Remedies Trial Update, September 22, 2025: Google's proposed remedies are like "putting a band-aid on a severed limb"

Day 1 of US v. Google remedies trial kicks off with opening arguments from both parties. DOJ’s Julia Tarver Wood opens the day with a Winston Churchill quote: “Those that fail to learn from history are doomed to repeat it.” She says this rings equally true today, where we are at a critical inflection point for digital advertising and the open web. Wood speaks to the calculated moves by Google, a recidivist monopolist, over the past 10 years to weaponize multiple interconnected and mutually-reinforcing monopolies. Why did Google do it? Power and incentive, she says. Unless structural measures are put in place to stop them, Google will continue to have both the motive and the means. She notes that this is particularly the case in the ad exchange market, where conduct is hidden in algorithms and code.

Wood speaks to the SCOTUS warning in International Salt against “toothless behavioral remedies” like those that Google has proposed, which she says would allow Google to travel new roads to the same monopoly.

She then speaks to DFP, where the “news just gets worse” - wordplay that is an allusion to Google’s 91% market share, and also the struggling newsrooms of its publisher customers. Wood describes the DOJ’s proposal for a phased divestiture of the final auction logic - which she refers to as the “source of power” - from DFP, and the contingent divestiture of what remains should the former be insufficient. Google’s remedies are, she suggests, “like putting a band-aid on a severed limb” because Google has “endless creativity” to engage in anti-competitive behavior. There’s an adorable demonstrative here of a game of whac-a-mole, where each mole represents a different anti-competitive antic - First Look, Last Look, UPR, and ???.

Wood unpacks the behavioral remedies proposals, which would prevent Google from creating the same DfP/AdX tie, and reminds the Court that an effective remedy would need to address adwords (Google Ads), which the Court has acknowledged as the “Golden Goose.” There’s another spiffy demonstrative using illustration of runners in a race, where DOJ demonstrates that Google’s remedies may be removing the obvious hurdles but does little to address Google’s massive headstart or the Google “ball and chain” slowing their rivals down - namely, their data disadvantage, and Google’s discriminatory and direct bidding.

Google, represented by Karen Dunn, centers its opening around two core arguments: (1) that the DOJ’s proposals are legally unfounded, and (2) that they are “destined to fail technically.” First, Google argues that DOJ’s proposals are an overreach, calling back to the recent Search ruling to remind the Court that “caution is key,” juxtaposing this against the DOJ’s “swings for the fences.” Dunn walks through how each of Google’s proposals addresses specific harms outlined during the liabilities phase, and would thus address publisher concerns. She cites to Plaintiff witness testimony - e.g. Stephanie Layser’s statements that “head-to-head competition” is what’s important to publishers, and a reference to a statement by Arnaud Creput that because prebid could act as an ad server, many publishers wouldn’t need GAM. We do not get to see these statements in context at this juncture, of course.

Dunn continues on the “legal error” line of argument, walking through why divestiture has never been ordered to remedy a tie, without an unlawful combination, in a complex, two-sided tech market, or where it takes the defendant off of the field of competition.

Dunn previews their “Google knows best” argument - that the DOJ’s remedies would pose a risk to consumer welfare; and speaks briefly to AI transforming adtech - again an attempt to harken back to the Search ruling. Then, we hear a more overt reference, quoting Judge Mehta on the need to approach remedies with “humility” in a rapidly changing tech market. Opening arguments end with Dunn stating that risk of consumer harm is enough to reject a remedy, and that Google has gone as far as it can in its proposals without breaking the tech.

Grant Whitmore, Advance Local

The first witness is a publisher - Advance Local’s Grant Whitmore. After re-orienting us on the impact of Google’s overlapping monopolies in the ad exchange and ad server markets, most of his direct examination delves into why he believes divestiture of AdX and DfP are necessary, and anything less would be insufficient to restore competition. He expressed his concerns with the DfP “black box” - that the final auction logic is “completely opaque” to publishers. Without this black box or the tie between AdX and DfP, AdX wouldn’t enjoy the preference it currently does.

We also hear about how difficult it is to do business with Google, generally. He explains that Google doesn’t accept any contract redlines, so his company ends up having no choice but to agree to things they never otherwise would agree to because they cannot lose access to AdX demand.

DOJ covers its bases on Google’s anticipated malware, privacy, and security arguments. Whitmore says these things are table-stakes and the company acquiring AdX would have a strong incentive to maintain standards. He isn’t worried about disruption during the migration, as adtech companies change hands all the time, and these transactions are typically seamless from the customer’s vantage point.

