive effects.22F^[23]

2. It requires some proof of consumer harm in unilateral conduct cases.

3. It excludes social harms that do not involve injury to competition.23F^[24]

4. It is agnostic concerning optimal market structures.24F^[25]

The EU approach to competition law shares certain features of this model but differs in significant ways. Like the U.S. approach, EU regulators and courts take into account evidence concerning likely impacts on price, output, quality, and innovation, although the probative threshold under the abuse of dominance cause of action is typically easier to satisfy as compared with the monopolization cause of action under the Sherman Act. Most fundamentally, EU competition law historically reflects the influence of an ordoliberal understanding of competitive markets that seeks to preserve decentralized market structures, rather than a consumer-welfare approach that prioritizes economic efficiency and is agnostic on market structure.25F[26] These substantive differences can account for the greater use by EU courts of “form-based” tests for single-firm conduct, as compared with U.S. antitrust case law, which typically applies some version of the rule-of-reason test outside cases of explicit collusion.26F[27]

Not so long ago, it had appeared that EU competition law was moving in the direction of U.S. antitrust law. In 2001, EU Competition Commissioner Mario Monti essentially endorsed consumer welfare as the objective of competition policy and stated that “U.S.-EU convergence is a key building bloc for a multi-lateral cooperation in antitrust,” and “we share the same goals … a competition policy which is based on sound economics and which has the protection of consumer interest as its primary concern.”27F[28] Concurrently, EU courts and regulators had made greater use of “effects-based” tests that analyze the conduct’s effect on competition, including procompetitive justifications to some extent.28F[29] At least in digital markets, this convergence has been halted by EU regulators’ embrace of a digital sovereignty narrative and aggressive response to the rise of U.S.-based digital platforms, culminating in the categorical prohibitions and preemptive approach set forth in the DMA.

The DMA adopts two key presumptions that depart from the U.S. antitrust model and, by design, certain elements of EU competition law. First, the DMA enables regulators to designate the largest platforms as “gatekeepers,” which are deemed to hold durable market power in specified “core platform services,” and impose special obligations on them. Second, the DMA largely prohibits “covered platform practices” engaged in by gatekeepers in those core platform services markets and precludes consideration of offsetting efficiencies, reflecting the view that those practices are almost certain to harm competition. This approach resembles the per se liability rules that had once characterized much of U.S. antitrust law and, as scholars documented extensively, often suppressed practices that promoted competition.29F[30] The DMA goes further and implements these per se prohibitions through an ex ante regulatory apparatus that preemptively targets the largest platforms, as opposed to evaluating anticompetitive conduct on a case-by-case basis.30F[31]

By design, the DMA avoids the risk of underenforcement in digital platform markets. However, given the absence of any balancing-test analysis, it necessarily elevates the risk of substantial overenforcement. This blunt approach reinstates the risk of false-positive error costs that U.S. antitrust has sought to avoid through the careful construction of tailored rule-of-reason tests.

Those error costs can have considerable impact on global technology markets, given the size of the European market and the firms that have already fallen within its regulatory scope. As of this writing, there are seven entities designated as gatekeepers under the DMA, which collectively constitute approximately one-third of the market capitalization of all companies included in the S&P 500.31F[32] Five are large U.S. platforms and the remainder are ByteDance (based in China) and Booking.com (based in the EU). The DMA entrusts regulators with exceptionally powerful remedies to target offending platforms. For violations of the DMA, regulators may assess fines representing up to 10 percent of a company’s worldwide annual revenues and up to 20 percent in the case of a repeat offense, in addition to behavioral and structural remedies.32F[33]

