Perplexity ran ads for fifteen months. Then it quit on trust.
How the first AI search engine to sell ads pulled the plug — and what their own executives said when they did.
In November 2024, Perplexity launched sponsored placements inside its AI answer engine. The pitch was clean: a publisher-friendly revenue share, contextual brand placement, lighter footprint than legacy display. The company told the market that subscription revenue alone could not support its growing publisher partner program. Ads, the company argued, were essential.
In February 2026, the ads were gone.
The timeline
November 2024. Perplexity launches its ad program. Stated ambition: a sustainable revenue-sharing model with publisher partners.
Throughout 2025. Brand applications outpace approvals — by a wide margin. Fewer than 0.5 percent of brands that applied to advertise on the platform were ever admitted.
Late 2025 – Early 2026. Taz Patel, the executive leading the ads effort, departs the company before the experiment runs its full course.
February 20, 2026. Campaign US and the Financial Times report that Perplexity has pulled the program. Executives go on the record. The reason is trust.
The quote
The decisive line — from a Perplexity executive, to the Financial Times:
“A user needs to believe this is the best possible answer.”
The executive added that ads risk making users “suspicious of everything.”
That sentence is the most important business document of the year for anyone selling, buying, or measuring inside AI engines. It is a public admission, from inside the category, that ads inside an answer cost more in trust than they earn in revenue.
The math
Three numbers explain the decision.
0.5 percent — the approval rate for brand applications. Perplexity tried to keep ad inventory premium and credible. The discipline killed scale.
$60 → $25 — ChatGPT ad CPMs in the same three-month window. Even the company most committed to AI advertising is watching price compress. Premium positioning was getting harder, not easier.
Two engines refusing. Anthropic publicly committed to keeping Claude ad-free, with an eight-figure Super Bowl campaign saying so. Google DeepMind chief Demis Hassabis is on the record that ads “have to be handled very carefully.” When two category leaders are signaling that ads are a trust problem — and the third, OpenAI, is selling them at scale — the middle position is the hardest to defend.
Perplexity sat in the middle. The middle didn’t hold.
What it means
Three things.
One — The trust math in answer engines is structurally different from search. Sponsored links sat next to answers. Sponsored placements inside a chatbot voice sit inside the answer. Once the voice takes a sponsor, the voice becomes a sales channel. Perplexity executives told the FT, in essence, that they could not sustain both.
Two — Citation Share just became more valuable. When the most trust-positioned AI engines — Claude, Gemini in chat mode, and now Perplexity — are not selling placement inside the answer, the only way to appear inside their answers is organic citation. Generative Engine Optimization is the channel. Earned media into trusted publications is the input.
Three — OpenAI’s bet is now the contrarian one. ChatGPT crossed $100 million in annualized ad revenue inside six weeks. The audience is enormous. The convenience is real. But OpenAI is now alone among the trust-positioned AI labs in selling inside the answer — and the longer it sells, the more Anthropic’s Super Bowl framing sticks.
The frame
Perplexity blinked because the trust tax was bigger than the revenue line. That is the structural lesson of the year — and it applies in three directions.
To the engines: ads inside the answer carry a credibility cost the rate card does not capture.
To the brands: the most valuable position inside an LLM today is one that money cannot buy. Earn it now.
To the industry: the question is no longer whether AI engines will sell ads. The question is which ones will pay the trust tax — and which ones will refuse, and turn refusal into positioning.
Perplexity refused. Read what they said.
Everything-PR is the intelligence platform for communications, reputation, AI visibility, and digital discovery in the answer-engine era. Thirty-plus publications. Publishing since 2009. Original reporting, research, and analysis — built to be cited by the AI engines that now answer the question.
