This website uses cookies

Read our Privacy policy and Terms of use for more information.

Welcome to Ads 2026: Where the Bots Are Fake, the IDs Are Broken, and Every Media CEO Is “Approaching Profitability”

This week in media felt like watching a casino continue operating after everyone realizes the slot machines are plugged into extension cords from RadioShack. The Trade Desk spent years terrorizing publishers into adopting UID2 only to discover nobody could tell when the IDs were broken. NBCUniversal is chest-thumping over Peacock growth that required the Super Bowl and the Olympics — basically the media equivalent of steroids and a favorable wind tunnel. Pinterest keeps bragging about AI-driven engagement while the internet quietly fills with bots clicking on ads for other bots. Google has now decided Reddit commenters are “experts,” which is either the death of search or the most honest thing Silicon Valley has ever done. And every agency CEO suddenly sounds like a guy who watched two OpenAI demos and now thinks firing copywriters is “innovation.”

The entire business is running on panic disguised as momentum. Streaming companies are stapling live sports onto money-losing apps and calling it strategy. Adtech firms are selling identity systems that may or may not work. Platforms are inflating growth with synthetic traffic while AI systems vacuum up whatever human content still exists online before replacing it with machine sludge. Somewhere, a consultant is billing $600 an hour to explain “signal loss” while three fake users argue with each other in a Pinterest comment section generated by Nvidia GPUs. Stay digital, everybody.

The UID2 Circus: Garbage In, Programmatic Out

The Trade Desk’s “Identity Solution” Apparently Doesn’t Check Identity

Publishers were pushed for years to adopt UID2 like it was oxygen for the future of CTV advertising, only for one major publisher to discover they’d been sending completely broken IDs into The Trade Desk’s pipes for three months without anyone noticing. Not TTD. Not buyers. Nobody. Even better, when the publisher fixed the encryption issue, revenue didn’t change at all — which is an incredible advertisement for a product supposedly essential for better targeting and higher CPMs. The whole thing exposes the dirty little secret of adtech identity: everyone’s pretending these systems are more sophisticated than they actually are while billions move around based on vibes, dashboards, and fear of missing out.

Peacock’s “Profitability” Is Just Comcast Grading on a Curve

NBCUniversal Keeps Declaring Victory for Finally Catching Up

NBCUniversal is acting like it reinvented television because sports temporarily inflated Peacock and ad revenue, but this isn’t breakout growth — it’s expensive catch-up after years of fumbling streaming while Netflix, YouTube, Amazon, and even Disney built actual scale. Yes, Peacock crossed $2 billion in quarterly revenue and gained subscribers during a quarter containing the Super Bowl and the Olympics — meaning NBCU basically paid a fortune to rent the most unfair growth cheat code in media. Strip out the mega-events and Comcast admitted growth falls to low single digits, which is less “streaming dominance” and more “sports dependency with a login screen.”

The funniest part is executives once again promising profitability is “approaching,” a phrase Peacock has stretched longer than a Marvel franchise. NBCU isn’t leading the streaming wars — it’s finally avoiding total embarrassment. The company spun off cable dead weight into Versant, stuffed every possible sporting event into Peacock, and now wants applause because losses are shrinking. Congratulations on achieving what Netflix did years ago without needing the Olympics every quarter to prop up the graph.

Martin Sorrell’s AI Pep Rally Comes With Layoff Energy

S4 Capital Discovers “Innovation” Looks a Lot Like Cost Cutting

Martin Sorrell rolled out another weak quarter wrapped in enough AI buzzwords to stun a venture capitalist into temporary silence, insisting S4’s future lies in AI-generated commercials and “output-based” pricing. Translation: agencies are trying to charge premium fees for work increasingly produced by software. Revenue dropped, clients remain cautious, and investors briefly hammered the stock because the market is starting to realize every holding company now sounds exactly the same — lots of talk about Runway, Nvidia, automation, and “hyper-personalization” while clients quietly ask why they’re still paying agency margins at all.

The entire ad business is entering its existential phase where executives must simultaneously claim AI is revolutionary while pretending it won’t destroy the labor model their companies depend on. Sorrell talks about efficiency like it’s innovation, but what agencies really want is fewer employees, faster production, and clients too dazzled by synthetic video demos to notice the business is becoming a glorified prompt factory.

