S2E15: The Open Exchange Myth

Full analysis

Judge Leonie Brinkema ordered Google to open up, not break up. Greg MacDonald and Geoff Wolinetz of "Leadership In" spent the episode asking whether that changes anything. Their answer, roughly: no. And the concentration math they cite backs them up more than the remedy does.

Here's the frame. Google has to make its ad exchange (AdX, where publisher ad space gets auctioned) work with rival publisher ad servers, build the Prebid integration it dodged for years, share more bid data, and live with a six-year compliance monitor. It keeps the ad server. It keeps the exchange. It keeps the money. Google will appeal, which pushes any real enforcement out by years.

This is hard to undo in one direction only. The ruling is on the books, but the remedy is soft and the timeline is slippery. Nothing here forces an operator to move this quarter. What's actually being decided isn't in the courtroom. It's whether publishers and agencies keep pouring money into open-web plumbing on the theory that it fights concentration, when the concentration numbers are going the wrong way.

The Market Analyst Look at the two numbers Greg MacDonald put on the table. Google, Meta, and Amazon took 53 cents of every US ad dollar in 2024 and 56 cents in 2025. That's the whole story of the last decade of "democratization" in one data point: the open web has been losing the share war while everyone congratulated themselves on header bidding. For a general reader: the three biggest platforms are eating more of the pie every year. The remedy does nothing to that trend, because interoperability doesn't move budgets, it moves plumbing. And EMARKETER's Nate Elliott pegs AI advertising at $32 billion in the US this year, almost all of which lands inside those same three. The direction is set.

The Skeptic Steelman the hosts and you land where Geoff Wolinetz did: if the punishment for illegally tying your ad server to your exchange is "play nicer in the sandbox at no cost," what stops the next player from doing exactly that? For the outsider: the court found Google cheated, then let it keep the winnings. But steelman the remedy too. Prebid integration and bid-data sharing are real, and independent SSPs have complained for a decade that they bid blind against Google. If the monitor has teeth, the information gap narrows. The catch is the appeal. A six-year monitor that starts in 2029 monitors a market that no longer exists. AI-driven buying will have redrawn the field by then.

The Operator Tuesday morning, nothing changes. That's where the hosts landed, and they're right. You cannot rebuild your supply paths around a remedy that Google is appealing and that may not bind for years. What breaks first if you try is your own roadmap: you'd be re-plumbing for AdX interoperability that could get stayed, reversed, or overtaken. The 90-day second-order effect is subtler. The 53-to-56 number gives your CFO a live question about why open-web integration spend keeps rising while its share keeps falling. That conversation happens whether or not the remedy ever bites.

The Customer / End User Take the small advertiser, because that's who the concentration actually traps. McDonald and Wallanets are right that Google Search and Meta together are roughly 35 of those 56 cents, and an SMB has no real substitute. Walmart's Vibe self-serve CTV platform and Universal Ads are not where a plumber buys leads next quarter. For the non-specialist: the corner business runs on Google and Meta because nothing else reaches customers at that price and simplicity. Interoperability between publisher ad servers is a large-publisher and SSP concern. It never touches the SMB. So the remedy leaves the most locked-in buyer exactly where they were.

The CFO Run the payback question on open-web infrastructure. You've spent years funding Prebid, header bidding, and supply-path optimization on the premise that they claw share back from the walled gardens. The share went from 53 to 56 the other way. That doesn't mean the spend was wasted. It means the justification has to change from "we're reversing concentration" to "we're protecting margin on the inventory we still control." Those are different budgets with different sizes. And the IAB lifting its 2025 growth forecast from about 9% to about 12% is a tailwind you shouldn't misread. More dollars in the market is not more dollars for you if the incremental spend flows to the same three names.

Where they part ways

The remedy fight. The Skeptic sees a real if delayed narrowing of Google's information advantage through Prebid and bid-data sharing. The Operator and Market Analyst see a monitor that starts too late to matter against AI-reshaped buying. Both can be right: the remedy works on paper and arrives after the market has moved on.

The concentration read. The Market Analyst treats 53-to-56 as proof the open-web thesis is dead. The CFO says it reframes the spend rather than killing it. Publishers don't fund SSP relationships only to fight Google. They fund them to keep control of their own yield.

What this actually hinges on

Two beliefs. First: does interoperability move budget, or just plumbing? Everything the hosts argue says plumbing, and budget follows audience and outcomes, both of which sit inside the walled gardens. Second: does the remedy arrive before AI buying rewrites the auction? On the timeline Wallanets described, no. The council leans hard toward "this changes little for years." Before anyone re-architects supply paths for AdX interoperability, verify the appeal status and whether the monitor's obligations are stayed pending appeal. Right now they most likely are not binding anything.

One caveat on the numbers. The 53% and 56% figures are McDonald's own pull, unsourced in the episode, and the exact methodology is unverified. The direction matches every published estimate. Treat the trend as solid and the decimals as approximate.

Prediction: The combined US ad-revenue share of Google, Meta, and Amazon will be flat or higher in 2026 versus the ~56% cited for 2025, and the Brinkema remedy will not be enforced against Google before the 2027 upfront selling season concludes.

Confidence: High — the appeal path defers enforcement and budget follows audience, not plumbing.

Confidence: High — the appeal buys Google years, and budget follows audience reach, not courtroom plumbing wins.

Why: The concentration figure moved from 53 to 56 cents in a single year while open-web infrastructure spending rose, which shows interoperability plumbing does not redirect budget. Advertisers buy the walled gardens for reach and measurable outcomes that live inside those platforms, and the remedy touches neither. Google has already announced it will appeal to the Fourth Circuit, and an appeal of a landmark antitrust remedy routinely delays enforcement by years, so no operator will see a functionally different AdX before mid-2027. The opposite outcome, share falling because of court-ordered openness, would require budget to chase publisher-ad-server integrations that most buyers never see, which is not how spend moves.

Revisit by 2027-09-30: We're right if the three-platform share of US ad spend for 2026 is reported at or above ~56% and Google's AdX interoperability obligations remain unenforced or stayed through the 2027 upfronts. We're wrong if that share drops meaningfully below 56% or the remedy takes binding effect on Google before the 2027 upfront season ends.

The plainer takeaway for operators: fund your open-web stack to defend the yield on inventory you actually own, not on the promise that a court just handed the open exchange a comeback. It didn't.