He's not scared.

He's totally fine.

The stock can be down 60% from its highs, the agencies can be quietly walking out the back door, the CTO position can be sitting empty like a haunted house nobody will move into, and Jeff Green will get on a stage in front of a thousand adoring ad tech believers and tell you, with a straight face and a confident blazer, that this is actually the best moment in the history of The Trade Desk. He's not scared. He just needs you to know that. Repeatedly. On friendly stages. With no follow-up questions. While mentioning every single competitor eating his lunch by name.

Look. I know what you're expecting. You want me to roast Ari. I'm not going to do it, and not because I'm soft. It's because I'd be factually wrong, and I don't do that.

Ari Paparo is legitimately one of the better interviewers this industry has ever produced. He built Marketecture into something real, something with some actual editorial teeth, which in ad tech media is roughly as common as a unicorn that also does your taxes. He knows the programmatic stack the way a chef knows a kitchen. He preps. He listens. He follows up. He pushes.

I want to be clear about something because I'm consistent about it. I'm not here to assassinate characters. Nothing in this piece is personal. This is about power structures. This is about what would actually make this industry function better. That's it. That's the whole game.

What Jeff Green did across from one of the sharpest interviewers in the business was deliver one of the most technically impressive performances of confidence over substance this industry has seen since the last time someone tried to explain header bidding as a consumer benefit. Ari did his job beautifully. Jeff Green is just very, very good at his job too. And his job, on this particular afternoon, was not answering questions.

The very first thing Jeff Green wanted to talk about was money. His money. Specifically, the "biggest purchase of my life," his words, an insider stock buy he revealed was the third largest insider purchase ever recorded. Not fourth. Not top ten. Third. He knew this ranking. He had looked it up. He mentioned it on stage the way someone mentions something they have been mentally rehearsing since the Uber ride over.

"Just a commentary on my convictions in our future," he said.

This is a very specific kind of rich person theater. You spend $150 million of your own money and you describe it as a footnote. A casual commentary. Like you're leaving a two-star Yelp review on a mediocre sushi place, not making the single largest financial commitment of your entire life.

Here is what that actually signals. When a CEO looks at the gap between what he's been saying publicly and what the market has been pricing in, and his solution is to personally spend $150 million closing that gap himself, that is not relaxed confidence. That is a man screaming into a void with his wallet. Which, to be fair, is a power move. But let's call it what it is.

Let's talk about the venue, because the venue is the message.

Jeff Green in 2025 and 2026 does not do hostile stages. He does not sit across from someone who is going to open with "your stock is down two thirds, what went wrong." He does not take questions from journalists covering the agency tensions, the missing CTO, or the gap between the AI narrative and what Kokai actually does under the hood.

What Jeff Green does is find rooms that already love him and then perform certainty for those rooms.

Marketecture is genuinely a good room. It's full of people who have spent years wanting The Trade Desk to succeed because the alternative, a world where Google and Meta and Amazon own every layer of the ad stack, is a world where everyone in that room has less power, less margin, and less reason to exist. So when Green walks out, he gets the warmth of people who are rooting for his story to be true. He is not just a CEO to this crowd.

He is a symbol. He is the guy who is supposed to win so that the open internet means something.

That is a lot of emotional weight to put on one man's quarterly earnings. But here we are.

And again, to be clear: Ari Paparo did not let him off easy. If this was planned, I was fooled. Because it seemed from the outside that Ari decided to go off script.

He pushed on Open Path. He named the AdWeek article and watched Green's face when he did it.

He identified the Claude beta as news in real time. He asked about Amazon, the Google ruling, Ventura, transparency versus outcomes. He did more actual journalism in forty minutes than most trade press does in a quarter. I respect it genuinely.

But Jeff Green is a black belt in performing candor without actually being candid. He has a tell. Once you see it you cannot unsee it. It is the warm smile that arrives just before the answer that is not an answer.

‘The "I can't comment on deals we haven't announced" delivered with the body language of a man who just announced the deal. The "I'm not sure if I was supposed to say that" from someone who has been in business for two decades and has never once accidentally said anything.

This is a very good act. It is a practiced act.

There is a useful diagnostic for understanding what a CEO is actually afraid of. Listen to what they bring up without being asked. Not the answers to hard questions. The volunteered information. The unprompted tangents.

By this measure, Jeff Green's brain in March 2026 looks like a whiteboard covered in red marker.

Amazon got mentioned so many times and with such elaborate analytical scaffolding that it stopped feeling like industry commentary and started feeling like a coping mechanism. Green argued at length that Amazon's DSP might not exist in five years. "I do want to be clear, I'm not saying Amazon isn't doing well in ads," he said. Which is precisely the sentence you say when you have been saying exactly that and someone has noticed.

Google got the full treatment. Sundar Pichai should be scared. The antitrust troubles are a case study. When Google leaves the open web it will create "the biggest opportunity we have ever seen." Green has been saying a version of this for years. The difference now is that Google's problems are real and documented. So he is not wrong. He is just perhaps overinvested in the specific mechanism of his salvation.