Pinterest’s Growth Story Gets Weird When the Internet Is Full of Bots

AI-Powered Discovery Is Great Until Nobody Knows Who’s Human

Pinterest posted strong user growth and over $1 billion in revenue while aggressively pitching its AI-powered “Taste Graph” as the future of personalized discovery, which sounds impressive until you remember half the modern internet increasingly consists of bots interacting with other bots. User growth exploded in lower-monetized regions while ad pricing declined, a combination that usually translates to: “tons of activity, questionable value.” But Wall Street hears “AI” and immediately starts levitating three inches above the conference call.

Pinterest insists personalization is driving engagement and shopping behavior, but the broader industry problem is impossible to ignore now. Platforms are rewarding scale over authenticity while synthetic traffic quietly floods ad systems already drowning in garbage impressions. The next phase of digital advertising may just be machines optimizing campaigns against fake audiences while executives congratulate themselves on “record engagement.”

Google Search Has Officially Become Reddit With Better Branding

AI Search Now Treats Forum Comments Like Expert Analysis

Google announced AI Search will increasingly surface Reddit posts, forum comments, and social chatter as “expert” guidance, which is either a clever evolution of search or the final collapse of institutional authority online. After spending decades organizing the web, Google now wants users to trust “some guy who once photographed the Northern Lights” as a primary source. The company says firsthand experience creates authenticity, but really this is what happens when AI systems run out of reliable content and start scraping the comment section for nutrients.

It’s also a quiet admission that Google broke the publisher ecosystem it depended on. The company spent years vaporizing referral traffic while trapping users inside search results, and now suddenly it’s pretending links and human expertise matter again. AI search isn’t replacing the web — it’s slowly turning the web into background material for Google’s answer machine while publishers get whatever scraps survive the summarization layer.

Threads Wants to Become Your Work Personality

Meta Is Turning Twitter’s Corpse Into Enterprise Social Media

Threads adding desktop DMs and expanding its web tools confirms Meta’s real ambition here: make Threads unavoidable for brands, creators, and social managers trapped online all day. This isn’t about culture or conversation anymore. It’s about turning Threads into a cleaner, advertiser-safe version of X where engagement can be packaged into Meta’s larger ad machine without Elon Musk randomly detonating the platform every Thursday.

The redesign sounds less like a social network refresh and more like another corporate dashboard — analytics, saved posts, messages, navigation tools. Meta doesn’t need Threads to be cool; it just needs it to be stable enough for marketers to move budgets. Which, compared to the current state of X, might actually be enough.

Pinterest Is Half AI Slop And They're Calling It "Growth"

The platform's own users are screaming that 95 out of 100 pins are garbage. Pinterest's response: more AI, more bots, more record numbers, and a securities fraud lawsuit on the side.

Let me tell you what's happening at Pinterest, because this is one of those stories that is so obviously absurd that the only reason it isn't the lead on every business page is that everyone in this industry has agreed, by mutual silent contract, not to look at it directly.

Pinterest just posted $1.008 billion in Q1 2026 revenue. Up 18%. 631 million monthly active users. Tenth straight quarter of double-digit user growth. The stock popped 15% after the report. CEO Bill Ready got on the call and talked about "differentiated visual search product experiences" and an "AI-powered ads platform." Wall Street levitated. Analysts upgraded. Everyone went home happy.

And here's the thing nobody on that earnings call mentioned: the platform is, by the testimony of its own users and journalists who have actually opened the app recently, a synthetic wasteland. A user quoted by 404 Media in February said it cleanly: "I can't even scroll through 100 pins without 95 out of them being some AI slop or theft."

Ninety-five out of a hundred. That is not me being dramatic. That is a user, on the record, describing the actual experience of using Pinterest in 2026. And it tracks with what every single piece of journalism on this platform has been saying for over a year.

Pinterest Is Not Fighting AI Slop. Pinterest Is The AI Slop.

This is the part that took me a minute to fully process, so let me lay it out, because it is genuinely remarkable.

You would assume, naturally, that the AI garbage flooding Pinterest is the work of bad actors. Spammy content farms. Bot operators. The usual suspects gaming an algorithm. And yes, that's part of it. But here's the kicker, courtesy of a December 2025 CNBC report by Jonathan Vanian and surfaced again in an April 2026 Plagiarism Today analysis: much of the AI content overrunning the site is coming from Pinterest itself in the form of ads. In October 2024, Pinterest launched a generative AI tool for advertisers to automate their creation of ads, including manipulating images. The number of ads and the amount of AI-generated content in those ads is very much by design.

