The SpaceX IPO: Why It Hasn't Happened Yet

Elon Musk once tweeted that taking Starlink public too soon would be, quote, very painful. For twenty years, the answer from SpaceX was always no. No bell. No ticker. No quarterly earnings call. June twelfth, two thousand twenty-six. Nasdaq. Ticker S-P-C-X. The largest initial public offering in history.

Wait, he actually did it? After all that?

He did it. And that's the question I want hanging over this whole thing... what finally changed his mind? Because the man had every reason not to.

Okay, but he'd been dangling this for years, right? I remember the email.

May nineteenth, two thousand nineteen. He writes to staff that it'll probably make sense to take Starlink public in about three years or so.

Three years. So, two thousand twenty-two.

Then twenty twenty-one arrives, and at an all-hands he says... I'm not sure exactly when that is, but maybe it will be like, I don't know, just guessing, three or four years from now.

So the three years is permanently three years away. It's a horizon.

And every time, he adds a condition. Starlink has to be in a smooth sailing situation. Good predictability. Hold onto that word. Predictability. It's the key to the whole episode.

So what flipped? Because he had all the money in the world privately. What's the actual mechanism?

The mechanism is a dish on somebody's roof. Everyone thinks SpaceX is a rocket company. Follow the revenue and it just... isn't. Two thousand twenty-three, Starlink had two point three million paying customers. End of twenty-four, four point four million. End of twenty-five, eight point nine million.

That's doubling. Every single year.

By the end of March this year, ten point three million subscribers across a hundred and fifty-five countries. By June, more than twelve million.

Twelve million people paying a rocket company a monthly bill. That's a utility.

And here's what makes an analyst sit up. Once the satellites are up there, each new subscriber costs almost nothing to add. The margins stop looking like hardware... and start looking like software.

So the rockets are the boring part?

The filing is brutal about it. Twenty twenty-five, Starlink generated eleven point four billion dollars out of eighteen point seven billion total. The space division, the one that dominates global launch, about four point one billion. First quarter this year, Starlink did three point three billion of four point seven. Launch made six hundred and nineteen million.

The most famous rockets on Earth are the side hustle.

And connectivity was the only division that turned a profit.

Then why wait at all? If it's that good, why not list in twenty-two like he said?

Because he didn't need to. This is what people miss about modern private companies. SpaceX found a way to pay its employees without ever ringing a bell. It's called an insider tender offer. The company lets staff and early investors sell existing shares to new buyers, at a price the company sets.

So... a private stock market. With one stock on it.

Exactly that. And watch the marks. December twenty twenty-four, a tender values it at three hundred and fifty billion. July twenty twenty-five, four hundred billion, at two hundred and twelve dollars a share. December twenty twenty-five, the Wall Street Journal reports a sale above four hundred dollars a share. Eight hundred billion.

It doubled in five months with no public market at all.

Same trick the big AI labs use. OpenAI, Anthropic. Private markets got deep enough that going public became optional.

And the downside is obvious. Quarterly earnings, lawyers, people yelling at you.

And Musk knows that world. Two thousand eighteen, he tweets he has funding secured to take Tesla private. He ends up settling fraud charges with the S-E-C and paying millions. More recently he's called the firms that advise big shareholders corporate terrorists.

So he genuinely hates the apparatus.

And the deeper fear was that public investors would never fund Mars. You can't put a payback date on a city on another planet. Shareholders want a number in ninety days.

So how does he square that?

He goes public without giving up control. The prospectus, filed confidentially on April first, confirms a dual-class structure. Musk holds roughly forty-two percent of the equity... and about seventy-nine percent of the votes.

Less than half the company. Four fifths of the power.

Super-voting shares. You buy in, you get one vote. He holds a different class. Reuters noted it reopened one of Wall Street's oldest fights about corporate governance.

So retail investors are, what, invited guests?

Invited in big numbers, though. Reporting around the deal pointed to an unusually large retail allocation, around thirty percent, at a valuation near one point seven five trillion dollars, raising up to seventy-five billion.

Because for years the only way in was those weird closed-end funds trading at insane premiums to what they actually owned.

Proxies. People bought anything orbit-adjacent because they couldn't buy the real thing. That premium is the tell. It measures how badly ordinary people wanted access to a company that had decided it didn't need them.

So back to the loop. What actually changed his mind?

Predictability. His own word. Twelve million people paying every month is predictable in a way rockets never are. He didn't lose the argument. He waited until the subscription business was big enough that Wall Street had to take Mars as part of the bundle.

The broadband bill pays for the spaceship.

And that's the thing sitting in your own budget right now. The most valuable asset a company can own isn't a spectacular product. It's your standing order. The payment you've stopped noticing.

Which explains why literally everything I own is a subscription now.

He built the most powerful rockets on Earth. The thing that finally dragged him to Wall Street... was a direct debit.