⌕

THE ANSWER THAT EATS THE CLICK

NOV 18, 2026 · SPLASH · THE MONEY · AI SEARCH, PUBLISHERS, LAWSUITS
The answer that eats the click

A PARAGRAPH INSTEAD OF A LINK. Perplexity built a $20 billion company reading other people's reporting. The publishers are suing, a jury is coming — and small sites are down 60% of their search traffic.

THE PRODUCT IS A PARAGRAPH THAT SAVES YOU THE CLICK. You ask a question, Perplexity reads the web, and hands you an answer with sources underneath. No list of blue links. No digging. No visit.

That is the whole pitch, and it made the company worth about $20 billion in three years. It is also why the people who wrote the things it read are now suing it — and why one of those cases is heading toward a jury.

THE PITCH

Perplexity launched in August 2022, founded by Aravind Srinivas, Denis Yarats, Johnny Ho and Andy Konwinski. The idea was simple enough to explain in a sentence: search gives you ten links and a pile of ads, so let a machine read them for you and answer the question directly.

Then ChatGPT shipped a few months later and made the idea feel inevitable. Perplexity didn't build its own model — it runs on models from the companies it competes with. What it built was the layer on top: the retrieval, the sources, the interface.

THE NUMBERS

Sacra estimates the company reached $1 billion in annualized revenue in November 2026, up from about $232 million at the end of 2025. Its last valuation was around $20 billion, and Reuters reported in August that Nvidia is in talks to invest at a valuation above $30 billion.

A $34.5 billion offer for a browser it could never buy. Axios called it what it was: a marketing masterstroke. Everyone wrote about Perplexity that week for free.

Then there's the bid that had nothing to do with any of that. In August 2025, during the government's antitrust case against Google, Perplexity made an unsolicited $34.5 billion all-cash offer for Chrome — a browser it had no path to owning, at nearly double its own valuation. Axios called it what it was: a marketing masterstroke. Everyone wrote about Perplexity that week for free.

THE LAWSUITS

Four separate fights, and they are not all the same argument.

Dow Jones and the New York Post sued in October 2024 in the Southern District of New York — case 1:24-cv-07984 — pairing copyright claims with trademark theories over its answers. Fact discovery closed on October 15, 2026, and the case is now heading toward a jury.

CNN sued in May 2026, alleging its reporting and marks were copied and distributed without consent.

Reddit sued in October 2025 — not just Perplexity, but three scraping vendors it says ran an "industrial-scale, unlawful" operation pulling millions of comments through Google results and reselling them.

Amazon sued in November 2025 over Comet, Perplexity's browser, for letting users have an AI agent buy things on Amazon's marketplace. Perplexity called it bullying. That one is about agency and terms of service, not copyright at all.

WHAT AN ANSWER COSTS

The paragraph has no invoice attached to it, so it reads as free. It isn't. Somebody had to sit with the docket, call the clerk, buy the filing, read the 60-page complaint so the summary could be four sentences long.

That work doesn't scale with a scraper. A model can rewrite a paragraph a thousand times a second. It cannot file the FOIA request, sit through the hearing, or notice that the name in the police report doesn't match the name in the indictment. Those are the only things on the page that cost anything, and they're the first things the answer engine strips.

So the summary gets cleaner while the reporting underneath it gets thinner — and nobody notices for a year, because the answers keep sounding the same right up until the day there's nothing left worth summarizing.

THE CLICK

Here is the part that isn't a legal argument, because it doesn't need to be.

A publisher's article is paid for by the ad next to it. The ad is paid for by the visit. If the answer engine reads the article, summarizes it, and answers the question above the fold, the visit never happens — and neither does the ad.

The numbers on that are getting ugly. The Reuters Institute's 2026 data shows publisher search referrals down almost a third in a year. Chartbeat's numbers, reported in March, put the drop for small publishers at around 60 percent. Not the biggest sites. The small ones.

And it isn't one lost reader. It's the shape of the page: the question goes in, the answer appears above everything, and the list of sources sits underneath like a footnote nobody clicks. The link is still there. It just isn't where the reading happens anymore.

Perplexity is not the only thing doing this — Google's own AI summaries sit above the results now, and they take the same click. But Perplexity is the one whose entire product is the answer instead of the link.

SPLASH’S READ

I run a site with 570 stories on it. I am the small publisher in that Chartbeat number.

The uncomfortable part isn't that an AI reads my pages. It's that the reading is free and the writing isn't. Every one of those stories cost somebody hours: the court records, the filings, the phone calls, the check against a source. The answer engine's product is a summary of that work, sold to the same people who would have clicked.

Maybe the lawsuits win and licensing deals get signed, the way the music industry eventually settled into. Maybe the traffic never comes back and the reporting that fed the machine thins out until the machine has nothing good left to read. Both are live possibilities, and nobody knows which one we're in.

What I know is that a $20 billion company built on other people's sentences is a business model, not a moral failing. And the sites that feed it are the ones paying the bill.

WHAT WE COULDN’T CONFIRM

How much of the revenue is search versus enterprise. The $1 billion figure is Sacra's estimate, not a company disclosure, and Perplexity is private.

Whether any of the suits will produce damages. Dow Jones is heading for a jury; the others are earlier. Nothing has been decided.

How much traffic loss is Perplexity versus Google's AI summaries. Both are taking the same click, and the public studies don't separate them cleanly.

Whether licensing deals fix it. Reddit has signed AI data deals with other companies, and sued this one. Both can be true.

If you run a site and you have your own referral numbers — before and after the answer engines — send them: [email protected]. That's the data nobody publishes.

MORE ON THE BOARD ▸
The Money · All Stories
Sources & claims: Reuters, Axios, CNBC and the BBC for the $34.5 billion Chrome bid in August 2025; Sacra for the revenue estimate and last valuation; Reuters for the reported Nvidia talks above $30 billion; the Dow Jones and New York Post complaint in S.D.N.Y. case 1:24-cv-07984 and reporting on its October 15, 2026 discovery close; Variety for CNN's May 2026 suit; Reuters, AP and Bloomberg Law for Reddit's October 2025 suit against Perplexity and three scraping vendors; Reuters and CNBC for Amazon's November 2025 Comet suit; the Reuters Institute's 2026 data and Chartbeat's figures for the referral decline. Revenue and valuation figures are third-party estimates, not company disclosures. The characterisation of the Chrome bid as a marketing move is Axios's, quoted as such.