Google does not have a final judgment against it in the ad tech case. Not today, and not on Wednesday night when AdExchanger told the industry that the final judgment "takes effect in 60 days." What exists is a 106-page memorandum opinion from Judge Leonie Brinkema, unsealed with no redactions after neither side asked for any.

What does not exist yet is the document that actually binds Google: the Final Judgment, which the parties owe the court by October 2, 2026.

That distinction is not lawyer trivia. It is the entire timeline of the remedy, and the timeline is the story.

Allison Schiff is one of the best explainers in this trade. I read her occassionally. Most of you read her. When she writes a TL;DR, a lot of people in ad tech treat it as the version of events they will repeat in the next pitch meeting, the next board update, the next LinkedIn post about how "the Google remedy is finally here." Which is exactly why the compressions matter.

A TL;DR that gets the date wrong, the legal vocabulary wrong and the scope wrong does not just misinform. It becomes the industry's working memory.

So let's quote the claims and check them against the record.

Here is the sentence that closes her piece: the final judgment "applies globally and takes effect in 60 days."

There is no final judgment.

On September 2, Brinkema told the parties to draft one within 30 days. That draft is due October 2. The court said the eventual judgment becomes effective 60 days after it is entered, not 60 days after the opinion was filed and not 60 days after it was unsealed.

Do the math the way a publisher's CFO will. If the parties file on October 2, and the court enters a judgment with no fighting over the language (lol), you are looking at effectiveness sometime in December at the earliest. If Google and the Justice Department spend a few weeks brawling over definitions, and they will, because the definitions are where the money lives, the clock slides into 2027 before it starts.

Then the real clock starts.

Reporting on the opinion puts roughly 12 months on AdX access for rival publisher ad servers and 12 to 15 months on the Prebid connections.

Stack those on top of the drafting and entry period and the honest answer to "when does a publisher actually get to route AdX demand through Prebid on equal terms" is somewhere around 2028.

"Takes effect in 60 days" makes it sound like publishers should be calling their SSP reps before Thanksgiving. The accurate version is that nothing operational changes this year. Google keeps running DFP and AdX exactly as it runs them today, and it keeps banking the margin while engineers build APIs on a court-supervised schedule.

If you are an independent SSP that spent the last eighteen months fundraising on the idea that Google was about to be pried open, the difference between 60 days and two-plus years is the difference between a growth story and a bridge round.

Schiff's refresher says the court found Google "guilty" of operating a monopoly.

This was a civil case. The Justice Department sued Google in January 2023 under the Sherman Act as a civil plaintiff. Brinkema found Google liable for monopolizing the publisher ad server and ad exchange markets for open-web display and for unlawfully tying DFP to AdX. Nobody was charged, nobody was convicted, nobody was sentenced. There is no guilt in the legal sense because there was no prosecution.

Is this a pedantic point? In a bar, yes. In a trade publication that lawyers, compliance officers and investor relations people read, no. The word "guilty" imports a set of consequences that do not exist here: criminal exposure, a verdict, a conviction record. The actual exposure is different and in some ways more annoying for Google, which is six years of court supervision with contempt power behind it.

It also matters because the industry is about to spend two years arguing over whether Google is complying. That argument happens in the language of civil decrees: parity, functional equivalence, reasonable access, dispute resolution. If the trade press starts from "guilty," the readers start from a morality play. The fight ahead is an engineering audit, not a sentencing.

This is the paragraph everyone is going to screenshot:

Schiff calls it irony. Prebid was built to let publishers get competing bids outside Google's control, Google refused to play with it for years, and "now Google is being ordered to join."

It is a great line. It is also a little too neat.

Google put Prebid interoperability on the table itself during the remedies phase. Google's own counterproposal was a behavioral package built around opening AdX to rival ad servers and to Prebid, killing First Look and Last Look, and not bringing back Unified Pricing Rules. Google's whole strategy was to offer the interoperability early so the court would have a reason to say no to the divestiture. It worked.

So the framing of Google being dragged kicking and screaming into Prebid misses what actually happened in that courtroom. Google chose which rope to be tied with. The court took Google's rope and tightened it: a worldwide scope, a six-year term, a Monitor, a Technical Committee, a documentation mandate for the DFP auction, prohibitions on AdWords bidding directly into DFP. Those additions are real and Google did not want them. But the headline concession, the Prebid integration, was the price Google was willing to pay to keep AdX.

And Schiff's own refresher gets halfway there. She writes that Google argued targeted behavioral commitments were enough, and then: "And Judge Brinkema agreed."

She agreed there would be no breakup. She did not agree Google's commitments were enough as offered. That is why there is a monitor with source code access and a six-year term. Collapsing those two things into "agreed" hands Google a cleaner win than the opinion actually gave it, which is a funny place for the trade press to land on the day the document finally became public.

The real irony is not that Google has to join Prebid. The real irony is that Prebid, a workaround built to escape Google, is now the load-bearing wall of a federal antitrust remedy against Google, and nobody asked Prebid whether it wanted the job. More on that behind the wall.

