How Palantir Was Founded

Palo Alto, two thousand three. A man who has just sold PayPal to eBay starts another company. And there are basically two sources of money in it. His own... and the Central Intelligence Agency.

The CIA. As in, an investor? Writing cheques like a venture fund?

Through its venture arm, yes. And then he names the company after a magic rock from Lord of the Rings.

Hang on. A magic rock.

The palantiri. The seeing-stones. Peter Thiel named the company after the stones in Tolkien that let whoever holds one see across great distances. For him, the name meant seeing hidden patterns across immense volumes of data.

Okay, but this is the bit I always snag on. In the books, those stones are a trap. You look in... and the thing on the other end starts working on you.

You're not the first to notice. The Ringer makes exactly that point — that Tolkien treated the seeing-stones as a parable about the dangers of surveillance technology. Thiel picked it anyway. Hold onto that, because at the end I want to ask you whether that was a blind spot... or the whole point.

Fine. So who else is in the room at the start?

Thiel incorporates Palantir in two thousand three. He brings in his Stanford Law School colleague Alex Karp as chief executive, a PayPal engineer called Nathan Gettings, and two Stanford students, Joe Lonsdale and Stephen Cohen. Five founders. Three of them barely out of school.

And Karp's the law school friend. Were they allies back then?

Kind of the opposite. Karp grew up outside Philadelphia, son of a Jewish father and a Black mother, never quite fit in at high school or at Haverford College. Then Stanford Law, where he met Thiel and argued loudly about politics. The arch-libertarian's pro-market philosophy against Karp's then-Marxist outlook.

A Marxist. Running the CIA's favourite software company.

Then-Marxist. And before Palantir he wasn't even in tech. Haverford degree, a law doctorate from Stanford, then a PhD in social theory in Frankfurt, in Germany. He's not a coder. He's a philosophy guy who ends up selling to spies.

So where does the actual product come from? You don't get from German social theory to counterterrorism software.

You get there from credit card fraud. PayPal was being eaten alive by fraudsters, and the trick they landed on wasn't pure automation. It was software that surfaced suspicious patterns, then let a human analyst look and judge.

Machine finds the weird thing, person decides if it's actually weird.

Exactly that. And after September eleventh, two thousand one, the pitch was: take the fraud trick, point it at terrorism, and protect civil liberties while you do it. Lonsdale later wrote they were, quote, very frustrated watching the government waste billions of dollars on technology that was way behind Silicon Valley and was not protecting civil liberties.

Which is a very convenient origin story for a company that later gets accused of building a surveillance machine.

It is. Although — normal venture capitalists wouldn't touch this thing. National security was a market they didn't understand, with unclear returns. So Thiel put in his own money. And In-Q-Tel, the CIA's venture fund, put in some of theirs.

How much are we talking?

Around two thousand five, In-Q-Tel began investing roughly two million dollars across multiple stages. You'll also see it reported lower — one timeline puts the seed at one point two five million dollars for equity. The exact figure is murky. Both versions are tiny.

Tiny for a company now worth... what, exactly?

Hundreds of billions, as of twenty twenty-six. But the money was never the point. The CIA helped Palantir refine Gotham, its first product, and became its first significant customer. In those early years, apart from Thiel, the agency was essentially the only investor in the business.

So the customer and the investor are the same organisation. That's... not a normal arrangement.

It's the single fact critics come back to, twenty years on. Supporters say that's how serious defence technology always gets built — somebody inside has to want it. Critics say it blurs the line between a vendor and the state it's meant to serve.

And concretely, what is an analyst doing with it? What's on the screen?

Picture a desk with five databases that don't speak to each other. Phone records here, travel manifests there, bank transfers somewhere else. Palantir's pitch: put them on one screen, link the same person across all of them, and see in minutes what used to take weeks.

Genuinely useful. And also exactly the thing you'd be nervous about if the data is about you.

Both true at the same time. That's why this company is so hard to argue about.

Was the business slow, then? Spies aren't a big market.

Painfully slow. Years of selling to agencies, no commercial customers. The break comes in two thousand nine, when JPMorgan Chase buys the software to detect fraud — the idea coming full circle, PayPal to Wall Street. Lonsdale leaves around then, keeping his stake. By twenty ten, Palantir is valued at a billion dollars.

One thing still nags at me. You said two thousand three. I've seen two thousand four.

Britannica says two thousand three, the incorporation. Some financial write-ups say two thousand four, when it really got going. For a company built on getting the facts straight, its own birthday has two answers.

Okay. So — the Tolkien name. Blind spot, or the point?

I think the point. Nothing here was hidden. A name taken from a story about the danger of seeing too much. A fraud trick from a payments company. A philosopher in charge. And a cheque from the intelligence community. Every fight people are having right now — policing, immigration data, war software, the share price — was already sitting there in two thousand three.

Written on the box.

Written on the box. They told us they were building a seeing-stone. We just never read what happens to the people who look into one.