Why Antitrust Enforcement in Tech Markets Is Harder Than It Looks

Antitrust law was built to address a specific set of problems in markets for physical goods and services—markets with tangible products, predictable cost structures, and clear price signals. The dominant tech platforms of the last two decades have created a new set of market structures—search, social media, operating systems, app stores, cloud infrastructure—that interact poorly with the tools and concepts that antitrust law developed over the previous century. The mismatch is real, and it explains why regulators in the US, EU, and elsewhere have struggled to translate concerns about platform power into enforceable and durable outcomes.

The Price-Based Harm Problem

Traditional antitrust analysis starts with price: does a firm’s market power allow it to charge consumers above-competitive prices? The consumer welfare standard—developed by the Chicago School of economics and dominant in US antitrust enforcement since the 1970s—focuses on whether consumers are harmed through higher prices or reduced output. If a market is characterised by low or zero prices to end users, the consumer welfare standard provides limited traction.

Google’s search engine, Facebook’s social network, and Snapchat are free to users. If the complaint is that these platforms have dominant market positions, the standard question—”are consumers paying too much?”—doesn’t apply. The answer is that users are paying with data and attention rather than money, but price-based harm analysis doesn’t map naturally onto these currencies. Antitrust regulators spent years grappling with how to adapt a framework built around price to platforms whose business models depend on free-to-user services funded by advertising.

The response from economists and regulators has been to reconceptualise harm: in two-sided markets, where a platform serves both users (who pay nothing) and advertisers (who pay for access to those users), the relevant market isn’t just the user-facing side. Harm to competition in advertising markets, harm through reduced privacy or quality of service, and harm to potential competitors through exclusionary practices can all be incorporated into analysis—but doing so requires adapting frameworks that weren’t designed for this structure, and reasonable economists can reach different conclusions about whether harm exists and how to measure it.

Network Effects and the Difficulty of Entry

Many tech markets exhibit strong network effects: a platform is more valuable to each user as more users join. Social networks are the clearest example—a social network with few users is useless, while one with all your friends and family is nearly indispensable. Search engines benefit from scale because more queries generate more data to improve results. Payment networks, operating systems, and app stores similarly benefit from critical mass on both sides.

Network effects create winner-take-most dynamics: once a platform reaches sufficient scale, it becomes difficult for competitors to displace it even if they offer a technically superior product, because users are reluctant to switch to a platform where their contacts aren’t present. This is sometimes called the “cold start problem”—a new entrant must somehow attract enough users to create value before those users have reason to join.

For antitrust purposes, the question is whether these network effects constitute a natural feature of a well-functioning market (the dominant firm is just the one that won fair competition and should be protected) or a barrier that locks in incumbents regardless of quality or innovation. The Chicago School view tends toward the former; more recent industrial organisation research suggests network effects can create durable lock-in that isn’t self-correcting through market forces, providing at least a theoretical justification for regulatory intervention.

Defining the Relevant Market

Antitrust analysis requires defining the relevant market—the set of products and services that are substitutes for the product under scrutiny. If you define the market narrowly, a firm can appear dominant; if you define it broadly, the same firm looks like a minor player. This definitional challenge is particularly acute in tech.

In the US Department of Justice’s antitrust case against Google, one question was how to define the market for search and search advertising. Google argued it competes with Amazon (for product searches), Yelp (for local business searches), YouTube (for video searches), and eventually AI assistants. The DOJ argued for a narrower definition of “general search” in which Google holds over 90% market share. Both framings are defensible, and the outcome of the case depended substantially on which definition the judge accepted.

Platform markets often span multiple product categories simultaneously, complicating market definition further. Apple’s App Store is a distribution market for software, but it’s also an advertising platform, a payment processor, and an operating system feature—depending on which lens you apply, the competitive landscape looks very different. Amazon is simultaneously a retailer, a marketplace, a logistics provider, and a cloud computing company, and its market position in each depends on how the boundaries between these services are drawn.

The Acquisitions Question

Critics of regulatory enforcement argue that the most consequential failure was not in challenging existing dominance but in allowing acquisitions that eliminated potential competitors before they became threats. Facebook’s acquisitions of Instagram (2012) and WhatsApp (2014), each approved by regulators at the time, are the most frequently cited examples. When approved, both acquisitions appeared to be between companies in somewhat different spaces—Instagram was a photo sharing app, WhatsApp a messaging service—that didn’t obviously compete directly with Facebook’s social network. In retrospect, both acquisitions eliminated nascent competitors that could have challenged Facebook’s dominance.

The problem is that merger review at the time of acquisition must assess future harm under uncertainty. Instagram had 30 million users and no revenue when Facebook acquired it for $1 billion—a price that seemed high at the time. Predicting that it would become a direct competitive threat to Facebook’s core social networking product required foresight that regulators didn’t have (or wasn’t weighted sufficiently in their analysis). The “kill zone” hypothesis—that potential competitors in sectors adjacent to platform dominants are systematically deterred from investing or acquired before they can threaten incumbents—is consistent with the empirical pattern even if causation is hard to prove case by case.

Remedies and Their Limits

Even winning an antitrust case doesn’t automatically resolve competition problems if the remedy is poorly designed. The standard remedies—structural (breaking up the company) or behavioral (prohibiting specific practices)—each have limitations in tech markets. Structural remedies require courts to engineer corporate separations that preserve value while creating independent competitors—a complex operation that courts are generally reluctant to order and that can take years to implement, by which time the market may have evolved substantially.

Behavioral remedies (consent decrees requiring the firm to change specific practices) have a mixed track record. They can be effective for well-defined, durable problems but are less effective when the specific problematic practices are easy to route around or when markets evolve quickly. Microsoft’s antitrust case resulted in behavioral remedies that allowed Windows to continue; by the time remedies were in effect, the browser wars had substantially resolved, and the more consequential competitive dynamics had moved elsewhere. The lesson many drew was that the remedy failed to address Microsoft’s underlying market power—though others argued the case was largely mooted by organic market developments.

The European Approach and Its Results

The European Union has taken more aggressive action against US tech platforms than the United States has, using competition law to impose fines and behavioral requirements on Google, Apple, Meta, and Amazon. The Digital Markets Act (DMA), which took effect in 2024, designated several major platforms as “gatekeepers” and imposed a set of requirements (interoperability, data portability, prohibition on self-preferencing) without needing to prove market harm case by case—a structural shift from reactive enforcement to proactive rules.

The early evidence on the DMA’s effects is mixed. Some requirements (interoperability, data portability) may genuinely reduce switching costs and benefit users. Others have faced implementation disputes and legal challenges. The fundamental challenge remains: writing technology-neutral rules that address current market power without inadvertently protecting incumbents or creating unexpected consequences requires predicting how markets will evolve, which even well-resourced regulators struggle to do accurately.

Antitrust enforcement in tech isn’t failing because regulators are captured or incompetent—it’s failing, where it is failing, because the tools were designed for different market structures and are being adapted in real time to a set of problems the original designers didn’t anticipate. That adaptation is ongoing, and the outcomes of current litigation and regulation will shape both market structure and the precedents available for future cases.