The European Union’s draft Merger Guidelines give a central role to dynamic competition in merger review. Some scholars have criticized dynamic competition as an analytical tool that seems to always discourage government intervention, given how quickly and unexpectedly—or dynamically—innovation can remake a market. Nicolas Petit, Selcukhan Unekbas, Bowman Heiden, and Pierre Regibeau argue this critique ignores the several large cases in which regulators used dynamic competition to intervene in a merger.
This article is part of a symposium on the European Commission’s draft Guidelines on the assessment of mergers under Council Regulation. The new merger guidelines mark their first systemic update since their first release in two parts in 2004 and 2008. You can read the rest of the contributions to the symposium as we publish them here. We encourage responses to our symposium, which can be submitted to promarket@chicagobooth.edu.
The European Union’s draft Merger Guidelines move dynamic competition to the center of merger analysis. Competition to change the basis of competition itself, often through technological change—which is what we mean by “dynamic competition”—infuses the assessment of market power, theories of harm, entry, and efficiencies, among others. This is a welcome development. Antitrust policy has always evolved with the economic reality it addresses. And the highly technological nature of competition today requires us to think about competition in more dynamic terms.
There is, however, a prominent criticism leveled against bringing dynamic competition into merger policy. Despite numerous merger challenges based on dynamic competition (like Dow/Dupont or Facebook/GIPHY), many maintain that dynamic competition is almost always marshalled to support less intervention. The rapid evolution of technology supposedly means government intervention always risks hampering innovation, and so dynamic competition becomes an excuse to discredit a theory of harm, or to clear a merger because of purported benefits to innovation. It never supports a merger challenge.
This view is both factually incorrect and is based on a misunderstanding of the nature of dynamic competition. Dynamic competition is not prescriptive. It does not call for more or less merger enforcement. It is descriptive and aims to help competition agencies make better enforcement decisions. Thus, dynamic competition can—and as we show, does—support intervention when the facts and the setting call for it. Indeed, to our knowledge, there are more well-documented cases where dynamic competition arguments were used to scupper a proposed transaction than cases where transactions that would otherwise have been blocked were allowed because of purported dynamic benefits.
This article demonstrates this by presenting four cases in which competition agencies in the United States, United Kingdom, and EU intervened on dynamic competition grounds. We did not pick these cases randomly. They were consistently mentioned at discussions held within the Dynamic Competition Initiative (DCI), an academic research venture with which several of the authors are affiliated. These cases should give confidence to the European Commission (EC)—and other agencies—that dynamic competition is neither a pro- nor anti-enforcement program. Rather, it is a dimension of competition worthy of consideration for any decision maker committed to the protection and promotion of innovation and competition in future products and technologies not yet fully developed.
Dynamic competition cases in the US
In the last fifteen years, the U.S. antitrust agencies successfully challenged two mergers on dynamic competition grounds. We review them in turn.
Applied Materials/Tokyo Electron: lessening dynamic competition through a loss in innovation potential
In 2013, semiconductor manufacturers Applied Materials (AMAT) and Tokyo Electron (TEL) proposed to enter into a “merger of equals.” The rationale for the transaction was to “accelerate development of breakthrough products” and “address future technology inflections,” particularly in the emerging market for mobile device chips. Until then, the U.S. agencies had mostly cleared mergers in the same industry, including in ASML/Cymer and Advantest/Verigy. In AMAT/TEL, however, the U.S. Department of Justice (DOJ) challenged the transaction. The parties abandoned their merger in 2015.
The theory of harm that supported the DOJ’s challenge is interesting. To find a substantial lessening of competition, the agency did not rely on a standard risk of anticompetitive price increase or on a finding of overlaps in several markets. The parties were not close competitors in existing products, consistent with broader trends of extreme specialization in semiconductor manufacturing equipment: AMAT led in deposition (layering thin materials onto silicon wafers) and TEL in coating and etching. From a static competition viewpoint, the deal was unproblematic.
Instead, the agency challenged the merger because “the combination would have diminished innovation” in next-generation semiconductor manufacturing equipment.
The agency did not predict which future products or services would not be supplied following the merger. Instead, the DOJ reasoned generally on firm-level capabilities to suggest that the merger would slow or reduce the supply of future products and services. The parties were, in the agency’s words, “the two largest competitors with the necessary know-how, resources, and ability to develop and supply high-volume non-lithography semiconductor manufacturing equipment.” They possessed “highly developed experience in material sciences and a broad array of existing chambers,” “engineering experience in designing and constructing mainframes,” “R&D facilities,” and “financial resources to take on more risky projects and persevere through setbacks.”
