Judge Leonie M. Brinkema's 106-page explanation for leaving Google's advertising exchange in Google's hands, entered on September 2, 2026 in the Eastern District of Virginia and sealed for two weeks, rests on findings that the two-page order of the same day did not disclose: the government never produced a buyer for AdX, no court has ever ordered a sale to cure an unlawful tie, and the case for a breakup "boils down to a lack of trust." In return, Google will carry a package of conduct rules for six years rather than the fifteen the government wanted, and the court has ordered that they apply everywhere Google's ad tech operates.

In Short

A federal judge has now written out why Google gets to keep AdX, the exchange at the centre of its ad tech business, rather than being made to sell it. That matters to anyone who buys or sells display ads through Google Ad Manager, because the remedy is a set of rules about how Google behaves, not a new owner. If you run a publisher site, the practical change is that Google's advertiser demand will have to show up in Prebid and in rival ad servers on the same terms it gets inside Google's own ad server, with a court-appointed monitor checking the work for as long as the judgment lasts. That is a long way from the breakup the government spent three and a half years asking for!

The reasoning the industry argued about before it could read it

The order that appeared on the docket on September 2, filed as Document 1857 in Case 1:23-cv-108, rejected three structural remedies, accepted most of the parties' behavioral proposals "as modified", and said nothing else. The reasoning sat in Document 1858, the memorandum opinion, which the court directed be filed under seal for 14 days "to enable the parties to review it for any needed redaction". PPC Land has now reviewed the full document.

The procedural arithmetic sits on the first page: four months of additional discovery after the April 17, 2025 liability ruling, an 11-day bench trial, 26 witnesses, more than 100 exhibits, then post-trial briefs and competing proposed final judgments. The two proposals overlapped so heavily that the opinion describes most of the behavioral remedies as "not seriously contested". What divided the parties was whether conduct rules were enough, and for how long: the plaintiffs wanted a 15-year judgment on top of a forced sale, Google asked for six years and no sale. The answer arrived in one sentence. "For the reasons discussed in this Memorandum Opinion, the Court finds that structural remedies are neither realistic nor needed," Judge Brinkema wrote.

Why the sale failed

The government asked for three things: the divestiture of AdX to a buyer approved by the plaintiffs "in their sole discretion", the open-sourcing of the decision-making code in Google's publisher ad server, DFP, under a neutral administrator, and a contingent sale of whatever remained of DFP if competition had not returned almost a decade later. The objections that sink the AdX sale run through all three.

The first is legal. Plaintiffs carry a "heavy burden" to justify the "radical structural" remedy of a forced divestiture, and the conduct condemned at trial was a tie between integrated products. "Where, as here, the gravamen of the violation was an unlawful tie among integrated technology assets, divestiture is unprecedented," according to the opinion, which notes that tying cases such as International Salt and Kodak ended with duties to sell or license on fair terms, not with forced sales.

The second is commercial. Paul Crisci, the plaintiffs' investment banking expert, testified that AdX "would be attractive to a credible and motivated pool of buyers", but his view extended only to initial interest. Three industry witnesses said they might be interested. Andrew Casale of Index Exchange saw "an opportunity to consolidate" but had not analysed feasibility. Rajeev Goel of PubMatic said his company did not "have enough information about AdX as an entity to know if [it] would be interested in bidding on it or not." James Avery of Kevel wanted the asset but conceded that Kevel lacked the funds on its own. "Tellingly, during the course of the remedies trial, Plaintiffs failed to identify a viable AdX acquirer," the court wrote, contrasting the record with the search case, where the plaintiffs "touted many likely buyers."

The third is the argument the plaintiffs actually made: that Google's record of pivoting from one practice to the next made any conduct order a "never-ending game of Whac-A-Mole". Casale testified that behavioral remedies "will help, but they do not provide guaranteed certainty" because "a very creative mind can develop another preferencing feature to implement into DFP tomorrow." The court was unmoved. "This Court is unaware of any court ever having required a remedy in an antitrust case to provide 'guaranteed certainty,'" the opinion states, before reducing the plaintiffs' position to its core: "the essence of Plaintiffs' rationale for seeking divestiture boils down to a lack of trust that Google will comply with an Order from this Court and an unrealistic desire for certainty."

