Judge Brinkema published her ruling on the remedies in the Google ad tech anti trust case yesterday, and the doomers are out in force. There are lots of places to read about this including https://www.bloomberg.com/news/articles/2026-09-02/google-avoids-ad-exchange-sale-as-judge-orders-tech-integration and
– but the long and short of it is that everyone is all upset that she didn’t advocate for any of the structural remedies where Google would have to divest business units. She instead leaned heavily on the behavioral remedies, where business units would have to change their behavior.
This is broadly being written about as big tech flexing its muscles against an ineffectual government, with a sprinkling of accusing the eminently capable and frankly impressive DOJ of going for an overreach resulting in an under delivery.
I don’t really like either of these readings. This is because while I sympathize with the difficulty of the behavioral remedies to rein in Google’s behavior, the market is moving and Google has already taken steps to obviate the dependency that created the tying (tying meaning using one business line, in this case Google Ad Exchange, to force people to use another business line, in this case Google Ad Manager / DFP). The structural remedy for this was to make Google basically get rid of both businesses – divest AdX, and open source some logic in DFP. The behavioral remedy was to undo the tie – offer AdX to people outside of DFP.
If we are to take the tying argument at face value, the behavioral remedy itself is sound. I am on the record saying as much. The only caveat I’d add, which converted me to the divest camp late in the game, is that there are myriad ways for Google to preference AdX bidding into GAM vs other ad servers – both legitimate and illegitimate – and given the nature of ad tech, it would be basically impossible to discern if the preference were real or fake.
But I don’t worry about that anymore. The reason I don’t worry about that anymore is because this whole boondoggle has taken long enough that Google has already architected its way out of it, and meanwhile, the whole industry is moving towards a new paradigm (double whammy for why AdX divestiture doesn’t matter).
Whammy 1 - New Industry Architecture
The role of an adserver is kinda basic – nauseatingly simplified, it is a place for publishers to manage advertising campaigns that run on their websites without having to change the code on the website every time you want a new campaign to serve. If you have 5 campaigns, you log into the adserver, put in 5 ads for the different advertisers, and the adserver chooses where they go.
One of the important parts about adserving, as we’ve learned from this case, is that the adserver is the last point of call for the vast majority of web publishers’ advertising activity. In the act of being the last point of call, it becomes the point where data needs to centralize. This has created a huge advantage for Google, which they then abused to manipulate bids blah blah blah. Once they got in there, they tended to stay there, because ad server migrations rank up there with “moving” and “hemorrhoids” as some of the worst things human beings have to deal with.
But this is organically changing with or without this ruling.
The way that it’s organically changing is that the act of putting those ads into the adserver is being automated by agents. The promise of this form of capital A Agentic, automated ad operations, is that a human doesn’t need to do these things anymore. In an agentic world, the stickiness of the adserver as a centralization point is massively reduced because operations in it are automated, and therefore moving operations out of it can be automated.
In an agentic first world the friction of moving adserving out of GAM and into an agentic first framework is massively reduced when compared to a pre-agentic world. I am optimistic that a Prebid / Open Source adserving environment, with AdX bidding into it, is a future that could come into being. It means the behavioral remedy is sufficient (with some reporting disclosure caveats), and when I’m wearing this hat, I feel very optimistic.
But then comes Whammy #2.
Whammy 2 - Google Unleashes their GAM PMs
In my time in the industry, Google PMs (product managers) have generally been mythical creatures, sequestered away in an ivory tower while Google sent semi-well informed account managers to communicate with us plebeians. This all started to change in 2019, when all of the sudden I started to see GAM PMs out in the wild, talking to people, asking questions about how the industry works, and generally engaging with us in a way that I (and lots of other people) had never seen Google doing before.
Oh, and by the way, and this sticks with me, these were wildly competent SMEs (subject matter experts). I’ve spoken to product people at lots of companies. These people were radically competent. AKA, these were smart product managers.
My read on this is that Google knew that GAM being an add-on to AdX, which it was for 90% of GAM’s customers, was going to come into question and wasn’t sustainable. Their PMs had been in their golden dungeon in mountain view and had missed header bidding, and it was time to fix that and to let them build new shit in GAM.
And what we’ve seen in the 7 years since then, is that GAM is adding new features at a breakneck pace while AdX seems to be the thing that’s stagnating (other than a half hearted attempt at curation, which the product person in me can explain in unkind words). Core to these new GAM features is that GAM is becoming a source of revenue unto itself, and this single handedly breaks every remedy we have.
You’ve seen me writing about Buyer Direct, in fact I became the de-facto buyer direct guy somehow at some point (I’m still waiting on at least a dinner Google). I deeply believe that the significance of Buyer Direct cannot be overstated – this is because Google inhabits a unique position, where now that they have the market share that GAM does, if they can turn GAM into an advertising platform unto itself, they can hold on to that position by making it untenable for publishers to move their activity outside of GAM. And this is the evil genius of buyer direct – Ad Exchanges can be abstracted, they can be forced to bid into other systems. A native tool for automating your direct sales? That has to function in your central brain, that has to be at the middle of your business. It is in fact more important than even AdX if it works. And if Google can get it off the ground prior to the behavioral remedies discussed above, I think it will seal the deal and we won’t see publishers migrating off of GAM. This is fucking nuts. It means the entire antitrust lawsuit might as well have been archaeological, and that all of the proposed remedies are basically for business models that existed in a prior iteration of the tech.
The behavioral remedies do absolutely nothing if GAM uses its market position to make itself indispensable to publishers as a source of liquidity.
And it looks to me like Google could make it happen. Because they happen to really fucking know what they’re doing over there, who knew?
Conclusion
So yea. I think you’re all outraged for the wrong reasons, the only remedy that could’ve solved this problem was full divestiture of GAM, and frankly divesting either GAM or AdX is so potentially disruptive to publisher’s businesses that I’m not sure the juice is worth the squeeze. Publishers don’t need a 25% revenue drop while we “figure this out.”
Our options here are simple – we settle in with Google as our adserving overlords, or another standard with similar penetration becomes the way that publishers prefer to manage their direct sales. We’ll need good collective reporting on AdX performance to make sure it behaves as well outside of GAM as inside, but we have a lot of vendors that can monitor for that. I don’t think the situation is totally dire – competing on the direct sales battlefield is vastly superior to competing on the network demand battlefield (aka, competing against AdWords, which is like competing against a god). But certainly the unleashed Google GAM PMs are not going to go quietly into the good night.