Schiff: "all of these requirements apply globally, not just in the US."

Half right, and the wrong half is the one this newsletter cares about.

The decree is worldwide in reach. Brinkema found worldwide markets and decided uniform global implementation was easier to administer than a US-only version. That part is accurate.

But the substance is not universal across Google's advertising business. It is confined to the markets found at liability: publisher ad servers and ad exchanges for open-web display. Reporting on the parties' language shows exclusions or special treatment for video, CTV and in-app inventory.

Read that list again. Video. CTV. In-app. Those are the three places the money is moving. Open-web display is the part of the business every holding company planning deck describes as flat, shrinking or "under pressure." The remedy is global the way a Blockbuster membership was global. It reaches everywhere the growth isn't.

"All of these requirements apply globally" tells a CTV publisher, a streaming app, or a mobile game developer that they just gained leverage over Google. They may have gained nothing. And it tells Google's compliance team exactly where the next argument goes: format classification. Is an outstream video unit on a news site "display"? Is a web-based CTV app "open web"? Is a rewarded video ad in a mobile web game "in-app"?

Every one of those questions is a place where a product can be reclassified, a format can migrate, and a remedy can quietly stop applying. The scope line is the most important sentence in the decree and the TL;DR turned it into a victory lap.

This piece compares AdExchanger's September 16 explainer against the unsealed Brinkema remedies opinion as described in the Justice Department's public account and independent reporting, including reporting on the parties' draft judgment language. The Final Judgment had not been filed at publication, so definitions, carve-outs and deadlines may change; the 12- and 15-month implementation figures are reported estimates, not court-ordered dates, and the video, CTV and in-app treatment rests on reported draft language. AdExchanger competes with ADOTAT for readers and I read and respect Schiff's work; I competed directly against executives at several of the SSPs named below during my years running Vizi. ADOTAT carries no paid placement in editorial and is funded by subscribers. If AdExchanger or Google sends a response, it will run here in full.

Before I keep swinging, let's be fair, because the piece deserves it.

Schiff got the center of gravity right. The remedy lives in four places: AdX and DFP interoperability with Prebid, real-time AdX bids to rival publisher ad servers, publisher access to winning and losing bid data, and restrictions on AdWords. That is the spine of the opinion and she found it in a few hundred words on a deadline.

She flagged "functionally equivalent," which is the phrase that will decide whether any of this works. A nominal API that is slower, sends thinner signals or responds less often than Google's internal pipe does not satisfy the remedy. The reported implementation language bars added latency, reduced response frequency and reduced information transmission. That is the single most important operational standard in the case and she put it in front of readers.

She got DV360 right, and even flagged that some people consider it a big miss.

She got the rejection of open-sourcing DFP right, and correctly noted that Google instead has to document how DFP picks a winner.

So this is not a hit piece on a reporter. It is a correction of a genre. The TL;DR format rewards the clean sentence, and antitrust decrees are made entirely of dirty ones. Every place Schiff compressed, the compression pointed in the same direction: toward a remedy that is sooner, broader and more muscular than the one Brinkema actually wrote.

The piece says a court-appointed "technical monitor" will keep watch for six years, with "full access to Google's employees, systems and source code."

The actual structure is bigger, more layered, and more breakable than that.

There is a Monitor. There is a three-member Technical Committee that advises the Monitor. There is an Internal Compliance Officer inside Google. There are periodic reports. And the government keeps its own inspection rights. The Monitor can inspect documents, data, systems, facilities, source code and algorithms, but that access is governed by reasonableness and confidentiality procedures.

"Full access" and "reasonable access under confidentiality procedures" are not the same sentence. The second one is the one Google's lawyers will be quoting in every dispute for six years. Reasonable according to whom? Confidential from whom? Can the Technical Committee share what it finds with the publishers who are supposedly benefiting? Can the Monitor bring in outside engineers? How fast?

We have seen this movie. The Microsoft interoperability decree turned into years of technical supervision over what counted as sufficient documentation, what counted as parity, and whether compatibility had to be maintained as products changed. Monitoring an interface is not a box-check. It is a permanent argument with a very well-funded counterparty.

And six years is not a fixed term. It is a floor Google can live with. The court rejected DOJ's 15-year ask, but it can extend the decree if Google hasn't fully satisfied it. That cuts both ways. A slow build could, in theory, earn a longer leash. It also means the six-year number in every headline is a guess about compliance, not a date on a calendar.

The reason this matters for publishers: when your Prebid integration comes in 40 milliseconds slower than the internal DFP path, you are not calling a "technical monitor." You are entering a multi-party dispute process involving a Monitor, a committee, a compliance officer and two sets of lawyers. Nobody on the publisher side should budget for that being fast.

The TL;DR says Google has to "share bid data – wins and losses – with publishers."

True. And that undersells the most commercially interesting part of the remedy by about half.