These capabilities were a source of competitive pressure, and the merger would have removed them. The parties’ capabilities were not tied to a single product but were deployable across a portfolio of related ones. This portability implied a reasonable likelihood of diversification into adjacent markets (a key insight of the resource-based theory of the firm on which the capability literature is built). The agency logically inferred that a reduction of innovation would imply a lessening of latent future competition. The case is a straight application of the concept of “innovation potential” as determined by the parties’ capabilities.
Today, the parties remain independent and continue to compete in the market for advanced semiconductor equipment. AMAT remains the largest semiconductor equipment maker by revenue, while TEL maintains its position as a close, top-four rival. Both firms continue to invest heavily in research and development, driven by the artificial intelligence revolution.
Visa/Plaid: Lessening dynamic competition through a loss in innovation competition and potential
In 2020, Visa announced plans to acquire Plaid, a fintech startup, for $5.3 billion. The DOJ challenged the deal, and the parties abandoned their transaction in 2021.
Like AMAT/TEL, the Visa/Plaid challenge illustrates a concern to safeguard dynamic competition, even if the core complaint by the DOJ focused on Visa’s dominant 70% market share in online debit processing. But the case is even more remarkable in that it tied a theory of harm to “innovation competition” and “innovation potential,” employing a concern from the 2010 Merger Guidelines that “a merger will diminish innovation competition by combining two of a very small number of firms with the strongest capabilities to successfully innovate in a specific direction.”
Visa is and was the leading processor of online debit payments, while Plaid is and was a financial data aggregator whose application programming interfaces (APIs) enabled consumers to share banking information with fintech applications such as Robinhood and Venmo. Plaid intended to leverage its database of around 200 million consumer bank accounts and its connections with 11,000 financial institutions, together with its technological capabilities, to launch a competing online debit-processing service (pay-by-bank). This would have allowed consumers to pay merchants directly through their banks. Plaid’s “pay-by-bank” service would have bypassed Visa and threatened its $2 billion in annual online debit revenue. The record showed that concerns of a lessening of innovation competition were real. In internal documents, Visa had described the deal as an “insurance policy” to protect its U.S. debit business. Other documents revealed plans to discontinue Plaid’s pay-by-bank development following the transaction.
In addition to concerns of loss of innovation competition, the Visa/Plaid challenge also features concerns of a lessening of dynamic competition through innovation potential. The agency pointed to evidence that the merger would eliminate Plaid’s “innovative potential,” powered by its “unique” access to a large network of fintech applications, consumers, and financial institutions. The DOJ’s concerns about innovation potential were backed by observations that Plaid had access to unique resources and technological capabilities. Perhaps it was these resources and capabilities that enabled Plaid to raise $13.4 billion in its IPO following the deal’s abandonment.
Dynamic competition in the EU and UK
The EU and U.K. agencies have also successfully challenged mergers on dynamic competition grounds. Two cases stand out: the EC’s review of Nvidia/Arm and theU.K. Competition and Markets Authority’s (CMA) review of Adobe/Figma.
Nvidia/Arm: maintaining dynamic competition through access to innovation inputs
In September 2020, chip designer Nvidia sought to acquire Arm, a licensor of semiconductor design technology, for $40 billion. The transaction was notified to several competition authorities in 2021, including the EC, and was abandoned in 2022 after multiple investigations.
The Nvidia/Arm case demonstrates that conventional theories of anticompetitive foreclosure go a long way towards protecting dynamic competition. The main theory of the case was that Arm’s intellectual property was an “important input in products competing with those of Nvidia,” including in “datacenters, automotive, and Internet of Things.” By changing Arm’s incentives to supply inputs to Nvidia’s rivals—a dynamic effect of the merger—the transaction would limit effective competition. The Commission would follow in 2022 a similar line of argument in Illumina/Grail, stating that biotech company Illumina’s acquisition of Grail, which claimed to have developed a path-breaking blood test for a large variety of cancer types, would affect Illumina’s incentive to provide DNA-deciphering technologies, which are key to the R&D efforts of Grail’s potential rivals.
