The ad industry is having its annual existential crisis, and this year's flavor is AI meets panic meets denial. Everyone from Google to Amazon to LiveRamp is racing to reinvent how ads work — Google's calling it "agentic commerce" (fancy for "we want a cut of every purchase you didn't know you were making"), Amazon's muscling into Prebid like a bouncer who just bought the club, and AppLovin is literally locking users' screens and calling it strategy. Meanwhile, the IAB is trying to figure out how to label AI-generated content without admitting that basically everything is AI-generated now, and forecasters are doing their usual trick of predicting growth today and decline tomorrow so they're never wrong.

But here's the real punchline: brands spent Super Bowl Sunday selling the warmth of being human while quietly using AI to make every frame of it. They want the machine's efficiency but not its reputation — AI is the secret boyfriend nobody wants to introduce to their parents. GEO is the new SEO because apparently gaming Google wasn't enough, now we need to game the chatbots too. OOH is celebrating that a baggage claim screen finally isn't categorized the same as a highway billboard. And through it all, Wall Street keeps asking "but does any of this convert?" while the audience is three beers deep, crying over a Budweiser puppy. The game hasn't changed — we've just found more expensive ways to argue about the score.

PQ Media says global ad spending is climbing nicely, marketing spending even more so, and everyone can exhale into their Q1 decks. The update slides neatly into MediaPost’s composite forecast, bumping the average just enough to make it feel like the vibes are back.

The logic is classic: big even-year events, consumers still swiping through rising prices, brands leaning in. And then the soft shoe—expect deceleration next year. Of course. Ad forecasting is the only business where you can be bullish and cautious in the same paragraph and still sound clairvoyant. Growth now, gravity later. Collect consulting fees throughout.

At Google, ads are no longer ads. They’re experiences. They inspire. They answer questions. They gently escort you from curiosity to checkout before you realize you were shopping. In her annual letter, Vidhya Srinivasan promises that AI Mode won’t just bolt sponsored links onto chat—it will reinvent what an ad is. Which is exactly what every platform says before inserting more ads.

The big play is agentic commerce, powered by UCP, rolling into integrations with Etsy, Wayfair, Shopify, Target, and Walmart. Translation: Google wants to own the moment between “hmm” and “buy now.” It’s not search anymore. It’s concierge capitalism. The pitch is seamless and secure. The subtext is toll booth.

After years of running its own side hustle auctions, Amazon is testing a Prebid adapter that drops its demand directly into the open-source framework publishers already use. No more parallel universe with TAM and UAM. Now it’s one big happy auction family. Except Google’s not invited to this particular dinner.

Publishers are cautiously optimistic, which in this industry means bracing for impact. Fold Amazon into the same pipes, same signals, same floor logic, and maybe—maybe—you get cleaner competition. Or maybe you just get a more efficient way to lose control. The real question isn’t whether Amazon bids aggressively. It’s whether anyone else gets to breathe.

The Out of Home Advertising Association of America has updated its OpenOOH taxonomy so programmatic buyers can stop pretending that all “retail” screens are created equal. A baggage-claim screen is now transit slash airport slash baggage claim. Revolutionary stuff. Somewhere, an OpenRTB spec just felt seen.

Out-of-home has been riding an 18-quarter growth streak, but most of it is still direct-sold and wrapped in legacy logic. Programmatic buyers want precision, not vibes. The new schema promises granularity and transparency, which in adtech is code for fewer excuses. If this pushes more OOH into the open marketplace, great. If not, we’ll just keep calling bar TVs “sorta CTV” and hope no one notices.

The Interactive Advertising Bureau has rolled out an AI disclosure framework designed to avoid both deception and label fatigue. In other words: tell consumers when AI is misleading, but don’t drown them in badges every time someone used a generative background.

It sounds simple until you get into digital twins, fabricated scenarios, and whether your AI Taylor Swift is endorsing a sweater or rewriting history. Human judgment is required, which is comforting until you remember humans run marketing departments. The core tension is obvious: transparency builds trust, but too much transparency exposes how much of the machine is already synthetic.

Investors have decided subscription software is doomed, and LiveRamp is caught in the blast radius. Despite solid revenue and improved profit, the market keeps side-eyeing anything that smells like SaaS. CEO Scott Howe insists AI is a force multiplier, not an executioner.

The company is nudging toward usage-based deals with Publicis Groupe and Uber, hedging against the seat-license model like everyone else. If AI eats software, it still needs data, identity and clean rooms. That’s the bet. Wall Street isn’t fully sold, but it’s at least listening. In this climate, that counts as a win.

Super Bowl ads made one thing clear: brands are still selling AI, but with the emotional caution of someone disclosing a complicated relationship status. Research from the IAB shows a trust gap, and suddenly human-made creative is back in vogue. Real film. Real people. Real warmth.

Some brands flirt with anti-AI positioning, others quietly use it everywhere except the press release. The tension is delicious. AI is faster and cheaper, but it also whispers layoffs, slop and surveillance. Expect more nuanced messaging, fewer chest-thumping AI boasts, and a future where not using AI becomes its own marketing flex.

Enter GEO, where startups like Evertune help brands influence how AI chatbots cite their products—and now, place programmatic ads on the very pages those bots reference. With hooks into The Trade Desk and Index Exchange, the pitch is simple: if you can’t win inside the model, win right after it.

It’s conquest warfare for the chatbot era. If an LLM praises your rival’s arch support, you’d better have a banner ready screaming new and improved arches. AI condenses the research funnel. GEO tries to wedge it back open—preferably with a CPM attached.

AppLovin CEO Adam Foroughi doesn’t mince words. His ads are full-screen lockups. Users can’t look away. Somewhere, a brand safety consultant just fainted. Investors, meanwhile, want proof that ecommerce is more than a side quest beyond mobile gaming.