The concession had come from the plaintiffs' own economist. Professor Robin Lee testified that behavioral remedies would suffice if he could be confident that Google "would actually act in complete good faith and follow an injunction." The justice system already answers that, in the court's view: Google is held in check by "not one, but two swords of Damocles", a court-supervised injunction backed by the contempt power, and the private lawsuits pending against the company. The opinion lists actions by PubMatic, OpenX, Magnite, Equativ, Index Exchange and Gannett in the Southern District of New York and notes, pointedly, that many were filed by witnesses who testified for the government.

Time closed the argument. Under the plaintiffs' own proposal, migrating AdX to an approved acquirer could take up to 540 calendar days after a sale was signed, followed by up to two more years to move AdX's customers, and none of that could start until a buyer was found. Every behavioral remedy, by contrast, would be in effect within 15 months. The causal link was thin as well: structural relief requires "a significant causal connection" between conduct and market power, and the record contained ample lawful conduct that helped Google, from Search-generated advertiser demand to the approved acquisitions of DoubleClick and Admeld. Whether structural remedies were needed was "arguably closer" than in the Microsoft and Google Search cases, the opinion acknowledges, but the evidence did not carry it, "especially where it is unclear whether they would actually further improve competition."

Seventeen years of knitting

The engineering record is where the plaintiffs' timetable fell apart. Goranka Bjedov, one of their experts, described the AdX migration as "a large migration, but it's also the kind of migrations that Google or Facebook ... do once or twice a year" and put the job at 18 months. A second expert, Weissman, called the source code "beautiful" and "very well structured".

Google's witnesses described something else. Adkins, a Google witness, compared the relationship between AdX, DFP and Google's core systems to knitting: code pulled in from elsewhere at compile time until the whole is "tightly coupled". According to the opinion, AdX and DFP have been woven into that infrastructure for more than 17 years, and the infrastructure involves more than 100 million lines of code, data centres and undersea cables. The comparators the court found persuasive were Google's own migrations. Moving DoubleClick onto Google's infrastructure took seven years. Replacing one AdX and DFP dependency took five years and 40 engineers. A Gmail dependency took eight years. Outside Google, LinkedIn's move to Microsoft's public cloud was abandoned after four years, and the "far simpler" partial migration of Netflix's DVD business to Amazon Web Services, which Bjedov herself offered as the closest analogue, took seven years. "Against these comparators, Plaintiffs' timeline estimates are unrealistic," the court wrote.

The plaintiffs had relied on an internal Google engineering analysis from 2023 and 2024, reported during the trial in September 2025, that found a business divestiture of AdX technically feasible, with two years as the "outer bound" for a transitional period. The opinion finds the study proves less than claimed: it covered only open and private auctions, "did not assume any infrastructure dependencies", and transferred no operational technology, only contracts, patent licences, revenue and a reference source code that Tim Craycroft, Google's vice-president for YouTube, app and display ads, described as "an instruction manual for rebuilding AdX."

Even a successful sale carried a product risk. Professor Jason Nieh, Google's technical expert, testified that in a best-case scenario the result "might be a product that is not comparable in terms of the functionality and scaling capabilities compared to AdX and DFP today." The court called it "technology that absolutely has to work for customers" and noted Elizabeth Douglas of wikiHow describing service disruptions as "really, really stressful" for a small publisher. For the contingent sale of DFP, one more number applied: DFP is free to more than 90 percent of its publisher customers, so "the possibility that costs will increase for publishers following divestiture is very real."

An auction that takes 10 hours to compute

The open-source proposal died on a narrower point. The plaintiffs argued that DFP's "final auction logic" is a black box that publishers and rival exchanges cannot audit, and witnesses from Advance Local, Index Exchange and the Daily Mail said so at trial. But the government never argued in the liability phase that the black box contributed to Google's monopolies, and the liability opinion made no such finding. Assertions "standing alone and unconnected to specific liability findings, cannot be utilized to justify specific remedial provisions," the court wrote. "For this reason alone, the Open-Source Auction will be rejected."