Alongside this conventional yet dynamic theory of harm, the EC also used standard economic reasoning to identify a risk of a substantial lessening of innovation. The EC considered that the merger between Nvidia and Arm would give rise to a hold-up problem. Arm’s business model consists of licensing its IP in chip design to traditional semiconductor firms like Broadcom, Qualcomm, and Nvidia, and to other technology firms that have vertically integrated into chip supply, like Google and Amazon. Arm and its licensees communicate extensively in order to develop tailor-made chips. The EC identified a risk that semiconductor firms that compete with Nvidia would be reluctant to collaborate with an Nvidia-owned Arm because of information spillover risks.
This was not a hypothetical concern, as Intel had a similar experience. To compete with Taiwan Semiconductor Manufacturing Company in chip manufacturing, in 2022 Intel tried to separate its manufacturing division from its design operations but has so far failed to attract customers. This is because Intel competes in chip design with its potential customers (e.g., Qualcomm), and the latter are reluctant to give their cutting-edge designs to Intel for manufacturing. A similar scenario could have played out for Arm, as the latter is a much more important partner for chip designers than Intel is for chip manufacturers. Hence, the EC’s worry in Nvidia/Arm was, from a preliminary point of view, plausible.
In hindsight, the EC got dynamic competition right in Nvidia /Arm. While the EC case focused on the risk of foreclosure of central processing units (CPUs) for datacenters, the decision protected dynamic competition by supporting innovation in chips for artificial intelligence and machine learning. By keeping Arm independent from chip suppliers like Nvidia, the decision maintained access to technological capabilities essential for the emergence of new types of AI-specialized chips—like Google’s tensor-processing units and Amazon’s Trainium—which are increasingly competitive with Nvidia’s GPUs.
Adobe/Figma: lessening dynamic competition through a loss in innovation investment or effort
The CMA’s 2022 challenge of software company Adobe’s proposed acquisition of digital design platform Figma is another illustration of dynamic competition and merger enforcement working hand in hand. The CMA’s concerns of harm to competition stemmed from a risk that the parties would abandon pre-merger investments or efforts in new product development.
In product design software, the CMA found that the merger would give rise to a loss of innovation investment. Figma held an 80% market share. Adobe had spent two years developing its own web-based tool, Project Spice, to compete more effectively with Figma. With only a 10% market share, Adobe had put the project in “maintenance mode” before the merger. The CMA relied on this fact to support its theory that the merger would harm to dynamic competition.
A similar risk of post-merger reductions in innovation effort also undergirded the CMA’s analysis of vector- and raster-editing software. Adobe’s Illustrator and Photoshop products were market-leading, while Figma’s products delivered limited competitive pressure. Yet, the CMA concluded that Figma was a “particularly credible dynamic competitor” to Adobe. This inference rested on a review of Figma’s resources and capabilities to overcome entry and expansion barriers. The CMA considered that Figma held the required technical capabilities and access to engineering talent. Figma also displayed a credible track record in terms of managing development costs and of time-to-market for new products. The CMA thus provisionally found that the merger would eliminate a dynamic competitive threat.
In retrospect, the CMA’s provisional findings were vindicated. In 2025, Figma launched Figma Draw, a vector-editing software that competes with Adobe’s Illustrator. Figma also went public in July 2025. To this day, both Adobe and Figma remain competitors in the design space, with Figma continuing to lead in collaborative interface design.
Conclusion: varieties of harm to dynamic competition
The four cases discussed above convey a clear message: dynamic competition is not biased against enforcement. On the contrary, a concern for the promotion and protection of dynamic competition inhabits the decisional practice of enforcers across the world.
The cases also suggest that there are varieties of harm to dynamic competition. Mergers can limit dynamic competition through a limitation of innovation competition (Visa/Plaid; Adobe/Figma), innovation potential (AMAT/TEL), innovation investment and effort (Adobe/Figma), or innovation inputs (ARM/Nvidia).
With its draft Merger Guidelines, the EC has taken on the challenge to bring dynamic competition into merger analysis. This is welcome. Frameworks to support the Guidelines must now be developed to specify the varieties of mechanisms that threaten dynamic competition. Combined with an equivalent effort to understand mergers’ benefits on innovation, we believe the new Guidelines will usher in a more analytically precise and socially beneficial approach to competition policy and enforcement.
Authors’ Disclosures: Pierre Regibeau is an affiliate of Analysis Group. Over the last eight years, he has not worked for any company active in the tech or biotech sectors. During his tenure as Chief Competition Economist at DG Competition from 2019-2023, he was involved in a number of merger cases involving dynamic theories of harm: Nvidia-Arm, Illumina-Grail, Microsoft-Activision, Google-Fitbit and Apple-Kustomer. The rest of the authors report no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