As for DfP, making the code open-source would be a “great start” though he thinks full divestiture is warranted. It would give publishers a better opportunity to see why bids are won, and mitigate the extent to which Google is making decisions for him influenced by their own competing interests. Again, no privacy, security, or malware concerns. He does feel that a true DfP spin out would create greater opportunity for new entrants, despite the DOJ’s more measured proposal of a phased auction logic divestiture and a contingent divestiture of the remainder of AdX.

Switching costs are not a concern, he says, because divestiture doesn’t mean that publishers themselves need to switch. If publishers do want to switch, he explains that while it is a bigger lift, it can be done - likening it to switching content management systems.

DOJ asks about the behavioral remedies proposals, which he mostly says are a good step but insufficient because “Google squeezes the balloon” and is adept at moving its advantage elsewhere. We hear about Google’s massive 1st party data advantage, and how the combination of data across its properties puts others at a massive disadvantage.

He unpacks the loophole in Google’s proposed remedies, which he says look good on the surface, but have minimal impact.  On Google’s proposal to not restrict real-time bids from being shared from AdX to a non-Google ad server, he explains that the part of his business that is growing isn’t open-web display advertising, but Private Marketplace deals (PMP). This is seemingly carved out from Google’s definition of Qualified Open-Web Display Advertising Inventory (QOWDA). These other deal types, like PMPs and Programmatic Guaranteed (PG) command higher CPMs, and DSPs and SSPs are bidding less on the open market. He also comments that Google is making a big push to activate more private deals in AdX. This private deal talk sets the stage for a very long and circular cross-examination by Bill Isaacson.

Judge Brinkema asks if AI will “handle” the types of negotiations done in these private auction, PMP and PG deals, and he explains it won’t as these deals still require participation of publishers.  The Judge does seem irritated during cross by the extent to which these concepts that weren’t extensively covered during the liabilities phase are in discussion here; but it does seem to become a bit more clear as the day progresses that this is a necessity of discussing the concept of divesting the ad server final auction logic.

The Judge also asks if the Qualified Open-Web Display Advertising Inventory didn’t have the limiting qualification to not apply to PG and similar deals, if that would alleviate his concerns. He says that it would not - Google could still put its thumb on the scale through its control of both the exchange and ad server.  On cross, Google tries to demonstrate that its proposals address his concerns by having AdX compete head-to-head with other exchanges to bid into DfP; but DOJ makes clear on redirect that this is only one buying path and nothing restricts Google from also having AdX bid directly into DfP. There was also a seemingly needless question on cross about his educational background which felt arrogant and superfluous, but alas I digress.

Andrew Casale, Index Exchange

Next up is Andrew Casale, who also testified during the liabilities phase, questioned by DOJ’s David Geiger. He speaks to why DfP is not a fair platform from his perspective, alluding to the black-box final auction logic, and the restriction of PG deals to AdX.  He says that AdX should be divested from Google, or a minimum separated from DfP and not owned by the same company. This would “unshackle” the ad server from AdX, putting it on a level playing field, competing in one auction with its rivals.

He speaks to Header Bidding, and how having AdX participate would increase competition, if AdX can’t separately also bid directly into DfP. He explains the benefits of open-source software, and that a neutral 3rd party overseeing open-source final auction logic would instill confidence. He references Prebid and IAB Tech Lab as two entities capable of “hosting” this open-source final auction logic.  The Judge asks if open-source software is more vulnerable to hackers, and he explains that it is not because just as the code is visible to bad actors, it is also exposed to the community which typically works to improve it for the benefit of all.

On conduct remedies, we once again get into the importance s of Google’s 1st party data, and how this data drives up the probability of higher CPMs. Restrictions on self-preferencing with this data, while important, don’t replace the need to divest AdX. We can’t know what new features Google might roll out tomorrow, he says, that would favor its product. He makes a reference to imaginary 3rd Look and 4th Look features, and the Judge smiles. He also calls Google’s proposal not to reinstate First and Last Look disingenuous; again pointing to the QOWDA carve-outs.

On cross-examination, Google works hard on impeachment on the basis of alleged inconsistencies with deposition testimony and prior statements; but to limited avail.  Dunn attempts to paint a picture that Casale changed his mind and started to favor divestiture after he spoke to the DOJ, and after he began contemplating acquisition by IndexExchange. Dunn asks if he realizes DOJ reserves the right to approve the acquiring entity, which appeared to be a subtle implication that he changed his opinion to curry favor. Paired with pointing to numerous podcast statements and general optimistic marketing puffery made much of cross feel like an abrasive and tedious reach. Casale won the day, explaining why his thinking on remedies has evolved over the past several months: “Generally speaking, I work through things in my head a lot.”