The feasibility findings are recorded anyway. The plaintiffs' definition of the code to be published swept in Enhanced Dynamic Allocation, Billable Event Rate adjustments, Reserve Price Optimization and Optimized Pricing Rules, functionality that Google's witnesses called the "heart and soul of any ad server" and that Glenn Berntson, Google's engineering director, said looked "an awful lot more" like the core of an ad server than a final auction. Google's own internal analyses had considered only the last comparison between direct and indirect demand, and even that was estimated at four years. Enhanced Dynamic Allocation depends on a seven-day window of historical bid distribution data. Google runs that calculation once a day, and according to Berntson it takes roughly 10 hours on 4,000 computers; for a single ad request from a single publisher it would take between 15 and 45 minutes. "Plaintiffs' portrayal of the Open-Source Auction as a 'plug and play' system is therefore a broad oversimplification," the opinion concludes.

Who would run the code? Publishers could host it, pay a managed service such as a Prebid Server provider, or run it inside their existing ad server, and each option would oblige them to evaluate the code and its updates. Only "a few of the largest publishers" could operate something "extremely complex" on their own, a Google witness, Levitte, testified. The proposal would also have reached the roughly 8,000 publishers who use Google Ad Manager only for direct sales.

What Google now has to build

The behavioral package sorts into five parts: an injunction, interoperability of AdX and DFP with Prebid and of AdX with rival ad servers, data sharing, non-discrimination rules, and constraints on Google's buying tools. Where the two sides proposed different deadlines, the court applies a rule of thumb: Google's shorter timeframe wins because its engineers are "intimately familiar with the products at issue", and where Google's timeframe is longer the parties must meet and agree.

The injunction

Google will be barred from enforcing any policy or contract term that ties DFP to AdX and from reimplementing First Look and Last Look, and must deprecate and not reimplement Unified Pricing Rules in DFP for all indirect transaction types. Publishers will be allowed to set different floor prices across buying tools and exchanges, and Google may not discount DFP fees in exchange for uniform pricing. The opinion notes that the plaintiffs' liability expert, Tim Simcoe, had calculated that eliminating Unified Pricing Rules alone would reduce AdX's take rate from 20 percent to roughly 16.6 percent, a level the court calls comparable to some competing exchanges. The plaintiffs had told the court at a May 2, 2025 hearing that these provisions "would absolutely address our concern about the prior illegal conduct."

Prebid gets both AdX and DFP

Prebid, the consortium of non-Google industry members that maintains the open-source header bidding software, becomes the pivot of the remedy. Google must build an application programming interface between AdX and Prebid so that Prebid can solicit real-time bids from AdX for all indirect open-web display inventory, and an interface plus a server-to-server connection between DFP and Prebid so that DFP publishers can route all indirect demand, and all programmatic direct demand, meaning programmatic guaranteed and preferred deals, through Prebid. The integrations must work "in a functionally equivalent manner and on materially identical terms" as DFP works with AdX, without additional latency and passing the same information AdX passes to buying tools.

Google's initial proposal was confined to open auctions until Craycroft committed on the stand "to enlarge the definition". The court draws the line at non-programmatic direct deals, quoting Prebid president Michael Racic that such deals "shouldn't be part of the final auction logic" because the price is already agreed. Google will also provide technical support to Prebid and build tools forcing publishers to make a mutually exclusive choice: call AdX from Prebid or from DFP, not both. The effect, in the court's description, is that DFP must share all pricing information, including the pricing of direct-sold campaigns, and the Google audience data it currently shares only with AdX, with every participant in a Prebid auction. Rival exchanges would compete "at parity", in Casale's words. The plaintiffs wanted DFP-Prebid interoperability within 18 months; Google committed to 12 to 15, and the court defers to Google.

Rival ad servers get AdX bids

Google must enable AdX to submit real-time bids to rival publisher ad servers "in the same manner it currently submits real-time bids to DFP" for all indirect demand, with technical support. This answers "the central competitive injury identified during the liability trial": publishers who wanted AdWords demand had to use DFP, and ad servers without AdX could not compete. Gannett had turned down offers to halve its ad serving fees for that reason; Equativ lost large media clients for the same one. Avery of Kevel testified that "the key thing that I think enables us to compete in the publisher ad server market is if we can get demand directly from AdX in a fair way." Here the government's deadline was shorter, six months against Google's twelve, so the parties must negotiate a date.

