Antitrust Law
Competition enforcement to prevent dominant firms from leveraging control over essential AI infrastructure to stifle competition and capture outsized economic value.
What it is:
Antitrust law would seek to prevent excessive market concentration, anti-competitive behavior, and abuse of dominant market positions. If a handful of firms control the infrastructure required to build and deploy AI, they effectively become gatekeepers for the broader economy's access to the most consequential technology of the era. This has direct implications for how AI's economic benefits are distributed: a concentrated AI industry can extract rents from every sector that depends on its infrastructure, set the terms on which smaller firms and startups access compute and models, and shape the direction of AI development toward applications that maximize the incumbents' revenue rather than broader social value.
AI-specific antitrust enforcement would focus on preventing control over the physical and digital infrastructure required to build frontier models by powerful established market players. Beyond traditional price-fixing concerns, regulators focus on vertical integration where cloud giants own the chips, data centers, and application layers, potentially locking out competitors. Other key issues include "acqui-hires" (hiring a startup's core team to avoid merger scrutiny), exclusive cloud partnerships that restrict model availability, and technical lock-ins like high data egress fees. The goal is to ensure that the AI stack remains modular and open, preventing a scenario where a few incumbents act as gatekeepers for the entire economy's infrastructure.
The challenge:
The challenge for antitrust enforcement is that AI markets move faster than regulatory processes. By the time an investigation concludes or a merger is reviewed, market dynamics may have already shifted; a dominant position may be entrenched or a nascent competitor may have been absorbed. Traditional antitrust tools, designed around clearly defined product markets, struggle with technology markets where the 'product' spans multiple layers and many services are offered at zero price, making it difficult to demonstrate consumer harm through the higher-prices framework that underpins most antitrust enforcement. Training frontier AI models requires enormous capital investment in compute, data, and talent, meaning that aggressive antitrust enforcement that fragments the industry could slow the development of the most capable systems.
Recommended Reading:
Real-world precedents:
The DOJ vs. Google ruling established that exclusive distribution deals violate antitrust law, specifically finding that Google maintained an illegal search monopoly by paying billions to companies like Apple and Samsung to make its search engine the default on their devices.
In the 1990s, United States v. Microsoft forced the unbundling of Microsoft’s Internet Explorer browser from the Windows operating system, as the court found Microsoft had deliberately sought to protect its operating system monopoly from the threat posed by cross-platform middleware like Netscape.