The Trade Desk, Amazon and the question: Who’s the Geisterfahrer?

There’s a German Geisterfahrer joke that goes as follows: A driver is on the motorway when the radio crackles: Caution, there is a ghost driver on the road. The driver looks up, peers through the windscreen, and mutters, “One? There are hundreds of them.”

Something about that joke keeps coming to mind when looking at DSPs. Especially after Jeff Green, CEO of The Trade Desk, recently made big waves (twice actually, once with his post and later, in more detail, at Marketecture Live) that he believes Amazon’s DSP may not exist in five years. Bold.

Admittedly, Jeff Green is not a disinterested observer. His business largely depends on being right about the competitive landscape, and while he has at times argued that Amazon is not a direct competitor, he also acknowledged that Amazon’s ad business is performing strongly. And yet, his current position remains.

Not because Amazon’s ad business is weak (it isn’t), but because the margin is in sponsored listings and Prime Video. The open internet is where Amazon’s antitrust exposure lives, where the regulatory risk accumulates and where the money doesn’t actually come from. So why maintain expensive open web infrastructure for a fraction of the return, when the O&O cash cow inventory is sitting there doing the real work?

And yet, recent evidence appears to point the other way. At least on the surface. Amazon has been busy growing its DSP business: Amazon + Netflix, Amazon + Spotify, Amazon + SiriusXM, Amazon + Microsoft, and the list keeps growing. One new, big, noisy collaboration announcement chases the next. Which is quite a turn for Amazon, having treated its ad business like a neglected side gig for years, to now going full-blown “we’re eating the entire open web” DSP-killer mode. However, none of these collaborations, none of this supply is really unique to Amazon, despite the hype suggesting otherwise.

Amazon’s DSP was never particularly remarkable. A year or two ago, it wouldn’t have been discussed in most media strategy meetings. Sure, it existed, strategically useful for what only Amazon can offer: shopper data, Prime Video inventory, the JBP sweetener. But a centerpiece? Never. It was always neither fish nor fowl. A lot like Adobe’s DSP, for those who remember. Or Microsoft’s Xandr, for those who’d rather not. Both companies decided the DSP business wasn’t worth the complexity and tossed their ad tech stacks aside when their core businesses needed the attention. Amazon runs AWS, a global retail operation and a growing entertainment business. The open internet is certainly not their main event. It never was.

But what if Amazon sticks with its DSP? Luckily there’s no need to roam far afield for a cautionary tale… just look at Google’s DV360. The DSP still exists, but since 2018, it’s been quietly redirecting spend toward YouTube. The open web got the infrastructure; YouTube got the money. Lose; win. Amazon, following the same logic, suddenly sounds possible, if not likely, if not even faster, because the pressure on their retail and AWS businesses makes open internet exposure a strategic problem, not just a margin problem.

And yet, local, international publishers and media groups alike are leaning harder into Amazon DSP as core infrastructure, outsourcing their pipes to a walled garden. Doing it, of course, for short-term revenue. The supply side desperation runs deep, especially in times of AI shrinking discovery. And even publishers that should know better, premium ones, the kind with recognizable brands and valuable audiences, eventually plug into every DSP available. Disney. Netflix. The ones that started with their own tech, their own exclusive partnerships, their own principles about protecting their supply and data. Revenue pressure apparently has a way of dissolving all principles.

On the advertiser side, Amazon’s aggressive pricing has helped attract ad budgets — and, as ever, supply follows demand. The FOMO of not being on the one platform where the money is moving proves stronger than the strategic case for staying off it. Which is, again, exactly how Google and Meta captured publishing the first time. The publishers knew the playbook. They followed it anyway.

Sounds familiar? The plot didn’t change much. Enter the market. Scale fast. Commoditize the incumbents. Then compress margins until the host has nothing left to offer that the platform couldn’t do without them. Because once a publisher routes its business through Amazon’s infrastructure, the advertiser relationship belongs to Amazon. The data belongs to Amazon. Publishers are left with a revenue share and the illusion of partnership, until the terms change. Those terms are Amazon’s to rewrite whenever the economics shift. Which they will. Because they always do. Which again, isn’t speculation; it’s the Google and Meta playbook, slightly repackaged with an Amazon Prime logo. After all, Amazon didn’t build its DSP to keep the open web alive. It built it for Amazon.

So who’s the ghost driver? The publishers, confidently heading into oncoming traffic while hundreds of headlights bear down? Amazon’s DSP itself, already driving the wrong way on a road its parent company is quietly planning to close? Or the advertisers lured in by the JBP, that will discover what “renegotiation” means when Amazon decides the terms?

Jeff G. or Jeff B.? Let’s circle back in five years…

This op-ed represents the views and opinions of the author and not of The Current, a division of The Trade Desk, or The Trade Desk. The appearance of the op-ed on The Current does not constitute an endorsement by The Current or The Trade Desk.