The package is not just AdX win and loss data. It includes exportable DFP configuration data, historical DFP data, export mechanisms, import documentation and technical support. Put plainly: a publisher can take its line items, its setup, its history and its yield logic out of Google and hand it to someone else.

That is not a reporting feature. That is a switching-cost remedy.

For fifteen years the reason nobody left DFP was not that DFP was beloved. It was that leaving meant rebuilding years of configuration by hand, losing historical data your yield team used to set floors, and walking away from AdX demand. The Prebid and rival-ad-server integrations address the demand problem. The data export addresses the "we'd have to rebuild everything" problem. Together they are the court's answer to lock-in.

Framing it as "Google has to share bid data" makes it sound like a transparency dashboard. It is actually a moving truck. And the companies who should be reading that section closely are not publishers. They are the rival ad servers, the analytics vendors and the yield optimization shops that can now build products on exported Google data.

That section is where new businesses get built. It got one clause.

Schiff notes DV360 faces no restrictions because the government never tied it tightly enough to the core conduct, and adds that depending on who you ask, that is a pretty big miss.

She is right, and I would go further. "Depending on who you ask" is doing a lot of work for a piece that also said there was "nothing surprising here."

Here is why it matters. The court accepted that AdWords was instrumental to the unlawful tie: AdWords demand flowing through AdX was the leverage that made DFP sticky. So AdWords got restrictions. But DV360, Google's DSP for agencies and big brands, is the buying tool that holding company trading desks actually use. The government's failure was evidentiary. It did not prove DV360's relationship to the AdX and DFP conduct well enough at liability, so the court would not reach it at remedies.

Translation: the self-serve small advertiser side of Google is fenced. The agency side of Google is not.

For anyone who covers holding company economics, that is not a footnote. Principal media, rebate structures and platform incentives run through DSP relationships. The part of Google's buy side that sits inside agency trading desks walked out of this remedy with no ownership-based restrictions at all. That is not a "depending on who you ask" item. That is a structural gap with a name.

The piece says AdWords is prohibited from "favoring Google's own ad tech tools."

Correct as far as it goes. What the TL;DR skipped is the exception that tells you how this will actually play out.

The prohibition is on ownership-based preference: AdWords can't favor a Google tool because Google owns it. But AdWords is still allowed to optimize for advertiser ROI. And the remedy does not compel AdWords to distribute equal volume through every exchange. The reported language preserves a direct integration path through Prebid.

Now imagine you are Google in 2028. AdWords routes the overwhelming majority of its open-web spend through AdX. A rival exchange complains. Google's answer is not "because we own AdX." Google's answer is "because AdX delivered better ROI for the advertiser," with a model, a dashboard and a Technical Committee presentation to prove it.

That is compliant. It may even be true. And there is no way for an outside exchange to prove otherwise without the kind of data and code access the decree limits to the Monitor.

Meanwhile the bigger trend doesn't care about the decree at all. AdWords demand has been drifting away from the open web toward Search, YouTube and Google's owned surfaces for years, for reasons Google frames as performance, user behavior and advertiser choice. The remedy does not reverse that. A publisher can win full, functionally equivalent Prebid access to AdX demand and still watch that demand shrink.

"Can't favor its own tools" is the rule. "Can optimize for ROI" is the practice. One of those made the TL;DR.

This is the section I'd have led with if I were writing a TL;DR, because it's what the court didn't order, and the missing pieces are where Google has room.

The anticircumvention language is mostly gone. DOJ asked for broad prohibitions on retaliation, interference, and any conduct designed to recreate the effect of the unlawful behavior. Brinkema rejected most of it as too vague. What survives is narrower: product-specific bans on the known tactics (First Look, Last Look, Unified Pricing Rules, direct AdWords bidding into DFP) and an employee anti-reprisal policy.

That means the decree is strongest against the last war. It bans the specific mechanisms the court already found. It is much less certain against a new mechanism that produces a similar result: a latency tweak, a signal-quality downgrade, an eligibility rule, a throttle, a reclassification of a format out of scope. Every critic who warned that interoperability only matters if someone can detect subtle degradation was pointing right here.

The escrow is gone too. DOJ proposed putting 50% of certain AdX and DFP net revenue into escrow, partly to fund publisher switching and to pay for administering the remedy. Rejected. So the rival ad servers who are supposed to benefit from all this interoperability have to fund their own onboarding, migration, support and publisher incentives.

The data export lowers the technical cost of switching. Nothing in the decree lowers the commercial cost. A mid-size publisher with two ad ops people is not moving off DFP because an API now exists. It moves if somebody pays for the migration and guarantees the yield won't dip for a quarter. DOJ tried to create that money. The court said no.

Neither of those rejections appears in the AdExchanger piece. Neither did DOJ's loss on disgorgement or the 15-year term, beyond a passing mention of the term. A TL;DR that lists everything Google has to do and nothing Google escaped is not a summary. It is a press release with better jokes.