Reconvergence: The Brussels Model and U.S. Antitrust Populists

The EU competition policy model, including its culmination in the DMA, is widely viewed among regulators and a significant portion of the scholarly community as a pioneering way to address risks to competition in digital markets. Indeed, both progressive and conservative antitrust populists among U.S. policymakers (including two U.S. senators who have published books on this issue) argue that large digital platforms pose a risk not only to competitive markets but also to democratic governance and freedom of speech due to the concentration of economic power and political influence.33F[34] Some of these concerns over industrial concentration are reflected in elements of the revised merger review guidelines issued by the Department of Justice (DOJ) and FTC under the Biden administration, which lower the “Herfindahl-Hirschman Index (HHI)” thresholds that trigger a rebuttable presumption of anticompetitive harm and limit significantly the ability to overcome a merger challenge by demonstrating countervailing efficiencies.34F[35] Current DOJ and FTC leadership has endorsed the revised guidelines.35F[36] And, as discussed, a California state-law proposal under serious consideration embraces “fairness” as a policy objective of antitrust law and, in certain respects, even goes beyond the EU competition policy model in addressing single-firm conduct.

This resurgent structuralist approach among certain U.S. policymakers and thought leaders in the academic and advocacy communities not only cuts across U.S. political boundaries, but also shares surprising commonalities with the European ordoliberal tradition insofar as it suggests that antitrust law should be deployed to preserve decentralized markets as a matter of principle, as distinguished from a structurally neutral approach that assesses contested practices based on demonstrated effects on competitive conditions.36F[37] This convergence of views, encompassing ordoliberals, progressive antitrust populists, and conservative antitrust populists across the political spectrum, supports various models of competition law that pursue redistributive, structuralist, or other fairness-related goals. All those models depart from the consumer-welfare model that has been in place in U.S. case law and agency guidelines for about a half-century and places antitrust law at a critical juncture not observed since the “Chicago turn” in the late 1970s.

An Inconvenient Fact for the Brussels Model

The EU model, and in particular the DMA, exerts considerable influence over both scholarly discussion and regulatory implementation of competition policy throughout the world. The United Kingdom has adopted legislation that implements an ex ante enforcement approach in digital markets (the Digital Markets, Competition and Consumers Act of 2024) and, in the smartphone market, so too has Japan (Act on Promotion of Competition for Specified Smartphone Software). DMA-like legislation is currently being considered in many other countries including Australia, Brazil, India, South Korea, and Turkey.37F[38] Under the Biden administration, certain elements of the EU model were reflected in legislation proposed in Congress (in particular, the American Innovation and Choice Online Act, which would have banned practices such as self-preferencing by dominant digital platforms), but which has so far stalled.38F[39]

Proponents of certain elements of the EU model, both in the United States and elsewhere, as well as antitrust populists among U.S. policymakers across the political spectrum rely on the assumption that the conventional U.S. antitrust model is a policy failure that has facilitated the entrenchment of a small number of leading digital platforms. However, this assumption must confront an inconvenient fact that suggests that it is the EU model, and specifically that model’s structuralist bias against large entities and concentrated markets as a matter of principle, that is more likely to lead an economy down the path of policy failure.

The EU model, and in particular its emphasis on preserving decentralized markets irrespective of actual impact on market efficiency and consumer welfare, appears to have a poor track record in supporting the dynamic forces of entrepreneurship and innovation that characterize the most robust tech economies. Indeed, since the late 1970s, a natural experiment in competition law has occurred. While the EU has tended to rely on structuralist presumptions and categorical prohibitions that seek to preserve decentralized market structures and consumer choice across multiple providers as a matter of principle (which culminates in the DMA’s differential treatment of the largest digital platforms), the United States has adopted a context-specific approach that in most cases demands evidence of harm to consumers before a violation may be found.

Unlike the EU model, the U.S. approach (at least for the past half-century) is agnostic concerning ideal market and firm structures and has specifically recognized that economies of scale and scope arising from large firm size (which are typically associated with increased concentration) can sometimes benefit consumers.39F[40]

This was the key, albeit now somewhat neglected, insight in landmark works published in the early 1970s by leading economists Harold Demsetz and Oliver Williamson. Both rejected form-based approaches to antitrust enforcement, arguing that judges and regulators should focus on effects, not form, in assessing the potential anticompetitive effects of contested business practices.40F[41] Starting at least with the landmark GTE Sylvania decision in 1977, U.S. antitrust case law has reflected the insights of Williamson and like-minded scholars by decreasing reliance on structural presumptions and, outside cases of explicit collusion, progressively discarding virtually all per se liability rules.41F[42] While it is true that this approach increases the costs of litigation, it also provides a tool to distinguish between practices that constitute competition on the merits and those that merely exclude competitors without any legitimate business justification.

