Remedies Closing Arguments, November 21, 2025: A critical inflection point for the future of the open internet and tech accountability

Today, the DOJ and Google presented closing arguments about their respective remedies proposals.

Judge Brinkema began the day by explaining that she anticipates this will not be the last set of arguments before the Court - so this is more likely opening closing arguments.

DOJ Closing Arguments:

DOJ’s Matthew Huppert delivered the first part of the government’s closing. He explains that we are here because Google systematically dismantled two adtech markets vital to the open internet, and has cast a long, menacing shadow over the adtech industry for too long.  It can be hard to picture in cases like this, he explains, a world without the monopolist. Google’s witnesses, which he reminds the Court were all compensated by Google, testified that Google should maintain its monopoly, and that Google alone can operate tools with quality, privacy, and security in mind. What this indicates is that Google disagrees with the liability findings — as Google’s economic expert, Dr. Lerner, seemed to indicate during the defense remedies case in chief. He reminds the Court that Googler Tim Craycroft testified that Google could not commit to do anything differently at this stage with regard to AdX’s take-rate, despite the unchanged 20% rate having been found supracompetitive.

Huppert paints a picture of hope. Google’s dominance is not inevitable. A brighter future for the open web is possible, and is the Court’s duty to facilitate. The question is how: through structural changes, or through behavioral remedies only, that would leave Google’s monopoly in place, but try to constrain it. A brighter future for the open web will never come to pass, he says, if Google’s illegally-obtained monopoly remains in place. Competition will never flourish in Google’s shadow.

He recaps the DOJ’s objectives framework for effective remedies. Effective remedies must (1) unfetter the market, (2) terminate the illegal monopoly, (3) deny the fruits of the violation, and (4) prevent re-monopolization. He cites the Supreme Court opinion in United Shoe to pose that the Court’s remedy must cause complete “extirpation” (i.e. eradication) of Google’s monopoly, “root and branch.”

An effective remedy must be “sure,” he says, as this assurance is necessary to give the public and the market confidence. Cumbersome conduct remedies are disfavored because they would have the Court acting as a central planner and entangle the government in remedies administration indefinitely. He raises the notion originating from the Supreme Court in its ruling in NCAA v. Alston and cited in Judge Mehta’s Search remedies ruling that “caution is key” in crafting remedies. He clarifies that contrary to Google’s framing of this precedent as supporting its argument against divestiture, what it actually means is that remedies need to be strong enough to eliminate the monopoly and simple enough to not entangle the government in managing business.

He calls back to the Court’s exchange with Google’s Craycroft where Judge Brinkema suggested that wiping out AdX might be a simple and elegant solution to demonstrate that structural solutions can indeed achieve this goal. The burden to Google only matters insofar as the Court is deciding between multiple effective remedies.

Huppert returns to Google’s apparent position that it won its monopolies fair and square, so it should be able to sustain them along with its high prices. This is a company, he says, that is not prepared to change. Google will do everything possible to maintain its monopoly. That would have everyone back before the Court litigating every single minutiae.

He walks through the ways in which DOJ’s remedies would restore competition. In the publisher ad server market, where all competition has been destroyed, this will be challenging. The Court is starting from scratch, and needs to foster the conditions for competing firms to enter the market. The remedies need to mitigate switching costs, and instill business confidence in the market. The DOJ’s proposals for AdX divestiture, open-sourcing the final auction logic, data portability, and escrow funds do exactly this. Even where Google’s proposals allegedly provide some of this, there are carve-outs - like limiting portability to data related to open-web display transactions. This alone is insufficient to enable publishers to switch, and thus for businesses to invest in building rival offerings.

Next, he turns to AdX divestiture. The barriers are limitations in access to Google Ads demand, the manipulation of DfP final auctions, and a lack of business confidence. The DOJ’s remedies of AdX divestiture and open-sourcing the final auction logic would resolve these issues by eliminating the self-preferencing incentive for Google Ads and DfP, and fostering business confidence. Structural remedies would provide the assurance that the DOJ’s witnesses across the adtech ecosystem indicated would be necessary to inspire confidence to switch, and to invest.

