AdTech Composite Growth Slows to 7.6% in Q2 2026

Full analysis

Your draft

Madison & Wall's ad-tech composite grew 7.6% in Q2 2026, down from 11% the quarter before and mid-teens through 2025, while EBITDA margins slipped to 31% from 32%. Total digital ad spend is still accelerating, but the gains are pooling at Google, Meta, and Amazon. The open web is where the deceleration lives.

What's being decided (briefing mode): Whether the open-web slowdown is a cyclical dip that mean-reverts or a structural re-rating of every ad-tech name whose revenue depends on programmatic inventory outside the walled gardens. This is a Type 1 read for anyone allocating capital or headcount against a multi-year plan, and a Type 2 read for revenue ops doing an in-quarter reforecast. Forcing function: Q3 planning and budget commitments land over the next 60 days.

The Market Analyst. One index reading is one reading, but the direction is the story. Growth halved from mid-teens to 7.6% in about three quarters, and it happened while total digital spend went the other way. That gap is the whole game. When the biggest platforms accelerate and the middleware decelerates, money reprices the middleware down, because investors pay for where the growth is going, not where it's been. Margins ticking from 32% to 31% is nothing on its own. Paired with the growth cut, it says pricing power is leaking. In plain terms: the companies that sit between advertisers and publishers are getting squeezed from both ends.

The Skeptic. Slow down. The composite is a blended bucket, and blending hides everything useful. AppLovin's mobile machine and a wheezing open-web SSP sit in the same number, so 7.6% describes neither. Q2 2025 was also a hard comp, fattened by cookie-deprecation-delay spend that isn't coming back and shouldn't be baselined. And 31% EBITDA is a healthy business, not a distressed one. The real question nobody's asking: is this a collapse, or is open-web programmatic just settling into a mid-single-digit utility that the market keeps pricing like it's dying? Those are very different bets.

The Operator. Tuesday morning, this is a reforecast, not a philosophy seminar. If you run yield at a mid-tier SSP, you feel it first: floors get tested, header bidding partners get cut, and publisher-direct deals jump the priority queue. If you run a DSP that isn't AppLovin or The Trade Desk, the 90-day risk is budget walking mid-flight as agency planners chase performance into closed ecosystems where the numbers look cleaner. Customer success should brace for churn conversations in August. The thing that breaks first isn't revenue, it's the assumption that last year's growth rate is your planning number.

The CFO. A margin point looks like rounding until you see what defends it. Holding 31% while growth halves means somebody is cutting to protect the line, and cost cuts don't compound the way revenue does. The opportunity cost question is brutal: every dollar of open-web sales and engineering spend now competes against retail media and CTV, where the growth actually is. If you're carrying open-web exposure hoping for mean reversion, you're financing a wait. Redeploying reads as admitting the loss, which is exactly why it won't happen fast enough.

Where they split. The Market Analyst and the Skeptic disagree on what 7.6% even means: a re-rating signal, or a blended average that describes no real company. That's the crux. The Operator and the CFO agree on the near-term move (reforecast, redeploy) but part on speed. The CFO sees loss aversion keeping capital stuck in open-web positions; the Operator sees teams anchored to 2025 growth rates that aren't the baseline anymore.

What it hinges on. Two things. First, whether the deceleration is broad across the open-web cohort or concentrated in a handful of pure-plays the index happens to blend in. Second, whether the walled-garden acceleration is taking budget from the open web or just growing a bigger pie faster. If it's the former, the Strategist's "horizontal middleware loses pricing power" read holds. If it's the latter, the Skeptic's normalization case survives.

What to verify before betting the plan on it. Get the composite decomposed. Look at open-web pure-plays separately from AppLovin, Viant's CTV mix, and anyone with retail signal. If the pure-plays are growing low single digits while the diversified names hold double digits, the market is sorting winners from utilities, and that's a structural split, not a cyclical dip. The council leans structural, but that call only firms up when someone shows the cohort split.

Prediction: Madison & Wall's Q3 2026 ad-tech composite, reported in the firm's Saturday summary in roughly November, will grow slower than the 7.6% posted in Q2, coming in at or below 7%.

Confidence: Medium. Deceleration is trending and walled-garden budget pull is structural, not one-off.

Why: The composite has stepped down two quarters running, from mid-teens through 2025 to 11% to 7.6%, and the cause M&W names is open-web weakness against accelerating walled gardens, which is a share shift, not a calendar quirk. Share shifts don't reverse in a quarter because agency planners chase the cleaner performance numbers inside Google, Meta, and Amazon, and those budget decisions get locked in Q3 planning cycles happening right now. The opposite outcome, a bounce back above 7.6%, would need open-web spend to re-accelerate or an easy comp, and Q3 2025 wasn't soft enough to hand the number an easy win. The one thing that could break this is a blended-index quirk where a diversified name like AppLovin or a CTV pivot drags the average up, which is exactly why the confidence is Medium, not High.

Revisit by 2026-11-30: We're right if the Q3 2026 composite prints at or below 7% year-over-year. We're wrong if it prints above 7.6%.