This divergence in approaches toward competition policy happens to track a divergence in economic and innovation performance in European and U.S. technology markets. During the 50-year period dating from the emergence of the now-conventional U.S. model starting in the late 1970s, the United States has produced world-leading technology platforms while supporting a broad-based information technology sector encompassing software, chip design, and hardware companies. Moreover, those technology platforms only emerged during the 1980s and 1990s and successfully challenged IBM, Xerox, RCA, and other now-lapsed incumbents. In critical industries such as telecommunications, computing, semiconductors, and pharmaceuticals, breakthrough innovation in the U.S. economy tended to shift away from the large vertically integrated corporations that had dominated postwar markets (a period that coincides with structuralist approaches to antitrust enforcement) to start-ups and smaller firms that partnered with, or challenged, incumbents.42F[43]

Famously, Joseph Schumpeter argued that market economies uniquely generate value through a “perennial gale of creative destruction” that challenges existing business and technological paradigms.43F[44] Contrary to the widely expressed view among certain policymakers that digital markets converge on entrenched monopolies, the U.S. economy has experienced repeated episodes of Schumpeterian creative destruction throughout the time of the purported policy failure attributed to the consumer-welfare model of antitrust enforcement. Tech giants formed in the early years of the digital economy, such as AOL, Yahoo!, and Blackberry, were overtaken by more innovative challengers. Today that cycle continues: Google now faces stiff competition in the answer-engine and search-adjacent markets from ChatGPT, Claude, and other AI-enabled chatbots; Facebook faces competition from TikTok, YouTube, and X; Amazon faces competition from Walmart (which is rapidly expanding online) generally and, in apparel, Shein and Temu; and Intel has struggled in the face of competition from both Nvidia, which has reinvented itself as a chip producer for the artificial intelligence market, and Qualcomm in mobile communications.

During this period, Europe has largely ceased to be a significant player in information technology markets, outside telecom infrastructure and certain chip design segments, although it remains a leader in biopharmaceuticals. In particular, the European market has failed to produce a single leading technology platform akin to Google, Amazon, Apple, Meta, or Microsoft, while its existing significant competitors in the information technology market, such as Siemens, Philips, Ericsson, and Nokia, have maintained more modest market positions as compared with U.S.-based tech companies. During this approximately 50-year period, Europe has experienced a growing innovation deficit compared with the United States and East Asia, which was identified in the influential Draghi Report as a central factor behind its productivity gap with the United States.44F[45] These observations should provide significant reason for caution in embracing what some U.S. policymakers and an increasing number of scholars laud as the EU’s “tough” approach toward competition-law enforcement.

The Overlooked Error Costs of “Big is Bad” Antitrust

To be clear, there are multiple factors that lie behind the divergence in economic outcomes and innovation performance between the United States and EU. However, there are several reasons to believe that the rigid “big is bad” approach, as well as associated preference for form-based categorical prohibitions rather than effects-based balancing analysis that tend to characterize EU competition law (with certain exceptions) and, to an even greater extent, the DMA, impose constraints that can inhibit a robust tech economy.

Why Big Is Sometimes Good

The European ordoliberal tradition’s idealization of small and medium-sized firms, as well as nostalgia among U.S. antitrust populists for a Jeffersonian economy of small farmers and merchants, can impose a significant cost in terms of underinvestment, under-innovation, and, in some cases, higher prices or reduced convenience for consumers. Specifically, a structuralist approach that condemns firms based solely or mostly due to size and conduct that increases market share, without a case-specific showing of adverse impact on competitive conditions and consumers, can inhibit firms from achieving the scale necessary to maximize cost savings for consumers or to assemble integrated product-and-service ecosystems that deliver transactional synergies for users. As shown in a landmark paper by recent Nobel prize winner Philippe Aghion and colleagues (and building on Schumpeter’s insights), market power up to a certain threshold can promote innovation by providing the incentives and revenue streams necessary to support R&D investments that may only yield new products after a considerable time lag and at a high risk of technical or commercial failure.45F[46]

These points are especially critical in platform markets, where scale and scope are both critical to achieving economic viability and delivering transactional efficiencies to users, including through network effects. Specifically, in platform markets, there is often a dramatic difference between fixed costs and marginal costs (sometimes almost zero), generating the opportunity to achieve economies of scale by spreading fixed costs over large production volumes. Subject to competitive conditions, some of those cost savings are passed on to consumers.