The notion of the industry’s lack of trust in Google was a hot topic during remedies trial. DOJ addresses it head-on.  Yes, it is true. After years under Google’s thumb, there is a trust deficit. But this is not subjective or idiosyncratic. Confidence in rules of fair play and competition on the merits are central to American markets. We see a slide showing witnesses from each side of the industry: publishers, buy and sell-side rivals, Prebid.org, and ad buyers all agreeing that only structural relief will be effective. The DOJ’s emphasis here is important following Mehta’s ruling in the Search case which critiqued a lack of sufficient evidence in the record from market participants.

Judge Brinkema raises her first question: one of the Court’s concerns about commercial reality is just that — commercial reality. The industry is rapidly changing, and time is of the essence. We don’t know what the market will look like in five years. It is unlikely, she says, given all of the private litigation Google is facing, that Google won’t appeal. While settling would otherwise seem possible, Google is an “impossible situation.” Whereas some of Google’s behavioral changes might go forward, irrespective of appeals, why would a more complex order involving a more “extreme” structural remedy be beneficial of in two years, the liability findings might be reversed?

DOJ explains that the timelines as Google has described them are deceiving. The government asked the Court to order closing date for AdX divestiture as being 15 months after the ruling. The benefits of the structural remedies on competition would begin to be seen on a comparable timeline to the 12-15 month timeline Google proposed for their behavioral remedies to be put in place.

Next, Huppert reminds the Court that absent divestiture, Google has the ability and incentive to manipulate its products and features in its favor in subtle but powerful ways.  Structural remedies are cleaner and less risky for this reason. It is not only difficult to conceive of all of these “roads” to the same monopoly (calling back to the opinion in International Salt), but to detect them. And, even once that has been done, it is also hard and costly to identify whether each may violate the Court’s order. He gives an example of latency. Google’s proposal specifies that when a publisher submits a bid request via another “qualifying” ad server or via prebid, it won’t intentionally introduce latency that does not otherwise exist as a technical reality. If Google causes AdX to send bids 400 milliseconds slower, was it intentional? Was it a “technical reality?” We’d be back in Court to figure that out - each and every time. Where Google says that its buying tools won’t prioritize bidding on AdX unless AdX provides more effective information, higher quality impressions, better fraud or privacy protections, the same is the case, despite these arguments already having been deemed pretextual by the Court. Google’s rational economic incentives will cause it to test every single boundary and “punctuation mark.”

Google has already shown its propensity for this. Throughout discovery, and even when it blew past the deadline to disclose trial witnesses with a pretextual excuse. Once Google gets the behavioral remedies it wants, it will test every boundary until the order is whittled down to a shadow of what it once was. Every day we hash out this behavior, he says, is a day the markets don’t get relief. And as is inscribed on the Courthouse entrance,  “justice delayed is justice denied.”

He closes by explaining that there is “more than one way to bake a cake” in terms of achieving the structural remedies. The DOJ is not dogmatic about this, as long as the same ends are met. And it cannot be said that divestiture is not possible since Google’s own internal analysis that it “fought tooth and nail” to try to keep out of the trial show that it is. He presents a chart demonstrating that the “business divestiture” Google contemplated is a divestiture. All of the elements that DOJ contemplates were included in Google’s analysis. DOJ would thus be happy for the Court to order the divestiture that Google has determined internally that it could do.

Huppert ends with a clear picture of exactly what is at stake here. If Google were to succeed in carving out an exemption from structural relief due to the technical complexity of divestiture, it would set a harmful and dangerous precedent. It is called big tech for a reason, he says. For antitrust laws to continue to protect markets, the Court cannot accept that complexity confers immunity.

Google Closing Arguments:

Karen Dunn delivers Google’s closing. She begins by addressing the Court’s earlier question on timeline. All of Google’s proposals can be done in 12 months, except for the Programmatic Guaranteed and Preferred Deal prebid integration, which would take 15. She then dives into attempting to sew doubt in the propriety and feasibility of the government’s structural remedies proposals. Dunn suggests the Plaintiffs are not familiar with the case law, or with Google’s proposals. She says that this is an unusual case where the parties are largely aligned on the behavioral remedies. (Of course, the DOJ has already shown that the devil is in the details, through their very effective latency example earlier on.) She also says it is unusual as all of the witnesses on both sides testified to the competitive benefits of the behavioral remedies. This logical fallacy whereby Google conflates what is beneficial with what is sufficient to restore competition recurs multiple times throughout the closing.

