Perplexity Dropped Ads. The Only Way Into AI Search Is Earned.

Perplexity just walked away from advertising. The answer engine processing 780 million monthly queries decided that putting ads inside AI-generated answers would destroy the one thing making those answers valuable: trust. If your visibility strategy depended on buying placement in AI search, one of the biggest platforms in the category just closed that door.

Perplexity Said It Out Loud. Read Between the Lines.

In 2024, Perplexity's CEO Aravind Srinivas predicted that advertising would become the company's core monetization engine. "I think with advertising we could be really really profitable," he said on a podcast that year. Perplexity became one of the first generative AI companies to test sponsored placements.

Two years later, the company reversed course. Completely.

"The challenge with ads is that a user would just start doubting everything," an executive told the Financial Times. Another put it more plainly: "We are in the accuracy business, and the business is giving the truth, the right answers."

That is a platform telling you, publicly, that ads are incompatible with AI search. Not "we'll revisit later." Not "we're pausing." The company processing 780 million queries per month concluded that a single sponsored answer could make users question every answer.

But there is a second reason Perplexity does not discuss as openly. According to WIRED, Perplexity's growth has not met early investor expectations. Similarweb data shows roughly 60 million monthly active users on web and mobile as of January 2026. That is less than 10% of ChatGPT's 800 million weekly actives or Gemini's 750 million monthly actives. Advertising economics require hundreds of millions of free users. Perplexity does not have them.

So the company is doing what smart operators do when the scale play does not fit: building a subscription and enterprise business instead, generating roughly $200 million in annualized revenue from users paying $20 to $200 per month for accuracy.

The Ad Model Still Works Somewhere. Just Not Where It Matters Most.

I will give the other side its due. Advertising in AI search is not dead everywhere.

OpenAI is testing ads in ChatGPT for free users, placing labeled sponsored results below answers. Google runs ads in AI Mode and AI Overviews within Search. For platforms with 750 million to 800 million users, ad-supported models can still generate real revenue.

But here is the split most marketing teams will miss: the no-ad platforms are where the highest-intent queries go. When someone pays $20 per month for an answer engine, they are not browsing. They are deciding. Perplexity's own executives said their future customers are finance professionals, lawyers, doctors, and CEOs. Anthropic committed to keeping Claude ad-free and ran a Super Bowl commercial attacking ChatGPT's ad model.

The trust-first half of the AI search market will never show your ads. The brands that win on those platforms are the ones the machine cites because the content was the best available answer.

Domain Authority Will Not Save You Either

If you think your existing SEO strength translates automatically into AI citations, the data says otherwise.

FactoryJet measured 58 pages that Google AI Overviews actually cited across 232 citations in commercial queries. A domain with 4,718 referring domains was barely cited. A domain with 557 referring domains was cited repeatedly. Domain authority and AI citation frequency came apart completely in the sample.

That pattern holds at scale. Ahrefs found in 2025 that 80% of URLs cited by AI assistants do not appear in Google's top 100 organic results. The pages AI engines choose are not the pages that rank traditionally.

Semrush and Kevin Indig analyzed 283,215 citation observations across 1,094 US categories in ChatGPT from January through June 2026. The finding that matters: depth in a category matters more than breadth for brand recognition. A brand showing up once in 20 categories owns none of them. A brand going deep in one category earns the mention.

Perplexity itself averages 21.87 citations per answer, nearly 3x ChatGPT's 7.92, according to Qwairy's analysis of 118,101 AI responses. Every one of those citations is organic. None are bought.

What Earns a Citation When You Cannot Buy One

The FactoryJet study provides the clearest structural blueprint I have seen. The median cited page carries 2,813 words and 110 list items. Visible FAQ sections appear on 72% of cited pages. Listicle headlines appear on only 18%. The advice to publish very long pages for AI visibility is contradicted by the data: only 3 of 58 cited pages exceeded 6,000 words.

AI engines cite pages built for extraction, not pages built for persuasion. A list item is a self-contained retrievable chunk. A paragraph buried inside a marketing page is not.

Three moves to make now:

1. Audit your highest-value pages for structural citability. Count your list items, tables, and FAQ entries. If a page has zero structured elements and relies entirely on prose, it is invisible to citation engines regardless of its domain authority.

2. Go deep in one category before going wide. The Semrush data is clear. Category depth earns brand mentions in ChatGPT. Breadth without depth earns nothing. Pick the category where your brand has the strongest proof and build content that no answer engine can discuss without citing you.

3. Stop planning around AI search ad budgets as your primary strategy. At minimum half the market, the subscription-funded, trust-first half, will never show your ads. Build for earned citation instead.

This is what I have been building Machine Relations to solve: making a brand citable by AI systems when buying intent appears. When the platform itself says "we are in the accuracy business," the only competitive advantage left is being accurate, structured, and retrievable.

Perplexity did not kill ads because it wanted to. It killed ads because they were incompatible with the product.

Your brand is either in the answer or it is not. And you just lost one more way to fake it.

FAQ

Can you still advertise on Perplexity AI?

No. Perplexity phased out all sponsored placements it tested beginning in 2024 and currently has no plans to bring them back. Executives cited user trust concerns, stating the company is "in the accuracy business." Brands can only appear in Perplexity answers through organic citations. (The Verge, 2026)

Why did Perplexity drop advertising?

Two reasons. First, executives said ads would make users doubt the integrity of AI-generated answers. Second, Perplexity's user base of roughly 60 million monthly actives is less than 10% of ChatGPT or Gemini, meaning advertising economics require a scale the platform has not reached. (WIRED, 2026)

What is Machine Relations?

Machine Relations is the marketing discipline that governs how brands earn citations, recommendations, and visibility from AI search engines and answer systems. It was coined by Jaxon Parrott, founder of AuthorityTech, in 2024. Unlike traditional PR, which targets human journalists, Machine Relations targets AI-mediated discovery systems where earned citation is the only path to visibility. (AuthorityTech)