Data leaves DFP, but only for open-web display

Google must give publishers their historical and configuration data from DFP, along with ongoing AdX bid data covering every winning and losing bid in each auction, and build an interface for exporting that data to rival ad servers. The court adopts it as a way of lowering the switching costs that make ad servers "sticky products", but rejects the government's demand that the data cover video and in-app formats: the plaintiffs built their liability case on the premise that open-web display is a distinct market and "cannot now abandon that distinction" to obtain remedies across every channel.

A data file for every ad served

Google's substitute for open-sourcing is documentation. Under its proposal, adopted as "reasonable", Google will provide technical documentation explaining DFP's final logic and a data file showing, for each ad served, the candidate prices and every adjustment used to pick the winner: the price of any direct deal considered, any programmatic bids, any bid adjustment applied and the category of that adjustment, "along with an explanation as to which is the winner and why was it selected," in Berntson's words. The plaintiffs objected that the proposal was too vague to test and would not let a rival ad server "kickstart a new business". The court rejects both objections and hands the format and frequency of disclosure to the parties, working with the Monitor and Technical Committee. "That will be the best vehicle for addressing the ad tech industry's call for transparency," the opinion states.

Non-discrimination, with carve-outs

Google must use and pass data signals on a non-discriminatory basis, AdX must pass bids to ad servers and header bidding auctions regardless of who owns them, and Google may not vary AdX's revenue share, add latency, respond to fewer bid requests or transmit less information depending on whether a publisher uses DFP, a rival server or Prebid. The court endorses the plaintiffs' bar on Google accepting payment for preferential routing of indirect demand, but finds the government's drafting "too indefinite": no compliance mechanisms, no exception for Google's privacy policies or governing law, and a definition of "Data Signals" as any information a buying tool uses to value an impression. "Google is entitled to know precisely what information it must share," the court wrote, and sent the parties away to draft language that works.

AdWords is in, DV360 is out

Google argued that no buy-side remedy was available because neither AdWords nor DV360 sits in a market where liability was found. The court accepts half of that. "Google reads the Court's liability opinion too narrowly," it wrote of AdWords, whose demand it had called the "golden goose" at the May 2025 hearing: restricting that demand to AdX was the primary source of Google's exchange monopoly. Leaving AdWords out would allow Google to recreate the tie by other means: Google Partner Bidding, or gBid Direct, already lets Google's buying tools bid directly into publisher auctions for mobile app inventory, and AdWords could vary its fees by ad server, as Lee testified it already does across exchanges, charging "more than double the take rate on a rival ad exchange than it does in AdX."

The remedy adopted is Google's narrower version. Google may not build any direct bidding integration between AdWords and DFP for indirect open-web display inventory, but it may integrate AdWords directly with Prebid. The plaintiffs' formulation, which would have let DFP transact with AdWords "only through an Ad Exchange", was rejected because it would have barred AdWords from doing what other buying tools already do through direct connections, an "anomalous competitive disadvantage" in Craycroft's words; The Trade Desk's OpenPath, which bids into the ad server and skips the exchange, is the example the opinion gives.

AdWords must also decide when, how and how much to bid without regard to whether the publisher uses Google's tools, and must not prioritise AdX on the basis of ownership. It may prioritise AdX to deliver advertiser return on investment, on factors such as information, impression quality, fraud protection or privacy protection, which the parties must now enumerate. First-party data from YouTube, Gmail and Search gets its own clause: Google may not use it to personalise an open-web display impression unless it does so without regard to whether the bid request came from AdX. "For now, this non-discrimination remedy merely puts in writing the status quo," with the door left open to modification.

DV360 escapes entirely. It "did not play a role during the liability phase," the opinion states, and Lee conceded it was not part of the markets he evaluated. Two figures on AdWords' habits close the section: since the 2015 AWBid feature, AdWords has integrated with about 50 third-party exchanges yet placed only about 3 percent of its impressions outside AdX, and Lee put the share of AdWords open-web display transactions on non-Google exchanges at 11 percent in 2022 and 5 percent in 2024.