Nevertheless, on these grounds, she asks how the defendants can possibly justify their “overreach” of proposing structural remedies. Dunn then brings up this notion of “trust.” The government’s proposals don’t just indicate that they don’t trust Google, she suggests, but that they do not trust the judicial process either. There was no attempt made by Dunn to address the DOJ’s illustrative example on latency, which in my view made this argument fall flat.

Of course Google will be accountable to an injunction, because it is publicly traded, she suggests. By that logic, Google would not have undertaken anticompetitive conduct in the first place, because it would not run the risk of being caught knowingly harming its customers. Alas, here we are.

She spends the bulk of the closing running through a list of ways that DOJ has failed to meet the heavy burden necessary to justify divestiture, and accusing DOJ of making disingenuous use of ellipses in citing legal precedent. Divestiture isn’t tailored, for example, because it would cause the Court to interfere with lawful practices, like interfering in Google’s ability to transact in growing markets like CTV.  All of the publishers said that access to real-time bids from AdX would be a condition for switching, so Google’s proposals give publishers what they need to switch, thereby restoring competition. Again, this conflates what is necessary with what is sufficient.

It is interesting to me that in making these arguments, the reference to DOJ’s rival witnesses were relatively unemotional, yet once again, there were direct, personal attacks on the publisher witnesses. Stephanie Layser is known as the Google Slayer, Dunn says, as she again suggests that Layser doesn’t have recent publisher experience. Twice, she speaks to items that Daily Mail’s Matthew Wheatland “complained” about, the same wording used during his aggressive cross-examination at trial. I have to wonder if Google’s skilled attorneys do not realize how this comes off. It is truly incredible just how much Google appears to hate its own customers, and how comfortable it is to express that contempt publicly.

She discusses the testimony of Elizabeth Douglas, the late-add WikiHow CEO. She omits the part about WikiHow’s content licensing deal with Google as she describes how Douglas, unlike DOJ’s publisher witnesses, understands the struggles of small publishers.

Dunn attempted to say that DOJ never presented evidence of initial buyer interest for AdX. DOJ objects, and Judge Brinkema says that she remembers the evidence. Dunn points to the Google Ads transactions flowing increasingly through YouTube amidst the decline of open-web display. Again, it is not lost on me that these time periods coincide with Google’s aggressive push of its AI product, Performance Max, which has Google’s own algorithms decide how to “optimize” the ad budgets of its millions of advertisers. Nevertheless, Google doubles down on its thinly-veiled threat that in the event of AdX divestiture, Google may simply stop bidding on open-web display.

DOJ Rebuttal:

Julia Tarver Wood delivers the government’s rebuttal.  She says that while behavioral remedies would slowly erode monopoly power, they would not terminate it. Given that time is of the essence and the future of the market is dynamic and uncertain, as Judge Brinkema has noted, the DOJ is pursuing future-proof structural remedies that truly free the markets.

Judge Brinkema interjects, asking why DOJ has not identified a potential buyer. DOJ’s Wood says that it is inappropriate to come in here and prematurely identify a buyer before an order. Judge Brinkema adds that if Microsoft, for example, were to offer to buy AdX, it would create similar problems, and likely require antitrust review that could cause delay. This is challenging for the Court as the proposal is still at a fairly abstract level, whereas an Order must be more concrete.

DOJ points to 4th Circuit precedent to explain that this is common. There is nothing that precludes due diligence from occurring pending appeal. She then rebuts Google’s assertion that DOJ’s proposal didn’t address Google’s update proposals. This makes no sense, since they were each filed on the same day. She then rebuts the point Dunn raised suggesting Plaintiff witnesses were not supportive of divestiture.