OpenAI and the AI companies got folded into an argument about how their massive capex will eventually drive them into advertising, at which point The Trade Desk will be there, neutral and ready, as the infrastructure through which those ads run. This is Green's big new bet. We will get to why it might be the most interesting and also the most dangerous idea he has ever had.

Paparo brought up the AdWeek article. Dentsu and WPP moving away from Open Path, citing transparency concerns.

Green's first move was the non-denial denial. "I won't comment on those two specifically." Which in the language of corporate communications translates directly to: it is true and I would prefer not to say so in front of a thousand people.

But then, instead of doing the smart thing and moving on, Green kept going.

And what he said next was breathtaking.

He accused the agencies of lacking transparency.

Yes. The agencies. His customers. The buyers. The people whose media budgets are the entire reason The Trade Desk exists as a public company. Dentsu and WPP are leaving because they cannot handle the blinding light of Jeff Green's transparency.

His actual words: "When you shine a light of transparency and if you have a way of monetizing that isn't conducive to transparency, you might not like the product."

Pause and appreciate the architecture of that sentence. They didn't leave because Open Path competes with their own supply chain economics. They didn't leave because the integration is difficult. They left because they are the shady ones.

The customer is the problem.

This is the kind of answer that only survives in a room that already believes you. In any other context the follow-up writes itself. "Jeff, your largest holding company customers are publicly walking away from a core strategic product and your explanation is that they personally lack the moral character to handle your standards?" Nobody asked. The room was warm. The moment passed.

The agencies noticed. They always notice.

Here is where the actual news happened, and Ari Paparo found it the way a good journalist finds things, by asking the obvious question nobody else had asked directly.

"Can someone go into Claude and create a campaign in the Trade Desk?"

Green's answer: "If you're a part of our closed beta? Yes."

Beat. Smile. "I'm not sure if I was supposed to say that."

He was supposed to say that. He absolutely was supposed to say that. This was the headline he came to make, delivered in the packaging of an accidental admission because accidental admissions feel more credible than press releases, and Jeff Green knows this better than almost anyone alive.

But here is what that headline actually means when you unwrap it.

The centerpiece of The Trade Desk's AI narrative, the thing Jeff Green "accidentally" revealed, is a beta integration with Anthropic's Claude. Not Trade Desk's Claude. Not a proprietary model built on their unique bidstream data. Not something they own, control, or can prevent anyone else from also having.

Any DSP can license Claude tomorrow. Google can use Claude. Amazon can use Claude. The startup that three former Trade Desk engineers launched last year in a WeWork in Austin can use Claude. There is no exclusivity here. There is no moat. There is a subscription dressed up as a strategic transformation, announced with the theater of an accidental disclosure.

And then there is the Anthropic problem, which Green himself described with perfect clarity while meaning to describe something else entirely.

"All of the AI companies have a serious problem," he said, "which is that their capex is massive and they have really high expectations on the way that they're going to make money."

He said this as a reason those companies will come to The Trade Desk for help.

Read it again as a description of Anthropic. Because it is also a description of Anthropic.

Anthropic has massive capex. Anthropic has enormous pressure to generate revenue at scale. And The Trade Desk has just handed Anthropic a live, operational demonstration of exactly how programmatic advertising works inside an LLM interface. With real campaigns. With real clients. With real money flowing through it.

Anthropic is now sitting in a classroom watching The Trade Desk teach the class, taking very detailed notes, next to a whiteboard that says "total addressable market: one trillion dollars."

That is not a partnership. That is a tutorial with a logo on it.

For a company that has spent a decade positioning itself as the neutral, unconflicted, structurally necessary alternative to every walled garden on the internet, "our AI strategy is a licensed integration with a third party model we do not own and cannot control" is a strange and revealing thing to lead with.

It is a rental car with Trade Desk bumper stickers on it.

And the question that did not get asked on that stage, that every serious investor and every serious journalist should be asking right now, is devastatingly simple: if the AI is rented, what exactly is the competitive advantage we are paying a premium multiple for?

That is what Part 2 is for.

Part 2 is called "The AI That Isn't." It is about what Kokai actually does when nobody is on stage describing it. It is about what it means that Jeff Green fired his CTO two years ago, decided the role did not need to be filled, and has been running technology strategy himself ever since at a company he is publicly calling an AI infrastructure business. It is about the gap between "distributed AI" as a phrase Jeff Green says and "distributed AI" as a thing that actually exists inside The Trade Desk's stack. And it is about what the people closest to that stack say when they are not on a friendly stage in front of a thousand people who want the story to be true. Spoiler: they use the words "machine learning" and "optimization" a lot. They do not use the word "transformation" very much at all.

Part 3 is called "The Loneliest Monopolist." It is about a man who has a $5.2 billion pay package, no CTO, a stock down 60%, agencies walking out the back door, and a trade press he has decided is the problem. It is about what happens to a founder CEO when the accountability structures around him get thin enough that the only check left is a journalist willing to read his own quotes back to him. And it is about the one question Jeff Green will not answer on any stage, friendly or otherwise: what exactly is The Trade Desk if the neutral party story stops being true?