The term, the start date and the reach

The plaintiffs asked for 15 years, extendable on request. Google asked for six, starting 120 days after entry. The court takes Google's term and the search court's start date. Six years accounts for the year it expects to spend establishing the Monitor and Technical Committee, gives competitors time to build capacity, and matches the term Judge Amit Mehta adopted in the search case and the five years of the Microsoft final judgment. A 15-year judgment in an industry of "constant and rapid change" would be, in the words of the Microsoft remand, "not unlike trying to shoe a galloping horse". If the judgment has not been fully satisfied after six years, the court can extend it. The effective date is 60 days after entry, except that appointing the Monitor and forming the Technical Committee begins immediately, and the court expects the behavioral remedies to be implemented within 15 months.

The geographic finding may matter more outside the United States than any other line in the document. Google argued, citing the Google Play injunction, that comity forbids a US court from ordering conduct abroad. The court disagrees. The liability opinion defined both markets as worldwide, ad tech providers "have built global infrastructure" and "view themselves as competing globally", and an injunction confined to the United States would risk adverse effects at home. Global application also means one set of product changes rather than regional variants, since Google charges consistent exchange fees and rolls out changes in every region at once. "For these reasons, the Final Judgment will apply beyond the United States and reach globally."

The monitor, the committee and the compliance officer

Both sides proposed a Monitor and a three-member Technical Committee, and the court agrees that appointed experts are the right way to resolve "inevitable disputes" without a judge refereeing each one. The Monitor will be chosen from a slate of three candidates, one party proposing and the other selecting, and will serve until the judgment expires, with power to interview any Google employee, inspect documents, source code and algorithms, access systems and premises, and hire staff at Google's expense. Each party picks one Technical Committee member and those two choose the third. Google pays all of them.

Where the proposals diverged, the court mostly sided with Google. The Monitor's work plan is due 45 days after appointment, not 30. Reports go to both parties in writing quarterly, rather than "at intervals and in a form specified by Plaintiffs". Google will have at least 30 days to respond to a finding of a potential violation, or to a complaint, before the matter reaches the court, a "broad right to object" that Judge Mehta also granted, with the warning that abuse could see it restricted. The plaintiffs' attempt to limit Google's objections to "malfeasance" by the Monitor, proven by clear and convincing evidence within 10 days, is dismissed in three words: "Plaintiffs' position is unreasonable."

Two of the plaintiffs' administrative proposals survive. Google must appoint an Internal Compliance Officer to distribute the judgment to officers and employees, collect annual written certifications, preserve documents and operate a confidential channel for reporting potential violations. Until the Monitor is in place, Google's chief financial officer and general counsel must sign an affidavit every 30 days describing compliance efforts, and the plaintiffs may inspect records and interview staff. The parties must file a status report within 90 days of the effective date.

What the court threw out

The escrow fund is gone. The plaintiffs wanted Google to deposit 50 percent of the net revenues of AdX and of DFP, counted from April 17, 2025, into an account controlled by the Monitor, to defray publishers' costs of switching ad servers among other purposes. The court found the "lack of specifics is fatal": no linkage between half of net revenues and the supracompetitive fees found at trial, no method for distributing money among switching publishers, and, for DFP, no finding of overcharging at all, since the liability opinion held that "Google has not exercised its monopoly power to raise DFP's prices." The court did reject Google's argument that disgorgement is unavailable under the Sherman Act.

The anti-retaliation, non-interference and anti-circumvention clauses fall for the reason Judge Mehta gave in the search case: Rule 65(d) requires an injunction to describe the restrained conduct in reasonable detail, and a ban on retaliation "in any form", on "any action" impeding implementation, or on conduct "designed to replicate the effect" of the violations offers "scant notice" of what is forbidden. An annual anti-reprisal policy for employees survives. A proposed compliance program covering antitrust, evidence preservation and privilege is rejected as untailored, because the court declined to sanction Google for spoliation at the liability stage, and any request for fees and costs must come by separate motion.

