Future of Marketing Briefing: The public ad tech era is over

That’s four sizable delistings inside a year, if Criteo closes — a clear signal of where ad tech sits right now.

Gone are the days when investors got themselves in a tizzy over ad networks, demand-side platforms and supply-side platforms on their own merits. Now they’re more likely to have their heads in their hands, or worse. The Trade Desk grew revenue just 3% in its second quarter, versus 19% a year earlier, and subsequently lost more than 20% of its value. CEO Jeff Green admitted the quarter “did not meet the standard we set for ourselves”. AppLovin, meanwhile, grew revenue 53% and still hit a 52-week low on a modest miss. Teads suspended its third quarter guidance entirely after a 17% decline.

These are some of the biggest, more durable businesses in advertising, and yet investors — rightly or wrongly — treat them like an endangered species anyway. Nothing persuades them otherwise. Not even double-digit growth. Nothing, that is, except being one of the few names still growing fast enough, in the right places, to earn the benefit of the doubt. What’s left of these businesses isn’t so much dying so much as splitting in two: infrastructure and measurement businesses better off private, and a shrinking set of platforms still growing fast enough for the market to keep faith. Magnite and PubMatic, both up 11% in the second quarter, are the clearest proof the market hasn’t given up on ad tech entirely. It’s just gotten far more selective about who earns that faith, and far less willing to extend it on credit. 

“It’s very easy to say the billboard’s going to exist in five years,” said Mark Boidman, a partner and the head of media and entertainment at Solomon Partners. “Not so easy to say that about ad tech. Will ad tech exist in the same way five years from now, with AI? Hard to say. Investors aren’t discounting these companies for what they did last quarter. They’re discounting them for what nobody can yet prove they’ll still be in five.”

Nobody can prove it, and the people building these businesses don’t pretend otherwise. 

“It’s a very fast-moving space, and the companies that are able to win are the ones that make the right bets and move very quickly,” said Chance Johnson, president of publicly traded ad tech firm Nexxen. 

The trick, he continued, isn’t answering the five-year question. It’s picking a direction and committing to it before the market forces an answer out of you. For Nexxen, that was owning more of the transaction than a point solution vendor ever could. Years of acquisition built a demand-side platform, a supply-side platform and the data connecting them under one roof. CTV, starting in 2019, and mobile in-app more recently, were accelerants for that bet, not the bet itself. 

Whether that bet has paid off depends on which quarters you look at.

The strong ones make Johnson’s case cleanly: Q1 2024 and Q2 2024 saw back to back quarters of solid revenue growth met with a rising stock, and Q4 2024 added another. The rough ones tell a messier story — and there were more of them, and more recently. From Q1 2025 through Q4 2025, growth and stock price were essentially never in sync. Decent revenue growth was met with a falling stock twice (once by 21% in a single day), a third quarter saw an early pop reverse within days, and by Q4 2025 revenue actually declined year-over-year while the stock rallied anyway.

That’s what makes Q1 and Q2 2026 interesting. It’s not that they’re unprecedented since growth and price lined up before too but that they snapped a four quarter run where the market seemed to trust nothing Nexxen reported. Record CTV growth, raised guidance and a stock that rose and kept rising followed through for two consecutive quarters right after the company’s roughest stretch. Granted, two quarters isn’t proof the full tech stack bet has worked. Not when the same alignment showed up before and didn’t last. But it is the first sign since early 2024 that skepticism might be easing rather than hardening.

“When something is purely commoditized they recognize very quickly that there’s not much there,” Johnson said of investors. “Even if there’s growth there, they question how sustainable that growth is.”

It’s the same problem The Trade Desk is grappling with. Revenue has stayed positive for seven straight quarters but decelerate hard — from 22% (Q4 2024) down to just 3% (Q2 2026) — and the stock has punished nearly every step down regardless of the number still being positive, Q4 2024 fell 33% on results that still grew 22%, Q2 2025 fell 39% despite 19% growth, Q3 2025 and Q4 2025 fell a more modest 4 to 6% as growth eased into the high teens and low teens, while Q1 2026 barely budged the stock (-2%) even as growth slowed further to 12%, and Q2 2026’s slide to 3% growth triggered another 22% rout. The lone exception was Q1 2025, which popped 19 to 30% on a relief rally after guidance reassurance. Unlike Nexxen, where good and bad quarters split roughly evenly between revenue and stock reaction, The Trade Desk’s pattern is blunter: growth never actually turned negative but the market has treated deceleration itself as the disqualifying event almost every single quarter.   

The public era isn’t over because ad tech stopped working. It’s over because the market stopped taking anyone’s word for which side of that line they’re actually on. Growth used to be enough to prove it. Now it’s just the opening bid — due again next quarter, and the one after that, until a company either proves it built a moat or the market decides it was selling access all along. 

The public ad tech era is over. Private equity isn’t ready to pick up the slack

Just because one door closes doesn’t mean another simply opens. The public ad tech era isn’t moving into a private one. Sure, buyers rarely pass up a discount sitting in plain sight but not all private capital is chasing the same thing, and not all of it is even free to chase. Private equity’s own numbers tell that story. Deal volumes in business services fell 29% year over year, first-time platforms fell 48% and assets held for more than five years now make up 26% of the market, according to SI Global’s 20266 Private Equity Insights Report. The firm tracked 80 private equity firms and 266 portfolio companies to get there. Almost a third of tracked investments are now “overdue” for an exit, up roughly 30% year-over-year. 

