Every now and then, adtech has a collective mood swing—the kind where everyone looks up from their dashboards like caffeinated cats suddenly realizing the furniture is rearranged.

That’s exactly what’s happening as Amazon DSP keeps beating The Trade Desk in head-to-head tests, and agencies start shifting budgets with the quiet shame of someone sneaking out of a bad date early.

Even the ever-combative LinkedIn crowd has settled into a rare moment of agreement:
Amazon DSP is simply performing better.

This isn’t subtle.

This is gravity.

Amazon isn’t winning because it’s clever.
Amazon is winning because it has data no one else has.

Real purchase history.
Real browsing behavior.
Real closed-loop attribution.

One friend didn’t soften it when I asked why Amazon’s outperformance is suddenly so obvious. He said, “If you can tie an ad to a purchase, why would you mess around with anything fuzzier?”

There was no sarcasm. It was just reality landing with a dull thud.

This is what happens when one DSP gets to play chess while everyone else is using refrigerator magnets.

The Trade Desk’s data is a mosaic held together with third-party segments, modeled behavior, and identity workarounds that look increasingly fragile in a world where privacy laws are squeezing the open web like a stress ball.

Probabilistic modeling used to be “good enough.”
Now it feels like lighting Shabbos candles with a wind machine on.

Dave put it more politely than he needed to:
“Modeled conversions are fine until you stack them against transactions. After that, everything else feels like guesswork.”

Translation: the math was always shaky.
The market is just finally saying it out loud.

People love to point at Amazon’s 1% DSP fee like it’s proof of moral purity.

Please.

The fee is the decoy.

The real power is Amazon’s control of the most coveted CTV inventory on the planet:
Prime Video.
Fire TV.
Twitch.
Premium publisher pipes.

Amazon doesn’t need to juice fees when it controls the placement, the audience, and the result.

It’s like bragging about cheap admission to a restaurant where the entrees cost $90.
Buyers still show up because the food is only served there.

As cookies crumble and third-party data fades into the sunset, there’s only one kind of DSP that survives:
the one with authenticated, deterministic identity tied to actual transactions.

Amazon built its entire system around that assumption long before the rest of the industry admitted the party was ending.

The Trade Desk built an elegant architecture for a world that no longer exists.

Amazon built a fortress.

Inside ADOTAT+, we break down the parts nobody discusses publicly — the off-the-record agency bake-offs, the ROAS gaps people pretend aren’t real, and the performance lifts that quietly shift millions in Q4 budgets before a single press release hits LinkedIn. You’re missing the deterministic playbook that explains why Amazon keeps winning commerce, CPG, and auto… and the probabilistic cracks that TTD is scrambling to spackle over with AI and identity patches no one wants to talk about.

You’re also missing the scenarios shaping 2025–2026 — what happens if Amazon tightens its CTV grip, what it looks like if The Trade Desk stabilizes the open web, and what advertisers will face if the market fractures into walled gardens and orphaned inventory. ADOTAT+ is where we map out those futures before they land in your QBR. If you care about where budgets are moving, not where people say they’re moving, this is the part you can’t afford to skip.