To illustrate the importance of scale, consider Google Android, the most widely used operating system in the global smartphone market. As a “free” operating system, Android effectively passes on scale-driven cost savings to consumers while extracting revenues from paying advertisers and in-app commissions. And to illustrate the importance of scope, consider the Google Workspace office suite. Through a series of acquisitions, Google created an integrated bundle of office productivity software applications that reduces transaction costs for users and now poses strong competition to what seemed to be the unbeatable MS Office suite.46F[47] In both cases, achieving economies of scale or scope have enabled efficiencies that deliver lower costs and yield transactional synergies for consumers and, in the case of the mobile operating systems and office software markets, challenged a once dominant incumbent.

To be clear, start-ups and other smaller R&D-intensive firms often play important roles in tech ecosystems as a source of disruptive innovation.[48] Yet a robust tech ecosystem demands a diversity of transactional forms, often consisting of large scale-efficient platforms that rely on a steady influx of smaller firms as a source of both complementary technologies, as illustrated by Google Workspace, or as a source of competitive challenge, as illustrated by the recent successful entry of OpenAI and Anthropic into the search ecosystem. There is no “one size fits all” for efficient market structures in any particular market, and only a case-specific approach to antitrust enforcement can allow competitive forces to adapt organizational forms in response to changing economic and technological conditions.

Are Digital Platforms Really Different?

Some commentators might argue that the DMA appropriately errs on the side of overenforcement since digital platforms are especially prone to acquire market power and then use business practices to deter entry by potential challengers. However, as it is far from a settled question that dominant tech platforms systematically enjoy entrenched leadership positions, a case-by-case inquiry would seem to be the appropriate decision rule. As noted previously, numerous examples can be provided of seemingly secure tech giants that rapidly lost market leadership to more innovative newcomers.47F[49] In other cases, incumbent platforms may compete with each other when entering new markets where no clear leader has yet been established.48F[50] This appears to be the case in the cloud computing market, where Amazon, Microsoft, Google, and more recently Oracle all maintain significantly sized offerings with comparable capacities. Even though each company has significant revenues, it is not clear that any individual firm can exert pricing power in the face of stiff competition and the ability of users to multi-home across several providers.49F[51]

The search ecosystem illustrates why it can be erroneous to presume that large market shares in a certain relevant market are immune from challenge. It may appear to be obvious that Google enjoys an unbeatable near monopoly in search services. Yet, this is not so obvious if the relevant market is expanded beyond general-purpose search to include specialized search services, such as travel and e-commerce, where Google has failed to secure leadership, or the chatbot/answer-engine market, where Google still lags behind OpenAI by a significant margin and Anthropic now offers stiff competition. Moreover, there is abundant capital to fund entry by independent answer-engine firms, such as OpenAI  and Anthropic, and answer-engine services offered by other platforms, such as Meta (Meta AI), and X (Grok). These developments indicate significant competition in the search and answer-engine ecosystem and are consistent with periodic challenges to incumbents throughout the history of U.S. technology markets. If that is the case, then the regulatory design of the DMA, and the powerful remedies it authorizes, may rely on an incomplete and in some cases inaccurate understanding of digital markets.

Why Interventionist Antitrust Is Self-Defeating

A signature characteristic of U.S. antitrust law since the early 1980s has been an insistence that, outside cases of explicit collusion, antitrust only supports intervention where there is evidence of harm to competition. This approach ensures that harm to competitors is not equated with harm to consumers. Both the European and U.S. varieties of structuralist antitrust depart from this approach and, in markets characterized by scale and scope economies, may take enforcement actions that seek to cultivate decentralized markets but increase prices, decrease quality, or discourage innovation. In other circumstances, a structuralist approach can even yield outcomes that harm both consumers and small businesses—a policy “lose-lose” without any offsetting trade-offs.