The market the court was ruling on

The opinion's factual background covers two developments since the liability trial: the decline of open-web display and the arrival of AI and supply-path optimization. Global digital ad spend reached 424 billion dollars in 2023 and is projected to exceed 605 billion dollars by the end of 2027, according to Crisci's testimony, and open-web display has held a roughly 30 percent share of ad format revenues while growing more slowly than retail media, streaming, native, audio and AI chat. The Daily Mail's Matthew Wheatland testified that display generates 50 to 60 percent of the publisher's digital revenue. Using AdWords as a sample, indirect open-web display fell from more than 40 percent of impressions in January 2019 to 11 percent in June 2025, while YouTube's share of non-search AdWords impressions rose from just under 30 percent to over 50 percent. One detail differs from the version Google filed a year ago: the September 2025 brief PPC Land covered at the time dated the 11 percent figure to January 2025, whereas the opinion, citing Craycroft's testimony and a trial exhibit, places it in June 2025.

AI, the court finds, has not yet disrupted ad tech the way the search court found it disrupting search; Goel called it "an enabler" of automation, and the opinion records only early experiments with ads inside chat interfaces. Supply-path optimization gets the more consequential treatment, because it bears on both the AdWords remedy and the length of the term. Goel testified that such activity "could be 75 percent of PubMatic's total buyer activity in the next several years", and Jay Friedman of Goodway Group said standalone exchanges "are likely not needed for very long." The trend, the opinion says, might blur the line between exchanges and ad servers, and it is one reason a 15-year judgment looked stale before it started.

Scale, finally, is quantified. Berntson testified that Google Ad Manager processes about 8.2 million ad requests and 60 million bid requests per second at peak, which Nieh translated into more than 600 billion ad requests and almost five trillion bid requests a day. AdX held 63 to 71 percent of global open-web display exchange transactions among exchanges that produced data between 2018 and 2022, nine times the 6 percent of its nearest rival, and charged a 20 percent take rate where rivals charged closer to 10 percent. DFP held 91 percent of the ad server market in 2022, and Advance Local's Whitmore told the court that "there isn't an alternative publisher ad server right now, and there might not be for a while."

Why the opinion matters for the marketing community

For a media buyer or a publisher, the document changes nothing in the account this week and a great deal over the next 15 months. Google Ad Manager remains one product under one owner, and AdX's fee is what it was yesterday. What the opinion fixes is the shape of the obligations the final judgment must contain, and the timetable by which they arrive.

The ruling was late by the court's own signalling. PPC Land's account of the closing arguments on November 21, 2025 recorded Judge Brinkema's doubts about a divestiture timeline in almost the terms the opinion now uses, and reporting on the trial's end noted that a ruling was expected in January or February 2026. The competing proposals the opinion adjudicates were set out in the final briefs of November 3, 2025, where the government asked for AdX to be sold within twelve months and Google offered conduct alone. Some of that conduct has already happened: Google removed Unified Pricing Rules from Ad Manager in December 2025, restoring buyer-specific floors before any remedy was ordered, and the Simcoe estimate of a 16.6 percent take rate without them puts a number on what publishers can measure over the coming quarters.

The pattern now holds across two courts. Judge Mehta rejected a Chrome divestiture on September 2, 2025 and ordered data sharing instead; his final judgment of December 5, 2025 runs for six years with a Technical Committee, the architecture Judge Brinkema borrows here. Google appealed that judgment and asked the D.C. Circuit on May 22, 2026 to reverse the liability finding outright, one reason the Virginia opinion treats an appeal as likely when weighing how long a forced sale would take. In Brussels, the European Commission fined Google 2.95 billion euros on September 5, 2025 for self-preferencing across the same products and published the provisional public version of that decision on January 14, 2026 while still weighing structural relief. Because the Virginia judgment reaches globally, the product changes it requires will be consistent across regions, according to the opinion, whatever the Commission decides.

The "two swords" passage also has an audience beyond Alexandria. The private suits the court cites as a check on Google's behaviour belong to a queue PPC Land has tracked since OpenX filed on August 4, 2025, followed by Magnite, Index Exchange, The Atlantic and, on August 3, 2026, Teads. Those plaintiffs now hold a federal opinion describing them as one of the two mechanisms that make a breakup unnecessary. The economic backdrop is visible in Alphabet's own numbers: Google Network revenue, the segment that pays open-web publishers, fell 1 percent to 7.3 billion dollars in the second quarter of 2026 after a 4 percent decline in the first, and publishers have reported impression declines of around 40 percent.

