For the first time, a United States court has called Google an illegal monopolist of the machinery that runs open-web advertising. Not the search box, not Android, but the ad exchange and the ad server that decide how a banner gets bought and sold on a news site or a recipe blog. That ruling landed in April 2025. The harder question, the one the industry is still waiting on, is what a court should do about it. One of the options on the table is to make Google sell AdX, the exchange that sits at the center of the open-web supply chain.
If you buy or sell digital advertising, this is not a spectator sport. The outcome could change which company runs the auction your impressions pass through, what it costs, and who can see inside it. Here is what has actually been decided, what has not, and what a forced sale would really mean.
Origin: how Google ended up in court
The case is United States v. Google LLC, filed on 24 January 2023 in the Eastern District of Virginia. The Department of Justice brought it alongside a group of state attorneys general, starting with eight states and growing as more joined (The Platform Law Blog). The complaint is not about Google's search ads or YouTube. It targets the plumbing of display advertising on the open web, the sites that are not Google, Meta, or Amazon properties.
To follow the argument you need three pieces of that plumbing. A publisher ad server is the software a website uses to manage its ad slots and decide what fills them. An ad exchange is the live auction where that inventory is sold to bidders. An advertiser ad network is the tool buyers use to push budget into those auctions. Google owns a market-leading product in each: DoubleClick for Publishers, now folded into Google Ad Manager, on the publisher side; AdX as the exchange; and Google Ads, formerly AdWords, as the buyer network. If the stack itself is unfamiliar, our ad-tech stack explainer lays out who sits where.
Much of this came from acquisitions. Google bought DoubleClick in 2008 and the supply-side firm AdMeld in 2011 (Mintz). The government's case was that Google then used its position on all three sides to wire the market in its own favor: tying its exchange to its ad server, steering auctions, and setting pricing rules that made it hard for rival exchanges to win. Publishers had spent years building header bidding precisely as a workaround to Google's auction advantages, which tells you the friction was real long before the DOJ got involved.
Present: what the April 2025 ruling found, and what it did not
On 17 April 2025, Judge Leonie Brinkema issued the liability ruling after a bench trial. This is the part to be precise about, because it is settled and the rest is not.
The court found Google liable on two counts. Under Section 2 of the Sherman Act, it ruled that Google had unlawfully monopolized two markets: the market for publisher ad servers and the market for open-web display ad exchanges (Simpson Thacher). Under Section 1, it found that Google had unlawfully tied two products together, requiring publishers who wanted real-time access to AdX demand to also use Google's ad server (CNBC). The court put Google's market share in publisher ad servers at roughly 90 percent, and AdX's share of the open-web exchange market at 55 to 65 percent, a level the judge accepted as monopoly power because AdX dwarfed its nearest competitor for years. It singled out specific tactics: "First Look" and "Last Look," which gave AdX preferential timing in the auction, and Unified Pricing Rules, which stopped publishers from setting higher price floors for competing exchanges.
The ruling was not a clean sweep. The DOJ also argued that Google monopolized a third market, advertiser ad networks, through Google Ads. The court rejected that claim, finding the government had not proven a distinct market there (TechCrunch). So the buy-side tool, the one most advertisers actually touch, was left out of the liability finding.
Here is the distinction that matters most. The April 2025 ruling decided liability only. It said Google broke the law. It did not say what Google must do about it. That second question goes to a separate remedies phase, and the two should not be blurred. A liability finding can be appealed, and Google has said it will appeal. The remedy is a different argument, argued later, on different evidence.
The remedies trial ran in September 2025, with closing arguments on 21 November 2025 (AdExchanger). The two sides asked for very different things.
The DOJ asked for a structural breakup. Its proposal: force Google to divest AdX, the exchange, with a roughly 15-month timeline to close the sale; open-source the final auction logic inside DFP; and, if those steps proved insufficient, divest the rest of the DFP publisher ad server too (Norton Rose Fulbright). Its core argument was that Google is a repeat offender and that behavioral rules would just invite years of monitoring and loophole-testing.
Google asked for behavioral remedies only. It proposed to make AdX real-time bids available to rival publisher ad servers, drop the contested First Look and Last Look practices, and build a server-to-server connection between DFP and Prebid, the open-source header bidding framework. Google argued most of this could be done within about 12 months, with one piece taking longer, and called a forced divestiture unworkable and unnecessary (Digiday).
As of late May 2026, the judge has not ruled on remedies. That is the current state of play, and it is worth saying plainly because headlines about a separate Google case can mislead. Google is also the defendant in United States v. Google LLC, the search antitrust case, a wholly different lawsuit in a different court before a different judge. In that case, Judge Amit Mehta found Google an illegal monopolist in general search in 2024 and, in September 2025, ordered behavioral remedies while declining to force a sale of Chrome. Both sides have since appealed it. None of that is this case. The ad-tech remedy is still pending.
