3 SEP 2026 — Judge Leonie Brinkema has rejected the Justice Department's proposal to force Google to divest AdX, ordering behavioural remedies including interoperability with rival platforms instead. One of her stated reasons is that nobody ever identified a credible buyer for AdX. A structural remedy needs somewhere for the asset to go, and at this scale there may be nowhere.
What was decided
Brinkema, sitting in the Eastern District of Virginia, found in April 2025 that Google had illegally maintained a monopoly in advertising technology. The remedy ruling rejects the divestiture of AdX, the open-sourcing of the final auction logic in DFP, and a contingent divestiture of the DFP remainder.
In their place she has ordered most of the behavioural remedies the parties proposed, requiring Google's advertising tools to interoperate with rival platforms rather than be sold. She reasoned that behavioural fixes would work faster than a breakup bogged down by years of appeals.
Both sides have 30 days to submit a jointly proposed final judgment. The fuller written ruling is sealed for 14 days for redactions, so the specific obligations are not yet public. Google's Lee-Anne Mulholland said the company was very pleased the court rejected the proposal to break apart tools that help small businesses reach new customers and grow.
No buyer is a remedy problem, not a legal one
The lack of a credible buyer for AdX was a decisive factor, and one most summaries will miss.
A court ordering a divestiture has to be satisfied that the divested business will operate as a competitor afterwards. That requires a buyer with the capital to acquire it, the technical capacity to run an ad exchange at global scale, and no competition problem of its own. The set of organisations meeting all three is very small, and several of them are the other companies a competition authority would least like to hand it to.
Which produces an uncomfortable result. A market can become so concentrated that the main remedy for concentration becomes impossible. The only companies able to buy the asset are the ones that would recreate the problem.
Behavioural remedies have a poor record and a real advantage
Brinkema's practical argument is strong. A divestiture order would have been appealed, and the appeals in cases of this kind run for years while the conduct continues unchanged. Behavioural obligations can take effect while the appeal proceeds.
The history of behavioural remedies in technology markets is poor. They require constant supervision and are hard to enforce when the defendant holds all the technical cards. Interoperability in particular is easy to state and hard to specify: the question is always whether an interface is genuinely equivalent to what the company's own products use, and answering it requires technical oversight that courts are not equipped to sustain.
The joint proposed judgment, due in 30 days, will determine whether this ruling has teeth. Vague obligations produce years of disputes about compliance; specific ones produce a compliance programme.
Winning the liability finding and losing the remedy
The Justice Department established that Google illegally maintained a monopoly, which is the hard part of an antitrust case and the part that usually determines the outcome. It then asked for the remedy it wanted and did not get it.
That sequence is more common than the coverage of either stage suggests, and it exposes something about how these cases are built. Proving liability is about past conduct. Devising a remedy is about a workable future, which requires the plaintiff to have already answered questions like who buys the asset and how it will operate.
Preparing that answer is a different discipline from proving the case, and it is generally under-resourced relative to the liability phase. A department that had arrived with a named, credible acquirer for AdX would have faced a materially different question from the one it lost.
Two Google cases keep getting merged
Coverage regularly blends this ruling with the separate search monopoly case before Judge Amit Mehta, in which divestiture of Chrome and Android was rejected in September 2025 and remedies covering default placement deals and search data sharing were ordered.
They are different cases, in different districts, before different judges, about different products, filed three years apart in 2020 and 2023. A remedy attributed to one is frequently a remedy from the other, and the resulting picture of what Google is actually required to do is wrong in both directions.
For anyone tracking Google's obligations, the distinction is critical. The ad-tech remedies apply to publishers and advertisers using AdX and DFP. The search-case remedies apply to distribution agreements and data access. Neither substitutes for the other.
What it means for advertisers and publishers here
Nothing changes this quarter. The ruling is sealed, the final judgment is 30 days away, and implementation follows after that.
If the sealed obligations deliver a genuine interoperability regime, publishers should prepare for it. A publisher currently locked into Google's stack because rival exchanges cannot bid on equal terms would gain the option to diversify demand sources, and the organisations that benefit are the ones that have already built the measurement to tell whether a change in exchange mix costs them revenue.
We reported on the FTC's allegation that Amazon described a second-price auction while running something else. Both cases turn on the same underlying condition: an advertiser cannot see the mechanics of the auction they are bidding into, and every remedy in this area is ultimately an attempt to make that visible.