This outcome can also be illustrated by policy actions taken by U.S., U.K., and EU regulators on the basis of the “killer acquisitions” theory put forward by certain antitrust scholars.50F[52] This theory proposes that dominant firms have an incentive to acquire start-ups that may pose a competitive challenge and then shut them down. Based on a single empirical study showing that this strategy may characterize about 5 percent of acquisitions in the pharmaceutical market, certain scholars have generalized the theory to platform/start-up acquisitions in digital markets.51F[53] Yet, there is extensive evidence showing that these transactions typically involve the efficient combination of complementary technologies, which are then distributed by the acquiror more rapidly and efficiently than the start-up could have done independently.52F[54] In fact, one of the authors of the empirical study on pharmaceutical acquisitions has recently co-authored an empirical paper finding that acquisitions in digital markets generally appear to promote rather than suppress the diffusion of acquired technologies.53F[55]

Notwithstanding this unsettled empirical grounding, regulators have identified “killer acquisitions” as a significant risk to competition and moved to block several acquisitions by large platforms of emerging companies. Examples include the U.K. regulator’s challenge in 2020 to Facebook’s acquisition of Giphy, a “GIF” digital images site; the FTC’s challenge in 2022 to Meta’s acquisition of Within Unlimited, the developer of a virtual-reality exercise app; and the EU regulator’s challenge in 2022 to Illumina’s acquisition of Grail, a cancer-diagnostics spinoff.54F[56] Under a foreclosure theory, EU regulators in 2023 publicly raised concerns over similarly situated acquisitions, including Amazon’s proposed acquisition of iRobot, the maker of the Roomba vacuum cleaner, and Qualcomm’s proposed acquisition of Autotalks, a mobility chip start-up.55F[57] Both transactions were abandoned. Given the relatively small size of the targets—and, in the Amazon/iRobot transaction, the absence of any direct overlap between the acquiror and target’s products—there is significant doubt that any of these transactions posed a sufficient risk of anticompetitive harm under U.S. antitrust standards to justify intervention.

These cases might be dismissed as “one-off” instances of regulatory overreach that pose little harm in the aggregate. Yet, the prospect of a merger challenge to platform/start-up acquisitions that were once thought to be antitrust “nonissues” can cause a chilling effect that discourages these transactions in general. Since acquisition is the most common monetization strategy for tech start-ups, any antitrust policy that limits the likelihood of an acquisition discourages venture-capital investment—hurting the very start-ups that regulators seek to protect.56F[58] While U.S. regulators appear to have largely ceased targeting platform/start-up acquisitions—even more so now, given that their new HSR (Hart-Scott-Rodino Act) reporting requirements, which were motivated in part by such concerns, have been struck down by a federal court—EU regulators continue to identify killer acquisitions as a focus of merger review policy.57F[59] This is an “own goal” as a matter of innovation policy since it encourages investors and start-ups to invest and locate outside the EU, which is seeking to bolster its lagging start-up ecosystem.

What’s more, some of these regulatory interventions caused significant harm to the start-ups that would otherwise have been acquired. After the EU’s “successful” challenge to Amazon’s acquisition of iRobot, the start-up experienced financial distress and subsequently filed for bankruptcy and now appears to have been acquired by a China-based firm.58F[60] The EU’s “successful” challenge to Illumina’s acquisition of Grail likely delayed deployment of its cancer diagnostics tests (and unnecessarily so, since the EU’s highest court ultimately ruled that the EU never had jurisdiction in the case at all).59F[61] And, after the EU’s challenge to Qualcomm’s acquisition of Autotalks, the start-up experienced financial difficulties and was ultimately acquired by Qualcomm in 2025 for one-fourth of the original acquisition price.60F[62] In other words, regulatory intervention simply delayed deployment of a valuable technology and transferred wealth from a start-up to an incumbent.