What the judge signaled in court is not encouraging for the breakup camp. During closing arguments, Brinkema raised repeated doubts about a structural order. She questioned who would even buy AdX, noting that an obvious candidate like Microsoft would bring its own antitrust review and fresh delay. She pointed out that Google will almost certainly appeal, and that a structural remedy would be hard to enforce while an appeal is pending, potentially leaving the market in limbo for years (PPC Land). Earlier in the remedies hearings she had grown impatient with abstract expert testimony, asking to hear from people who actually knew what a divestiture would require (AdExchanger). Trial watchers came away expecting hard behavioral conditions as the likely outcome, with a structural breakup held in reserve rather than ordered outright (AdExchanger). Nothing is decided. That is informed reading of a courtroom, not a verdict.
The liability finding is already doing work outside this courtroom, though. In November 2025, a federal court in New York gave the April ruling preclusive effect in private litigation, meaning publishers suing Google for damages do not have to re-prove the monopoly from scratch (DLA Piper). The pressure is not confined to the United States either. In September 2025, the European Commission fined Google EUR 2.95 billion over self-preferencing in ad tech, ordered it to end the conflicts of interest in its stack, and kept open the possibility that only a structural separation will resolve the problem (Loyens & Loeff).
Future and impact: what a forced sale of AdX would actually mean
Set aside whether a divestiture happens. Suppose it does. What changes for the people who actually buy and sell ads?
The theory of the case is straightforward. AdX is the largest exchange for open-web display, and it sits inside the same company that owns the dominant publisher ad server and a dominant buyer tool. Separate the exchange, and you remove the incentive to steer auctions toward Google's own boxes. In principle that means more genuine competition between exchanges, more pressure on the fees each layer takes, and more transparency, because an independent exchange has no reason to hide how an auction cleared. Independent ad servers become viable again once they are no longer competing against a product subsidized by exchange revenue. Several publishers, including names like News Corp and Advance Local, told the court a clean structural break could actually be less disruptive than a long list of behavioral rules to police (Digiday).
The disruption is the other half of the story, and it is not small. Google Ad Manager is deeply woven into how a large share of publishers run their business, and unwinding AdX from DFP is an engineering project, not a signature. There is also a buyer problem. A divested AdX has to go to someone, and as the judge noted, the credible candidates are limited and several would trigger their own competition concerns. Google has argued that DFP is offered free to many publishers and questioned whether a new owner would keep it that way (AdExchanger). That points to a real risk: if Google's ad serving was effectively subsidized, a standalone version could cost more, and the publishers least able to absorb that are the small and niche ones the remedy is partly meant to protect (AdMonsters).
There is a quieter risk too. A forced separation could push Google's buy side further inward, toward spending more inside YouTube and its other owned properties, which would pull demand away from the open-web exchanges the remedy is trying to revive (AdExchanger). And in the gap between an order and its execution, standards could fracture as competing players push their own approaches, creating instability before any cleaner market settles in.
For practitioners, the honest near-term advice is that very little changes fast. A remedies ruling is still pending. Whatever it says will be appealed, and the appeal alone can run for years. The realistic planning horizon is to treat heavy dependence on a single vendor's stack as a risk worth diversifying against, to push for fee and auction transparency from whoever runs your exchange, and to watch the docket rather than the headlines. Implementation partners, Perform Digital among them, are increasingly asked to design ad stacks that can survive a vendor change without a rebuild, which is a reasonable response to an outcome no one can yet predict.
The line to hold in your head is the one this post opened with. The court has found the monopoly. It has not yet prescribed the cure. A forced sale of AdX is a real possibility and a genuinely consequential one, but as of now it is a proposal a skeptical judge is weighing, not a thing that has happened. Treat it that way.
Council summary
This post argues that the US ad-tech case against Google has produced one settled fact and one open question, and that practitioners should not let the two run together. The settled fact is the April 2025 liability ruling: a federal court found Google an illegal monopolist of the publisher ad server and the open-web exchange, and an illegal tier of the two. The open question is the remedy, where the DOJ wants a forced sale of AdX and Google wants behavioral fixes, and where a visibly skeptical Judge Brinkema has yet to rule. The reader's takeaway is to plan calmly: a divestiture is a real and consequential possibility, not a done deal, the search case is a separate matter that should not be cited as precedent here, and the sane hedge is to reduce single-vendor dependence and demand auction transparency rather than bet on any one outcome.
Comments