Perplexity, the AI search startup valued at $18 billion, has pulled the plug on advertising. Executives confirmed at a media roundtable on Monday that the company phased out sponsored answers late last year and has no plans to bring them back, according to the Financial Times.
The reversal is striking because Perplexity was among the first AI companies to try this. Back in November 2024, it launched labeled sponsored answers beneath chatbot responses, complete with a blog post explaining that subscriptions alone couldn't fund publisher revenue-sharing. That argument apparently didn't survive contact with reality.
Why they bailed
"A user needs to believe this is the best possible answer, to keep using the product and be willing to pay for it," one Perplexity executive told the Financial Times. Fair enough, but the more telling quote came next: "The challenge with ads is that a user would just start doubting everything."
That word, "everything," does a lot of work here. Traditional search engines trained users to scroll past blue sponsored links. Nobody confused a Google ad with an organic result (well, fewer people did). But when an AI chatbot gives you a single authoritative answer, any commercial presence poisons the well. The format itself can't tolerate the ambiguity.
The company also lost its ad chief, Taz Patel, who departed in August 2025 after just nine months. And the ad numbers were grim: Benzinga reported the company generated roughly $20,000 in ad revenue during Q4 2024. Twenty thousand dollars. For a company valued at $18 billion. At some point, the math makes the philosophical argument easy.
The industry picks sides
Perplexity's retreat lands in the middle of what's become AI's loudest business model argument. OpenAI started testing ads in ChatGPT earlier this month, charging a reported $60 CPM with a $200,000 minimum buy. Google already runs ads in AI Mode and AI Overviews, though it's kept Gemini ad-free so far.
Then there's Anthropic, which turned the whole debate into a Super Bowl spectacle. The company aired satirical ads showing chatbots pivoting mid-therapy-session to hawk cougar dating sites, with the tagline "Ads are coming to AI. But not to Claude." Sam Altman called the campaign "clearly dishonest." Anthropic's Claude downloads jumped 32% in the days following.
So the lines are drawn: OpenAI and Google on the ad-supported side, Perplexity and Anthropic on the subscription side. The interesting wrinkle is that Perplexity tried both and chose. That carries a different kind of weight than Anthropic, which never ran ads to begin with.
Can subscriptions actually pay for this?
Perplexity claims over 100 million users and $200 million in annualized revenue as of late 2025, with paid tiers running from $20 to $200 per month. That $200 million figure represents roughly 4.7x year-over-year growth, which sounds impressive until you consider the infrastructure costs of running AI search at this scale. 780 million monthly queries don't serve themselves.
The company is now targeting enterprise customers with a five-person sales team, going after finance professionals, doctors, and CEOs who presumably care more about answer accuracy than the average user. "We are in the accuracy business, and the business is giving the truth, the right answers," another executive said, which is the kind of statement that sounds great until you have to explain it to investors watching OpenAI's ad revenue ramp up.
One executive left the door slightly ajar, telling the Financial Times they might "never ever need to do ads." That "might" is doing as much work as the "everything" from earlier.
Perplexity's bet is clear enough: users who pay for answers won't tolerate even the appearance of commercial influence. Whether $200 million in ARR and a five-person enterprise team can sustain an $18 billion valuation without ad revenue is the question that follows. OpenAI's ad rollout to US free-tier and Go subscribers is ongoing, so the industry will have comparative data before long.