By design. Read that again. Pinterest is not a victim of AI slop. Pinterest is a vector for it. They built the firehose, pointed it at their own users, and turned it on.

And it gets worse. Pinterest is training its AI models on human-created content and then using that model to add more AI-generated content to the site. They train on the humans. They generate slop from the humans' work. They flood the platform with the slop. The humans leave. The slop remains. The Taste Graph trains on the wreckage. Repeat.

This is a perpetual motion machine for digital garbage, and Pinterest is calling it innovation.

"It's Impossible To Talk To A Single Human"

The 404 Media investigation from February 2026 is devastating if you take five minutes to read it. Artist Tiana Oreglia, who has been on Pinterest for years, told them: "I feel like, increasingly, it's impossible to talk to a single human [at Pinterest]. Along with being filled with AI images that have been completely ruining the platform, Pinterest has implemented terrible AI moderation that the community is up in arms about. It's banning people randomly and I keep getting takedown notices for pins."

Nobody on the other end. Bots talking to bots. AI moderating AI. Real humans getting banned in the crossfire while the slop sails right through.

Artist Min Zakuga reported her decade-old hand-drawn artwork was flagged as "AI modified." "Even artwork from 10-13 years ago will still be labeled by Pinterest as AI, with them knowing full well something from 10 years ago could not possibly be AI."

So let's get this straight. Pinterest's AI moderation system:

  • Cannot reliably detect actual AI content. Slop continues to dominate feeds.

  • Falsely flags real human art from before generative AI existed as AI-generated.

  • Bans human accounts with appeals processes that eat hours of artists' time.

  • Allows bot operators, like the one documented on Reddit, to run thousands of automated clicks per session without throttling, when X and Facebook would catch the same activity in minutes.

This is not a content moderation system. This is a sorting hat for chaos. And Pinterest's CTO Matt Madrigal describes it as "striking the right balance between human creativity and AI innovation."

The right balance. The right balance is 95 AI pins out of 100, apparently. That's the balance the user above described. That's the balance Pinterest has chosen, because every quarter it gets to put bigger numbers on a slide deck.

The Numbers Are Real. The Users May Not Be.

Pinterest's official MAU definition, in their SEC filings, counts authenticated accounts, not unique humans. If one person operates four accounts in 30 days, the language permits Pinterest to count that as four MAUs. There is no deduplication requirement. There is no "actual person" check.

Combine that with the Reddit bot operator who bragged about running automated activity on Pinterest for a full year, unthrottled, with video evidence of the platform never flagging him. Combine that with Pinterest's own SEC disclosure that 5.2% of ad requests in 2025 were blocked as invalid traffic (the figure for what they caught, which is always lower than what got through).

Now look at the geography of the "growth":

United States and Canada ARPU: $7.12. User growth: roughly 4%. Europe ARPU: $1.17. User growth: roughly 7%. Rest of World ARPU: $0.20. User growth: roughly 15%.

The fastest-growing user segment generates twenty cents per user per quarter. That is roughly 2% of the value of a U.S. user. The "record growth" Pinterest brags about is concentrated in a demographic where each "user" is worth pocket change, and where the platform's ability to distinguish a person from a bot is, charitably, unproven.

Meanwhile, Pinterest had to acknowledge in its Q1 2026 results that ad impressions grew 24% but average ad pricing declined 5%. More inventory, less valuable inventory. That is the textbook signature of a platform whose audience quality is degrading even as the headline numbers grow.

The Lawsuit Nobody Talks About On The Earnings Call

There is a securities fraud class action against Pinterest, Uziel v. Pinterest, in the Northern District of California, covering the period from February 2025 through February 2026. The complaint alleges Pinterest made materially false and misleading statements about advertising revenue resilience. Three corrective disclosures during the class period dropped the stock from the mid-$30s to $15.42, with cumulative per-share losses of $12.77 and a 15% global workforce reduction.

The lawsuit isn't specifically about bot-inflated MAUs. It's about ad revenue. But it establishes, on the legal record, that the gap between Pinterest's narrative and Pinterest's reality has grown wide enough to produce litigation. And it didn't make the earnings call.

The Taste Graph Is Training On A Polluted Signal

Here is the structural absurdity at the center of the Pinterest pitch.

Pinterest sells advertisers on its "Taste Graph", hundreds of billions of interactions, AI-powered personalization, real intent from real shoppers. Bill Ready talks about it like it's the holy grail of digital advertising. High intent, ready to buy, AI knows what they want.