Conversion rates are climbing, generative AI is cranking out creative for merchants, and the ambition is enormous. But replicating the self-reinforcing loop of gaming advertisers in ecommerce is another beast entirely. It’s a bold bet that better models plus captive attention equals outsized returns. If it works, it’s genius. If not, that’s a very expensive lockup.

Stay Bold. Stay Curious. Know More Than You Did Yesterday.

Adam Foroughi would like you to know that everything is great. The ads are converting. The AI is humming. The ecommerce pivot is definitely not a desperate lunge away from a mobile gaming business that increasingly smells like it was built on quicksand. Everything is fine. The stock is down 40% from its highs, but everything is fine.

Let me tell you what's weird.

I've been covering ad tech for a long time. I've watched companies rise on narratives that made no sense and collapse when someone finally Googled the fundamentals. But I have never — never — encountered a company where the reaction from every corner of the industry is the same whispered confession: "We know something is wrong. We just can't prove it yet."

Investors tell me this. Regulators tell me this. I've had members of Congress — actual sitting members of Congress — say some version of "that company isn't right" to me off the record. The SEC is poking around. State attorneys general in Delaware, Oregon, and Connecticut are poking around. Multiple securities-fraud class actions are very much alive. And yet AppLovin just... keeps going. Like a cockroach in a tuxedo at a black-tie gala, daring someone to say something out loud.

Here's what Foroughi built, and credit where it's due — it's genuinely clever if you don't look too hard.

AppLovin's core trick is a self-reinforcing loop: it owns the games, it owns the ad network, it owns the measurement, and it owns the optimization engine. You're the advertiser, the publisher, the referee, and the guy selling the scoreboard. In mobile gaming, this created a perpetual motion machine where every dollar spent inside the ecosystem validated the next dollar. Growth looked incredible because AppLovin was grading its own homework.

The AI layer — the much-hyped AXON engine — made this loop tighter. Better models, better predictions, more efficient spend, higher returns, rinse, repeat. Wall Street loved it. AppLovin joined the S&P 500 in late 2025 trading at multiples that assumed the company had essentially solved advertising forever.

Then people started reading the fine print.

There's a buffet of problems here, and I genuinely cannot decide which one is the most damning, so let's just walk through the greatest hits:

The "Array" situation. Lawsuits and research reports describe AppLovin's distribution product as something out of a malware playbook — silent app installations, dark-pattern UX, users ending up with apps they never asked for after what amounts to an accidental tap. If your growth metric depends on people who didn't mean to download your app, that's not user acquisition. That's a hostage situation.

The Meta data allegations. Culper Research and Fuzzy Panda Research dropped reports in February 2025 claiming AppLovin reverse-engineered Meta's ad data to build a "persistent identity graph" — essentially strip-mining Facebook's ecosystem to fuel their own targeting engine in potential violation of platform rules and every privacy expectation a normal person might have. AppLovin's response was essentially "nuh-uh." Compelling stuff.

The click spoofing. Multiple class actions allege AppLovin inflated its growth numbers through good old-fashioned click fraud — the ad tech equivalent of a restaurant reviewing itself on Yelp, except the restaurant is publicly traded and in the S&P 500.

The money-laundering sideshow. In January 2026, a short-seller called CapitalWatch published a report alleging that major shareholders Hao and Ling Tang were using AppLovin's ad business as a laundromat. AppLovin called it "conspiratorial." The stock dropped 6%. Then in early February, CapitalWatch apologized and walked most of it back, saying their claims about "criminal syndicates" were inaccurate. The stock bounced 14% in a day. The whole episode was bizarre — less a smoking gun than a guy yelling "fire" in a theater, then sheepishly admitting he saw a fog machine. But the fact that a 6% sell-off happened on the rumor tells you everything about how much faith the market actually has in this company's narrative.

Here's what I keep coming back to: the vibes are terrible and everyone knows it.

The SEC investigation isn't about the money-laundering nonsense — it's about whether AppLovin misled investors on how its technology collects and uses data. That's the real game. That's the thing that, if it goes sideways, doesn't just ding the stock — it breaks the machine. Because AppLovin's entire value proposition is "our AI is better because our data is better." If the data was collected in ways that violate CCPA, GDPR, or the half-dozen new privacy frameworks coming online, then the AI isn't better. It's just more illegal.

Multiple state AGs circling the same questions only reinforces the point. This isn't one regulator with a vendetta. This is a pattern.

And yet — and this is the part that makes me want to put my head through a wall — AppLovin keeps pitching the ecommerce pivot like the house isn't on fire. Full-screen lockup ads! Generative AI creative for merchants! We're going to do to Shopify sellers what we did to mobile gaming! Which... given what we now know about what they allegedly did to mobile gaming... is maybe not the flex they think it is.

Look, I'll be fair. None of this is proven. The SEC hasn't found wrongdoing. The class actions are allegations. The money-laundering report got retracted. AppLovin says everything is false and misleading, and maybe it is. Maybe Adam Foroughi really did build a better mousetrap and everyone from short-sellers to state attorneys general just doesn't understand the technology.

But I've been doing this long enough to know what it looks like when an entire industry — investors, regulators, competitors, Hill staffers — all independently arrive at the same conclusion: something here doesn't add up.

Maybe they're all wrong. Maybe AppLovin is the most misunderstood company in tech, a victim of jealous competitors and lazy short-sellers.

Or maybe — just maybe — when everyone in the room smells smoke, there's a fire.

The stock is down 40% from its highs. The SEC is asking questions. The AGs are asking questions. The class-action lawyers are definitely asking questions. And Adam Foroughi's answer, as near as I can tell, is: "Have you seen our conversion rates?"

Cool. Cool cool cool.

I guess we'll find out.