The FTC sets its sights on generative AI

If you control thousands of GPUs, build top-tier frontier models, and package them neatly into your existing operating system or enterprise software, antitrust regulators are already taking notes.

Staff from the Federal Trade Commission's (FTC) Bureau of Competition and Office of Technology recently laid out the antitrust hazards surrounding generative AI on the agency's official technology blog. The message is clear: generative AI could either disrupt entrenched incumbents or cement their dominance, and the FTC is gearing up for aggressive enforcement.

Four ways giants could face regulatory fire

The FTC spotlighted four common business practices that could trigger government intervention.

First up is exclusive dealing. The agency warned that incumbents offering both compute services and AI applications could use exclusive arrangements to lock up GPU supplies, distribution pipelines, or key customer accounts. In antitrust precedent, exclusive arrangements are viewed as particularly risky when a firm holds more than 30% of the market.

Second is tying. If tech giants force buyers of dominant legacy products to take their generative AI add-ons as well, it reduces the market value of standalone competitors and distorts fair play.

Third is bundling. Offering productivity suites, web browsers, or cloud infrastructure alongside generative AI at a deeply discounted price makes it brutally hard for focused startups with a single product to survive.

Finally, regulators are monitoring acquisitions. That does not just mean buying out competing model labs. The FTC is actively scrutinizing deals involving foundational datasets, rare engineering talent, and cutting-edge computing infrastructure, including non-horizontal mergers.

Why it matters

For AI startups and enterprise buyers, this regulatory stance is a major shield. If enforced, it prevents hyperscalers from simply leveraging legacy monopolies to kill innovative standalone tools. Businesses shopping for software will get to pick the best model for the job instead of being forced into whatever their cloud provider bundled.

For tech giants, the compliance calculus just got complicated. Any company bundling AI upgrades into existing enterprise software or signing long-term chip reservations will need to document immediate pro-competitive benefits, like clear cost savings or improved product value, to survive regulatory audits.

We think the FTC is moving much earlier here than it did during the rise of mobile app stores and social platforms, which is smart. The risk, however, is that training state-of-the-art models genuinely demands massive capital and tightly integrated infrastructure. If regulators overcorrect on strategic partnerships, it looks like they could inadvertently slow down the pace of enterprise AI adoption.