Antitrust and AI: Key Takeaways From My Congressional Testimony
Earlier this month, I was honored to testify before the House Judiciary Subcommittee on the Administrative State, Regulatory Reform, and Antitrust in its hearing “Artificial Intelligence: Examining Trends in Innovation and Competition.” In my testimony, I explained that the AI industry is dynamically competitive and growing rapidly, which suggests heavy-handed antitrust enforcement is premature. Moreover, against the oft-heard refrain of digital “market failure,” AI represents the next gale of creative destruction that, like the Internet, will likely prove a boon for both consumers and the American economy more broadly. Rather than engage in knee-jerk antitrust enforcement—or, on the contrary, abandon the antitrust enterprise entirely—the current antitrust legal framework is sufficiently well-founded and adaptable to ride the AI wave.
Competition at the key levels of the AI stack—foundation models, cloud computing, and chips—is healthy and consistent with the Schumpeterian dynamics that typify high-tech industries. Entrants like OpenAI and Anthropic are leaders at the model level and face robust competition not just from other startups like Mistral AI and Cohere, but large digital incumbents like X, Google, and Meta. Indeed, the cloud industry itself is a testament to Big Tech firms, rather than resting on their laurels, innovating and competing, and continues to see entry with firms like Coreweave attempting to capitalize on AI. And, at the chip level, Nvidia represents a textbook case of a large tech incumbent reinventing itself with a revolutionary new product that leapfrogs the big winners in earlier technological waves (i.e., mobile and PC), and its success has spurred entrants both domestically and globally, not least from China.
As such, although the antitrust enforcers in the Biden administration liked to analogize the current moment in AI with the beginning of the digital revolution three decades ago, the circumstances could not be more different. Instead of a “Wintel-duopoly” 2.0, the AI space features a range of firms (some big, others much smaller) competing across the various levels of the stack. This reflects the cumulative nature of the Schumpeterian competition that has long defined the high-tech space: With each new market that emerges (e.g., search, e-commerce, social media), the number of tech players with the incentives and scale to compete for the next wave of innovation grows. It’s a virtuous cycle. Moreover, and here again in contrast to the Wintel era and broader “unipolar moment,” the U.S. is no longer the only player in the game. Several countries are intensely focused on developing their AI industries, especially China, which seeks to wrest global techno-economic leadership from the U.S. by winning in AI.
Concerns about rampant anticompetitive conduct stifling innovation also don’t translate well from the beginning of the digital revolution to today. The DOJ’s landmark case against Microsoft was premised upon the company using exclusionary conduct to suffocate the next big wave of creative destruction, the “Internet Tidal Wave.” Yet, none of the current Big Tech antitrust cases are premised on a Big Tech firm using its alleged monopoly power to stifle the AI revolution. Indeed, DOJ strenuously argued that AI was not a constraint on Google’s purported search monopoly, despite what is increasingly clear evidence to the contrary and which Google will likely point out in the upcoming remedies trial. Meanwhile, at the FTC, a trial will soon begin against Meta for acquisitions that occurred well over a decade ago—hardly the sort of Microsoft-esqueforward-looking enforcement to protect the next generation of innovation.
In fact, instead of trying to strangle AI in the cradle, large digital incumbents are doing the exact opposite—actively investing in its growth. Frankly, the figures are staggering: In January, Microsoft, along with Nvidia, Arm, Oracle, and OpenAI, announced a $500 billion investment over the next four years in the U.S. to build new AI infrastructure. A week later, Meta also announced that it would be putting “hundreds of billions of dollars” behind AI in the long term, with $60 billion in 2025 alone. Not to be outdone, Google subsequently revealed that it allotted $75 billion in 2025 for AI and cloud infrastructure, with Amazon similarly committing $100 billion in capital expenditures for the year, the “vast majority” of which will go to AI. Shortly thereafter, Apple let slip that it would be spending more than $500 billion in the U.S. over four years to support a wide range of initiatives, with a big focus on AI.
And yet, rather than see these actions for what they are—proof of a thriving AI industry—in its final days, the Biden administration issued a report warning that AI investments by Big Tech, and specifically partnerships like those between Microsoft and OpenAI, “may cross the line into anti-competitive conduct.” Indeed, at the hearing, Biden-appointed FTC Commissioner Bedoya testified that these partnerships may incentivize large digital cloud providers “to limit access to their platforms away from AI startups who they’re not partnered with.” Troublingly, the Trump FTC appears to be doubling down on the Biden FTC’s broad AI-centered investigation into Microsoft, both with respect to the sort of vertical concerns mentioned by Bedoya, as well as a potential horizontal reduction in competition by virtue of Microsoft’s stake in OpenAI.
The reality is that these partnerships, which typically involve a large cloud service provider getting equity and revenue-sharing, a commitment to use its cloud services, and certain resource-sharing and potential joint development opportunities, are unlikely to run afoul of the antitrust laws. First, not only does the robust competition at the cloud level indicate that substantial foreclosure of cloud resources is unlikely, but the model level is particularly dynamic and competitive, which strongly suggests that such partnerships are unlikely to substantially lessen competition. Second, these partnerships have considerable procompetitive benefits, such as providing AI startups with long-term patient capital and facilitating other relationship-specific commitments that allow them to grow more quickly and sustainably.
In sum, the AI industry’s structure, conduct, and performance at this time present no justification for hasty antitrust enforcement, including against the very investments and partnerships that are helping to drive the AI revolution forward. To be sure, AI is already being heralded as the next great general-purpose technology that will broadly affect our economy and society, including the practice of antitrust law. But it is too soon to blithely assume that, unlike the Internet, AI will lead to undesirable effects throughout the economy radically, such as increased market concentration, more anticompetitive conduct, or persistent market failures. Rather, the antitrust laws must be ready to police anticompetitive behavior just as they have amidst past industrial changes, allowing the AI revolution in America to flourish and keep America the world’s techno-economic leader in the 21st century.