We asked Joe Hine, a partner at SI Global, to unpack what’s actually happening beneath those numbers. 

Hine said: “The private equity market’s got issues but at the lower mid-marklet end, the demand is still really hot. 

What he means: Not all private equity is stuck. First-time deals — say, a founder selling for the first time or a slightly clunky business with good bones that a fund can size up quickly — are still moving easily, because that capital was raised recently and isn;t weighed down by a return target set at 2021 to 2022 prices. The squeeze, in other words, is concentrated further up the chain. 

Hine said: “Anyone that put money in and around that time [2021-22] is struggling to get it out because they can’t get their turn. They need two to three times return. 

What he means: This is really the crux of the SI Global data. Multiples ran 30 to 50% in that era when, in Hine’s words, ad tech saw the best growth and therefore the highest prices the industry had seen “in a generation”. Funds that deployed then underwrote a two to three times return on those valuations. Years of war after tariff turmoil after tech shift plus now AI meant the growth never came back to justify those prices. That capital isn’t idle by choice. It’s stuck waiting for a number the market won’t currently pay. 

Hine said: “Private equity is going to have to look through a different lens as to how they manage to exit businesses. What you’ll see is people buying things for a debt return as well as an equity return.”

What he means: If a fund can’t hit the two to three times equity return it needs, it doesn’t have to sell at a loss to exist. It can restructure the deal so debt does the work instead. A 10% return on loan notes plus a 10% dividend gets a buyer to 20% annually without needing the underlying business to grow into a higher valuation at all. It’s a way of getting paid without waiting for — or needing — the market to agree the company is worth more than it currently thinks. 

What we’ve heard

“An agent will automate a repetitive task. And if you have waste in the system, you’re automating a repetitive task where waste continues to exist. So before we even go to a buying agent we have to clean up what we want. Would we consider buying agents in the future? Yes, we would but it’s more important that we try to get the best quality impression at the lowest possible cost and then at that point we try to automate that repetitive task.”

— Paras Shah, Georgia-Pacific’s senior director of digital media

Numbers to know

70%: Percentage of U.S. social buyers that have made a purchase on TikTok in the past 12 months

24%: Percentage of consumers that find new products to purchase on YouTube

1 billion: Total number of monthly users that Google’s Gemini app now has

$567 million: The amount Meta was fined by a U.S. judge in New Mexico for its failure to warn the public about the dangers its platforms pose to children — it’s largest fine over child safety to date

What we’re reading

Superintelligence is a dragon

Platformer’s Casey Newton argues Zuckerberg’s “The Future is for Everyone” manifesto is a policy wishlist dressed as vision — faster data centers, looser training-data rules, legal cover for distillation, all Meta’s commercial interests. The core critique: it reframes safety as a distribution problem rather than a control problem, just as frontier models are escaping constraint in testing.

Restaurants, pubs and theatres ban Meta’s ‘spy glasses’ over privacy fears

U.K. venues are banning Meta’s £359 smart glasses over privacy: Jeremy King has barred them from his London restaurants, Soho House and ATG Theaters will ask people to remove them, and Wetherspoons won’t allow recording. It follows cases of non-consensual filming, though Meta says sales have tripled past 7 million worldwide, according to The Guardian.

TikTok Lays Off 250 Employees in Office That Includes Content Moderation

TikTok’s U.S. joint venture is cutting 250 jobs and closing its Nashville office on Oct. 5, including content moderation staff, who were logged out of their devices the morning they were told, per the New York Times. The company cited only streaming, though AI is displacing that work across the industry.

Can the apps stop AI slop?

Almost every major platform now brakes AI content, from Pinterest’s feed toggles and TikTok labelling to Snapchat pulling AI videos from Spotlight. Platforms overestimated the appetite for AI features, and an outright ban looks like the next step, per Social Media Today.

What we’ve covered

OpenAI is coming for SMB advertisers

Six live job postings show OpenAI building a dedicated SMB ad unit spanning growth, data science, demand strategy and sales ops — plus a Dublin-based vendor manager to police outsourced sales firms, the model Google and Meta have already run. EMarketer’s Nate Elliott notes SMBs make up the majority of ad revenue at both, so any serious ad business has to go there.

Lifestyle brand Stanley 1913 adapts its marketing for the AI search era

Stanley 1913’s influencer-led visual marketing built awareness with humans but left LLMs nothing to parse, so it’s rebuilding the product information layer — FAQs, care instructions, usage guides, explicit feature-to-benefit copy. It’s testing Shopify and Google’s Universal Commerce Protocol to get its catalog into chat, with Yotpo baselining citations.

Experian retires the Audigent brand, folding it into Experian Marketing Services

Twenty months post-acquisition, Experian has folded the Audigent brand into Experian Marketing Services, calling it consolidation and declining to comment on layoffs. Curation had a clear definition when the deal closed in 2024; now everyone in the supply chain claims the label. Eyeota’s Marc Fanelli likens it to DMPs and identity — categories that disappeared by becoming infrastructure.

Why Georgia-Pacific bet the future of programmatic on supply, not demand

Georgia-Pacific has cut its SSP roster by 80%, from 30-plus to around six, after exhausting DSP-side gains. It brought in SWYM to score inventory before the bid rather than after — CPMs down 17% to 44%, viewability up 11%, completion up 7%. Fees are flagged as the next pressure point on SSPs.