The open question is enforcement. The opinion leaves the specifics of the non-discrimination rules, the AdX-to-ad-server deadline, the advertiser-value exception and the format of DFP's documentation to negotiation between the parties, with the Monitor and Technical Committee as the backstop. A joint proposed final judgment is due by October 2, 2026; any provision still disputed must be filed in both versions, labelled by proponent, without argument. Only then will the industry know the dates on which AdX bids begin to arrive in Prebid.

Timeline

  • January 24, 2023: The Department of Justice and state attorneys general file the ad tech antitrust suit in the Eastern District of Virginia
  • April 17, 2025: Judge Brinkema finds Google monopolised the publisher ad server and ad exchange markets for open-web display and unlawfully tied DFP to AdX
  • May 2, 2025: At a hearing, the plaintiffs acknowledge that the proposed injunction provisions would address the prior illegal conduct
  • September 2, 2025: Judge Mehta orders data sharing rather than a Chrome divestiture in the search case
  • September 5, 2025: The European Commission fines Google 2.95 billion euros in its ad tech case
  • September 5, 2025: Google files a memorandum warning that open-web display is in rapid decline
  • September 22 to October 6, 2025: The 11-day remedies trial hears 26 witnesses
  • September 2025: Trial testimony shows Google had modelled shutting AdX down
  • November 3, 2025: The parties file post-trial briefs and competing proposed final judgments
  • November 21, 2025: Closing arguments conclude, with the bench questioning the practicality of an AdX sale
  • December 5, 2025: Final judgment is entered in the search case with a term of six years and a Technical Committee
  • December 2025: Google removes Unified Pricing Rules from Ad Manager
  • January 14, 2026: The Commission publishes the provisional public version of its ad tech decision
  • January 16, 2026: Google files its notice of appeal against the search judgment
  • May 22, 2026: Google asks the D.C. Circuit to reverse the search liability finding in full
  • July 22, 2026: Alphabet reports Google Network revenue down 1 percent to 7.3 billion dollars for the second quarter
  • August 3, 2026: Teads files a follow-on complaint against Google in New York
  • September 2, 2026: Judge Brinkema enters a two-page order rejecting all three structural remedies; the 106-page memorandum opinion is filed under seal
  • September 16, 2026: The 14-day sealing period for redaction requests ends
  • October 2, 2026: Deadline for the parties to file one jointly proposed final judgment

Summary

Who: Judge Leonie M. Brinkema of the United States District Court for the Eastern District of Virginia; Google LLC; the United States Department of Justice and plaintiff states; Prebid; rival publisher ad servers and exchanges including Kevel, Equativ, Index Exchange and PubMatic; publishers including the Daily Mail, wikiHow and Advance Local; and a court-appointed Monitor and Technical Committee still to be selected.

What: A 106-page memorandum opinion rejecting the divestiture of AdX, the open-sourcing of DFP's final auction logic and the contingent divestiture of DFP Remainder, and adopting, with modifications, behavioral remedies covering the AdX-DFP tie, AdX and DFP interoperability with Prebid, AdX bids for rival ad servers, data sharing limited to open-web display, per-ad DFP documentation, non-discrimination rules and limits on AdWords, for six years with global reach.

When: Entered on September 2, 2026 and filed under seal for 14 days; a joint proposed final judgment is due within 30 days, by October 2, 2026, with the judgment to take effect 60 days after entry and the behavioral remedies expected within 15 months.

Where: The Alexandria Division of the Eastern District of Virginia, in Case 1:23-cv-108, with obligations that the court ordered to apply globally.

Why: The court found that a forced sale is unprecedented for a tying violation, that no viable AdX buyer was identified, that the plaintiffs' case rested on distrust rather than evidence that conduct remedies would fail, that the government's engineering timelines were unrealistic against Google's own multi-year migrations, and that behavioral remedies backed by a monitor, the contempt power and pending private lawsuits would restore competition faster than a divestiture contested through appeal.