This blog is the latest coverage in a series on Artificial Intelligence and the Google cases. You can read more of our coverage here, here, here, here, here, here, here, here, here, here, and here.
Imagine: a company brings a product with mass appeal to market, wowing consumers, and earning monopoly power. A generation passes, and another round of innovation emerges. The dominant firm sees where the market is headed, and uses its monopoly power in the first market to squash upstart rivals in the second. This is an obvious competition problem. An order is issued: to curtail its monopoly power, the dominant firm is required to let rivals interoperate with its monopoly product. In theory, this would allow competition to flourish without the severity of a breakup. But complaints begin almost as soon as the ink is dry. Because the dominant firm has every economic incentive to exclude its rivals, technical problems, logistical roadblocks, and petty disputes pile up fast.
Is this Google, 2026? No, it’s AT&T, 1968. (And Live Nation, 2010).
This week, the Court in the Google AdTech case rejected divestiture of AdX and instead appears inclined to require interoperability. (No official remedy is public yet.) This follows a similar approach to the Search decision, and other Big Tech monopolization cases, favoring behavioral remedies over structural changes. Will on-paper interoperability be enough to stop Big Tech’s anticompetitive conduct in practice? Given the hard shift into artificial intelligence—another emerging wave of innovation built on top of monopoly products—the stakes could not be higher.
AT&T and the Carterfone Trap
In 1968, the FCC issued its famous Carterfone decision, requiring AT&T to allow competing ancillary devices (like phones and answering machines) to interoperate with the copper wire network. But instead of solving the problem, it delivered twenty years of enforcer distraction and litigation, with precious little interoperability in practice. That frustration led directly to the AT&T breakup in 1984. The two decades that followed 1984 could not have looked more different than the previous: a flourishing of innovation built on top of, and genuinely interoperating with, the telephone network—computers, modems, and the Internet. The Third Industrial Revolution.
The trouble with Carterfone was never the theory. Requiring AT&T to let non-harmful equipment attach to its network was, on its face, exactly the kind of narrowly tailored fix that lets the market do the rest of the work. The trouble was execution. AT&T dragged out technical certification standards. It insisted that competitors’ equipment connect only through “protective connecting arrangements,” which only AT&T could sell, install, and price. When MCI tried to offer competing long-distance service under the same interconnection logic, AT&T fought that too, through sprawling federal antitrust litigation. The FCC, for its part, spent the better part of two decades running the Computer Inquiries, an entire regulatory apparatus built around the single recurring question of how the phone company was, this year, using its network to disadvantage the market next door. Congress, the FCC, and the courts were always one step behind a company with every incentive to stay ahead of the competition.
Only divestiture ended the cycle. The 1984 dissolution famously carved the Bell System into regional operating companies. It also spun off Western Electric, AT&T’s manufacturing arm and the maker of the very ancillary devices that competitors had spent decades trying to challenge. Thus, splitting off the regional networks curtailed AT&T’s power to dictate terms to the equipment market. And spinning off Western Electric eliminated AT&T’s incentive to squash competing devices in the first place.
The 80s and 90s saw a flurry of answering machines, fax machines, computers, and modems—all of which existed in 1968, but were prevented from coming to market. These of course became the foundation for the Internet and digital markets.
Encore: Live Nation Repeats the Carterfone Saga
If AT&T feels too vintage, Live Nation and Ticketmaster tell a similar story. In 2010, DOJ approved the Live Nation–Ticketmaster merger subject to a behavioral consent decree: a firewall between Ticketmaster’s ticketing operation and Live Nation’s concert promotion business, and a prohibition on retaliating against venues that used a rival ticketing platform. In other words, a behavioral remedy eerily similar to Carterfone’s interoperability mandate.
Ask Taylor Swift fans how that played out.
By May 2024, the DOJ and thirty states sued to break up Live Nation outright, alleging the company continued to leverage its dominant position over venues and artists to entrench exactly the anticompetitive conduct the 2010 decree was designed to prevent. A decade and a half of compliance fights and half-hearted interoperability, in the government’s own telling, simply bought Live Nation a longer runway to squash competition.
AT&T and Live Nation were different “Eras” and distinct products. But the track record is the same: on-paper interoperability did not solve the competition problem in practice.
Google’s Turn
Which brings us back to Google. In the Ad Tech case, the Court found that Google illegally monopolized markets, including by tying its publisher ad server (where publishers list advertising space for sale) to AdX, its exchange that facilitates the sale. Tying claims seem particularly well suited to a divestiture remedy: as with AT&T, sell off one of the two tied products, and the incentive to squash competition in the other disappears.
The Court did not order that relief. Instead, it appears inclined toward behavioral remedies. government proposed. (The public version of the full remedies has not been released yet). For example, the government earlier proposed interoperability and data sharing requirements, which the Court may ultimately order. This sounds reasonable enough in theory. If Google has required AdX users to source publisher inventory only through Google’s own server (and vice versa), an order allowing rivals to interoperate (thus eliminating Google’s requirement) should, on paper, address the antitrust problem.
The catch, as with Carterfone and Live Nation/Ticketmaster, is that Google has little incentive to make that interoperability frictionless. The risk is not that Google will violate the letter of the order, but rather that Google gets very good, very quickly, at finding the tactics the order doesn’t specifically mention.
That risk lands at an inconvenient moment, since Google is building out the cloud infrastructure, foundation models, and distribution deals that will shape artificial intelligence markets. As one example, a live question every AI company is wrestling with right now is how to monetize AI. Google, like its rivals, is already testing how to introduce advertising into AI-generated results. Superficially, that question may only be loosely connected to the facts of the AdTech case. But Google’s incentives as it makes AI-related choices (including how to monetize AI products through advertising) look very different than they would if either the Ad Tech or Search courts had ordered structural relief. How those incentives produce the next round of antitrust headaches is a question antitrust enforcers will be watching closely.