Wood emphasizes that the testimony of Google’s witnesses alleging that divestiture would be too complex is inconsistent with Google’s own documents - a pattern this Court has seen before during liabilities. She turns to tackling the topic of AI, citing the Court’s own finding that given Google’s vast data, scale, and sophistication, AI developments will further benefit its open-web display advertising business. She explains that AI is hurting publishers. Google, who claims to be the steward of the open web put AI overviews at the top so nobody has to click through to the publisher’s website. That is the AI apocalypse WikiHow’s Douglas referred to.

Wood closes by asserting that genuinely it is hard to imagine a case more deserving of structural remedies than this one, where monopolies were derived not from a recidivist monopolist being “better,” but by cheating.

We are, she says, at a critical inflection for the open internet, and of antitrust enforcement, including for big tech. The world is watching. On behalf of the industry witnesses that came to testify against a monopolist, which took courage to do, she urges the Court to have the courage to treat the disease.

Google Surrebuttal

Dunn begins by suggesting that DOJ did not respond to any of the allegations that it has not met the heavy burden required for divestiture. She spends the surrebuttal arguing that the DOJ is relying on faulty legal grounds for their proposal - that there has never been divestiture ordered to remedy a tie, to remedy unlawful combination, in a complex, two-sided technology market, or that takes a competitor off of the field of competition.

She argues that Google’s search advertiser base was lawfully acquired in part, and thus DOJ has not sufficiently illustrated a significant causal connection in proposing divestiture, pointing to Judge Mehta’s Search ruling.

She returns to the idea of “trust,” saying that while DOJ says it is not simply about trust, what they cite suggests otherwise, overlooking the DOJ’s arguments about rational behavior and economic incentives. This premise is used to draw a reductive parallel to the Microsoft antitrust case. If it were sufficient to say that Microsoft “has proven untrustworthy” in the past, the divestiture order would have been upheld, she suggests.

Dunn ends by noting that someday, people will ask what this litigation will have achieved. Was a messy divestiture really necessary, she asks, when both parties agreed on behavioral solutions.

Conclusion

Judge Brinkema concludes by acknowledging the excellent advocacy on both sides, and expressing appreciation for the civility in such a high-stakes case. She says that she will need to decide on the “core issue” - namely, structural and behavioral, or behavioral only. Then, she will need to craft the order. Even on behavioral remedies, there are a lot of differences between the proposals, she notes, asking about whether the parties have come to an agreement on the Monitor mechanics. DOJ’s Wood says that they are very close, with one exception. One side submits a slate of candidates, and the other side selects among them. They have not yet agreed on which side submits versus selects.

What’s to come?

Now, we wait. Judge Brinkema indicated the decision will likely arrive in the new year. While that seems to suggest Q1, as we’ve seen from liabilities, it could take longer. From there, it sounds like there may be additional arguments before the court, not to mention the years-long appeals process that Judge Brinkema pragmatically acknowledged is nearly certain to follow.

🌅 So, think of today not as the end, but as a new beginning.

The DOJ’s somber reminder about the stakes of this trial are impossible to overstate. Weak remedies could serve to further embolden big tech, and cause a crisis of confidence in the ability of US antitrust law to protect the market from their abuses. Yet, that is only a part of a much bigger, more optimistic picture.

Irrespective of the outcome, we will continue to see companies forge ahead with private litigation for the harms they've suffered. Google will have to think twice about engaging in dodgy conduct, like rolling out self-serving features or making sneaky, subtle changes that its customers do not want and did not ask for. Now, all eyes are on Google - and that's not only a result of the brilliant work of the DOJ and the courage of the witnesses, but thanks to all of you outside of the courtroom that have read our newsletters and followed along.

That is already a huge win.

For tech accountability.

For publishers that have suffered under Google's reign.

For advertisers who were overcharged to boost Google's own margins.

For rivals who would have innovated had they had a fair shot.

For the future of journalism, which relies upon putting an end to big tech's unbridled extraction.

For democracy, which rests upon an informed public, and a free and independent press - each of which Google’s conduct has undermined.

And for all of us, who have had our shared reality manipulated and our autonomy undermined for commercial gain by a handful of companies with too much power.

Now, we are taking back control.

‍