U.S. District Court Judge James Boasberg’s decision dismissing the Federal Trade Commission’s (FTC) lawsuit against Meta, which accused the company of maintaining a monopoly in the market for “personal social networking,” may have closed the door on one era of activist antitrust enforcement against so-called “Big Tech.”
If enforcers do not learn from the mistakes that doomed the FTC’s case, however, we could soon be in for another protracted, yet ultimately fruitless battle targeting the still-emerging market for artificial-intelligence (AI) services.
Boasberg’s opinion begins with Heraclitus and the idea that no man can ever step into the same river twice. The social-media market has changed, even since the FTC first filed suit in 2020 seeking to force Meta to divest Instagram and WhatsApp. When Meta acquired Instagram in 2012, TikTok did not exist. By 2021, TikTok had become one of Instagram’s primary competitors.
Ultimately, the agency’s theories couldn’t survive contact with evidence of how consumers actually behave and how markets evolve. Whatever one thinks about whether the government should have blocked Meta’s acquisition of Instagram in 2012 or WhatsApp in 2014, the question before the court in this case was whether Meta possesses monopoly power today. On that question, the FTC produced no evidence that could overcome what consumers actually do.
The FTC tried to define a narrow market called “personal social networking,” which basically comprises platforms for connecting with friends and family. Conveniently for the FTC, that excluded TikTok and YouTube, because those apps supposedly serve entertainment purposes. This distinction would have allowed the agency to claim Meta held a dominant share of a market containing only itself, Snapchat, and the obscure MeWe platform.
Boasberg flatly rejected this approach. The evidence showed consumers readily switch among Facebook, Instagram, TikTok, and YouTube. When economists paid participants to reduce their Facebook usage, the app they turned to most was YouTube. When Meta’s apps went down for several hours in October 2021, users flocked to TikTok more than any other platform. That far exceeded the migration to Snapchat, supposedly Meta’s closest competitor under the FTC’s theory.
Market definition may sound like technical lawyering, but it determines quite a bit in an antitrust case. A firm cannot be a monopolist if the market includes strong competitors. By requiring evidence of how consumers actually behave, rather than accepting theories about what different apps are "for,” Boasberg followed sound economics.
The FTC’s static approach to market definition matters beyond social media. The next wave of antitrust enforcement, nearly all observers would agree, will target AI. Regulators in the United States and Europe are already investigating competition concerns in AI compute access, model licensing, and cloud infrastructure.
Social media changed faster than the agency’s theories could accommodate. AI markets are evolving faster still.
OpenAI released ChatGPT in November 2022. Within two years, Google, Anthropic, Meta, Microsoft, and dozens of startups had deployed competing models. China’s DeepSeek recently matched frontier performance at a fraction of the training cost. Just as in the Meta case, an antitrust framework that defines AI markets by what products are “for,” rather than what product substitutions consumers actually make, would freeze this competition in place.
Cases targeting Microsoft’s investment in OpenAI or cloud providers’ arrangements with AI companies would fail if built on the same approach that failed in Meta. The FTC’s case treated Meta’s 2012 acquisitions as if markets were frozen at the moment of purchase. If applied to AI, this would treat chatbots as the permanent form in an industry where consumer interaction itself keeps changing.
Judge Boasberg demanded evidence of how consumers actually behave. Future enforcers should heed him.