The End of the “End of History” in Antitrust

Only about a decade ago, it appeared that antitrust law was converging on its own “end of history”—a technocratic model anchored in the consumer welfare standard and economic methods, which still preserved a significant range of views over the application of that standard in particular circumstances. This consensus has eroded considerably. Regulators, legislators, and courts today face the choice between, on the one hand, the conventional model that seeks to preserve the competitive process and, on the other hand, alternative models that support using competition law for redistributive, structuralist, and other policy objectives. Those alternative models encompass the EU’s quasi-administrative apparatus as reflected in the DMA, certain applications of competition law by EU courts, and new enforcement models advocated in the United States by both progressive and conservative populists.

To be sure, policymakers are understandably concerned about the size of large tech platforms as well as high concentration levels in certain segments of the digital economy more generally, and seek tools that enhance the antitrust agencies’ enforcement powers. The EU model, and the DMA in particular, offers a powerful ex ante toolkit that lowers the evidentiary burden to take action against what are perceived to be anticompetitive practices by entrenched monopolies. In the United States, both the progressive populist approach and the emergent populist conservative approach seek to restore structuralist presumptions that would facilitate litigation against leading tech platforms and incumbents more generally. To various degrees, U.S. antitrust populists embrace those presumptions in an effort to restore or preserve decentralized markets or achieve redistributive outcomes as a matter of principle, even if doing so may have adverse effects on pricing, output, quality, or innovation.

There should be significant hesitation in adopting antitrust models that seek to impact the outcomes of the competitive process to favor certain market structures, rather than simply preserving the free play of competitive forces. Structuralist “big is bad” approaches to antitrust have little grounding in the economics literature, which rejects any simple relationship between concentration levels, vertical integration, and effects on pricing, quality, and innovation. Much of current antitrust commentary seems to overlook that the shift in antitrust law starting in the late 1970s reflected decades of economic scholarship showing there was no consistent relationship between market structure and performance; rather, these relationships vary depending on market-specific factors—making per se rules to identify anticompetitive conduct usually inappropriate.61F[63] The relationship among these variables is especially complex when antitrust policy must make a trade-off between short-term effects on pricing and long-term effects on innovation.62F[64] As Harold Demsetz observed, competition takes place on various dimensions and there is no single rule that can maximize competition on every dimension simultaneously. It is impossible for a central planner to determine the efficient mix of competitive forms at any point in time or to assess whether any such mix has been realized in the relevant market.63F[65]

Given the centrality of scale, scope, and innovation in digital markets, coupled with the inherent uncertainty in implementing antitrust law beyond simple cases of outright collusion, these considerations suggest that the DMA’s per se-like approach, and skepticism toward balancing tests more generally, may endanger efficient transactional and organizational structures in digital markets. The rigidity of the EU policy approach, as reflected in EU competition law (with some important exceptions) and even more so in the DMA, may account in some part for the underperformance of the European tech economy as compared with the outperformance of the U.S. tech economy over the past several decades. The EU’s categorical approach toward competition enforcement may have constrained economic dynamism by inhibiting transactional and organizational innovation. Without making any causal assertions, it is nonetheless suggestive that the U.S. tech economy has outperformed during the period that runs concurrently with the adoption of an efficiency-driven model by U.S. courts and agencies when implementing the antitrust laws, and that there are strong theoretical bases for understanding why the latter development may have contributed to the former.

To be clear, it is critical to scrutinize closely whether leading tech platforms are engaging in exclusionary practices rather than “competition on the merits,” even if a platform acquired its leadership position through technological innovation and business ingenuity. The conventional antitrust model, as reflected in decades of federal case law, has robust analytical capacities to address those concerns through nuanced and evidence-based methodologies that balance the error costs arising from both under-enforcement and overenforcement. It will always be a work in progress that leaves much room for factually and economically informed debate among courts, policymakers, and scholars. However, it remains true that jurisdictions that abandon this analytical framework and revert to “big is bad” intuitions risk making a policy error of significant proportions.

Endnotes