business
Google's Ad Empire Is About to Be Sentenced. The Judge Has Already Said She Is Skeptical of the Remedy That Would Actually Work.
Judge Brinkema's questions from the bench during November closing arguments telegraphed a behavioral package. Every comparable behavioral package, including the one Judge Mehta imposed on Google Search five months ago, has left the underlying monopoly intact.
Judge Leonie Brinkema is going to sentence Google’s ad tech business this week or next, and the questions she asked from the bench during November 21 closing arguments give the market a fairer-than-usual read on what is coming. According to trial coverage, she pressed the Justice Department on its failure to identify a plausible AdX buyer; she told Google’s lawyers the order needed to be “far more down-to-earth and concrete” than what either side had given her; and she said she was deciding “whether to order both structural and behavioral remedies or only behavioral remedies.” When a federal judge frames the choice that way, after spending an hour pressing the government on a missing buyer, the path to “both” gets harder to see. That is not a prediction she has made; it is the inference this article draws from the questions she chose to ask.
That is the news. The bench transcript points toward a conduct-only remedy. The case that produced the most sweeping antitrust liability finding against an American technology company since the 2000 Microsoft trial is, on the questions Brinkema chose to press in November, likely to produce a sentence that, on the historical record of every comparable tech antitrust order, leaves the monopoly in place. That reading is this article’s inference from her bench questions, not a forecast she has stated; judges routinely test both sides without telegraphing a ruling. What follows is the case for taking the questions she asked seriously as a signal.
Brinkema found Google liable on April 17, 2025 for monopolizing publisher ad servers and ad exchanges and for unlawfully tying the two products together. Her opinion described AdX as the “glue that sealed DFP inventory to AdWords demand,” and quoted Google’s own engineers acknowledging that the exchange’s 20 percent take rate extracted “irrationally high rents” from publishers. The DOJ asked for divestiture of AdX and, if necessary, of DFP. Google offered a package of behavioral concessions: open real-time bids to rival ad servers, abandon Unified Pricing Rules, share data files with publishers on request, build a server-to-server connection between DFP and Prebid. Closing arguments wrapped November 21. Brinkema told the parties she had begun drafting. She set a self-imposed March 31 deadline and missed it. The ruling is now overdue by more than a month.
By the numbers
- Google’s ad tech business generated $30 billion in 2024 revenue, double the figure from a decade earlier, in a US sector worth more than $300 billion.
- AdWords captures 89 percent of Google’s display ad spend versus 63 percent for DV360, and 46 percent of all open-web display advertising flows through AdWords.
- In 60 percent of AdX auctions there was zero competitive bidding, with more than half featuring only AdWords as a bidder.
- DOJ trial estimates put the customer migration time required to transfer AdX to a buyer at roughly four years; Google’s own implementation timeline for its proposed behavioral remedies is one year.
- The 2001 Microsoft consent decree’s behavioral provisions expired in 2007 with Microsoft’s Windows monopoly intact.
What Brinkema actually said
The argument that the bench questions point toward a conduct-only outcome rests on three exchanges in November.
First, on the buyer problem. After the DOJ’s rebuttal, Brinkema noted the government had failed to “identify a plausible AdX buyer” and observed that any large acquirer, with Microsoft mentioned by name, would face its own lengthy antitrust review. That review, she said, could outrun the implementation of any divestiture order, particularly if Google appeals. The DOJ’s answer, that the assets would find a market, did not appear to satisfy her.
Second, on commercial reality. Brinkema said the trial record left her “at a fairly abstract level” and warned that any structural order would have to be “far more down-to-earth and concrete.” That is a judge signaling she does not believe the government has done the operational work to make divestiture run.
Third, on settlement. She told the parties she could “see a settlement,” then conceded that the volume of pending private suits made one unlikely. The aside cuts both ways. A judge raising settlement at closing might be signaling preference for the lighter remedy she could ratify in a deal; she might equally be flagging that her structural order would be appealed for years and a negotiated outcome would land faster. Read alongside the buyer-feasibility and “down-to-earth” exchanges, the comment fits the first reading more readily than the second, but the inference is the article’s, not the judge’s.
Why behavioral remedies do not work in tech monopoly cases
The benchmark is no longer Microsoft 2001. It is Google Search 2025, the case decided five months ago by Judge Amit Mehta in the same federal courthouse system. Mehta found Google liable for monopolizing general search and search advertising in August 2024. He issued his remedies opinion September 2, 2025, and entered final judgment December 5. He rejected the DOJ’s request for a Chrome divestiture. He declined to impose a choice screen, citing the European experience showing such screens did not change competitive conditions. He ordered data sharing with qualified competitors and capped default search agreements at one year without auto-renewal, but allowed Google to continue paying for default placement.
The University of Iowa’s Herbert Hovenkamp and Stanford’s A. Douglas Melamed, two of the most cited antitrust scholars in the federal courts, wrote that some of Mehta’s explanations for the narrow scope of his payment restrictions seem “inconsistent with his liability decision last year.” Mehta had found at the liability phase that default agreements harmed competition by foreclosing rivals from distribution. At the remedy phase he permitted the same agreements with minor procedural changes. The scholars’ point is concrete: the liability decision rested on the conclusion that Google’s agreements impaired competition, yet the remedy permits Google to keep paying for default status on a shorter contract. Google could continue paying Apple to be Safari’s default for 95 percent of users and the order would be satisfied.
The 2001 Microsoft decree fared no better. The settlement required API disclosure and a three-person technical committee with source-code access. It did not require code changes and did not prevent future bundling. It expired in 2007, with Wharton’s review of the remedy concluding it had failed to produce “a meaningful choice among competing PC operating systems.” Andrew Chin, a North Carolina law professor who studied the decree’s compliance reports, argued it had effectively granted Microsoft a “special antitrust immunity” to license Windows on terms that destroyed competition.
What Brinkema’s order is likely to contain
The signals from the bench, read together with the structure of Google’s proposal, point toward a ruling that requires Google to make AdX real-time bids available to rival ad servers, drop Unified Pricing Rules, share data files with publishers, and submit to a court-appointed monitor for a period of years. The package would draw heavily from Google’s own proposal because Brinkema told the DOJ in November she did not see how its package would actually run. The buyer-feasibility exchange suggests the ruling is unlikely to order immediate divestiture of AdX; the more probable shape, on the available signals, is one that reserves divestiture as a contingent remedy if conduct relief fails, in language that mirrors Mehta’s reservation in the search case. That contingency is the door Brinkema can leave open without writing the order she has signaled she does not want to write. The forecast above is the article’s reading; the actual order could go elsewhere.
That outcome is what the trade press has called Google’s pattern of losing in court and winning anyway. The pattern is now consolidated as a doctrine. Federal courts have found Google liable for two distinct monopolies in 18 months. The remedy in the first case did not change the conduct. The remedy in the second is about to follow the same template, written by a different judge in a different district, against the same defendant. The Amazon trial expected later this year and the Apple App Store remedy phase will both inherit it. The question for the next decade of antitrust enforcement is not whether plaintiffs can win liability findings against the largest technology platforms. It is whether any sitting federal judge will order the structural relief their own findings imply. Brinkema’s overdue ruling is the next data point. The questions she asked in November suggest the answer.