But the Taste Graph is trained on:

  • Bot interactions the platform admittedly fails to throttle

  • AI-generated pins that systematically outrank human content

  • AI-generated ads that Pinterest itself produced via its own generative tools

  • Synthetic engagement on synthetic content linking to synthetic blogs

  • The clicks of users who got fooled before they realized the pin was fake

The signal is polluted. The optimization target is fictional. The "personalization" is matching synthetic humans to synthetic content. And Pinterest is selling that as insight into commercial intent.

When Bill Ready stood on the Q4 2025 earnings call and claimed Pinterest sees 80 billion searches per month, more than ChatGPT, "one of the largest search destinations in the world", he didn't disclose what percentage of those searches came from human fingers on human phones. Because he doesn't have to. And because the answer, whatever it is, is almost certainly worse than the slide deck suggests.

The Reason Nobody Calls This Out

Here's the part that should make every advertiser reading this uncomfortable, and most of them won't be, because the entire industry is running on a shared agreement not to look.

Imperva's 2025 report: 51% of internet traffic is automated. Bad bots: 37%. A March 2026 Human Security report: AI-agent traffic up 187% in early 2025. Global ad fraud losses projected at $63 billion in 2026. Fraudlogix examined 105.7 billion impressions in 2025 and found 20.64% global invalid traffic.

One in five impressions. That is the universe Pinterest operates inside. And Pinterest's own filters caught only 5.2%. The math, charitably, does not work.

But nobody wants to do the math. Advertisers don't want to admit a meaningful slice of their spend is feeding bots. Platforms don't want to disclose how porous their counting is. Investors don't want to interrogate the engagement metrics under the AI pitch, because the AI pitch is what's holding the multiples up. The AI pitch is the multiple.

So Pinterest gets to stand on stage every quarter and announce record users, record searches, record AI-powered everything, and the room nods and the stock pops, and meanwhile the artists who built the platform are getting banned by bots, and the bots are getting promoted by bots, and the bots are buying ad inventory from bots, and somewhere in the middle of it, a real human is trying to find a recipe and getting served an AI-generated photo of food that doesn't exist, linking to an AI-written blog with an AI-generated author photo, monetized by ads written by an AI tool Pinterest sold to the spammer.

The Dirty Secret

Pinterest's growth is real in the sense that the numbers are accurate. Revenue is real. The billion dollars exists. The 631 million accounts exist.

What is not clear, what Pinterest will not tell you, what the SEC filings carefully decline to specify, is how many of those accounts are people, how many of those searches involve a human, and how many of those interactions reflect anything other than machines optimizing against machines while the executives at the top of the org chart count the engagement and pretend it's an audience.

The platform that built itself on human inspiration, on real people pinning real things they wanted to make and buy and remember, has decided that synthetic activity is just as good as long as the dashboards keep going up and to the right.

Pinterest doesn't want anyone to know it's drowning in slop. Pinterest has made the slop a product line. And every quarter the gap between what they're selling advertisers and what's actually on the platform gets a little wider, and one of these days, somebody with real money is going to do the math out loud.

When that happens, the "record growth" story will look exactly like what it is: a Potemkin village built out of bots, run by AI, sold as the future of human discovery.

I'll be here when it falls.

You're reading the free edition.
ADOTAT+ is where the work happens.

The full ADOTAT Scorecards. Vendor comparisons that name winners and losers, with the analysis your planning team needs before Q3 budget meetings.

The takes nobody else will publish. Independent means we don't answer to vendors, agencies, or the trade press establishment.

The archive. Every Scorecard, every analysis, searchable. The reference layer for how this industry actually works.

Read by senior leaders at every major holding company agency, GroupM, Publicis, Omnicom, IPG, Havas, Dentsu, and inside the buying and strategy teams at Netflix, Roku, Pinterest, Disney, NBCUniversal, Spotify, Amazon, Meta, and most of the platforms shaping where ad dollars actually go.

If your competitors are reading ADOTAT+ and you aren't, you're planning with less information than they are.

$99/month. $999/year (save $189). Expense it.

Founding member rate. Locks in for life.

Also, no ads. Zero ads. No Ads.

logo

Subscribe to our premium content at ADOTAT+ to read the rest.

Become a paying subscriber to get access to this post and other subscriber-only content.

Upgrade

Keep Reading