SpaceX stock (SPCX): what are you paying $2 trillion for?

0:00 Everyone's hyped. Nobody can say why.

Everyone's hyped about SpaceX. Nobody can quite say why. So that's the question: what's the hype about? And by the end of this, you'll know exactly what you're paying two trillion dollars for. Three stops: what they promise, what actually pays the bills, and what it costs you.

0:20 What SpaceX actually does

But first. What are we actually buying? Quick version. They build rockets. Lots of them. Last year they launched more than the rest of the planet combined. The big one, Starship, reached orbit for the first time in September. They run Starlink. Internet from space. Twelve million subscribers.

And since February: an AI business. Data centers they rent out, and a chatbot, Grok. That part already changed its name once this year. They went public in June. Priced at 135. Hit 211 on day three. 108 in August. Now around 160. And Elon Musk controls 46 percent of the shares. And 82 percent of the votes. Their own prospectus: you get no say. Page 61.

So. What's everyone so excited about? Here's the thing. People think they're buying a rocket company. Starship reached orbit for the first time. Historic. The stock went down that day. Starlink doubled its subscribers. Nobody noticed. It was in the same report as the AI bill.

So what moves it? AI. Take away the IPO week, and the four biggest days since were all AI. AI costs: down. AI hype: up. And the CEO agrees. Quote: 'SpaceX is a super intelligence company.' For those who missed it: 'SI' is the new 'AI'. On September 29th: the White House ordered every federal agency to stop saying 'AI'. Because 'artificial' sounds fake.

Turns out it's not just women. He grabs technocrats by the kitten too. Hahaha. So the rocket is the brochure. What people are buying is an AI company.

1:53 Hype check

Hype check. Same formula as last time. The market says this company is worth 2.18 trillion dollars. What they own, minus what they owe: 127 billion. Out of 2,180. And 86 of those 127 billion is the IPO cash. So two thirds of what's real... is your own money. So 94 percent of the price is hype. So what are you paying two trillion for? Next year.

Five times. On one call. Where have I heard that before? Remember the robotaxi? 2019: next year. 2020: next year. This year: still next year. Eleven years in a row. Hahaha. Let's make it twelve.

2:36 The promises: "next year"

First stop: the prospectus. The sales document SpaceX had to file before the IPO. Written by them. The one you're supposed to read before you buy. The largest market in human history. Their words. 28 and a half trillion dollars. Musk says a trillion dollars in revenue by 2030. Last year they did 19 billion.

Fifty-three times. In four years. How? Same call: a hundred billion by December, 'if we did nothing'. Five times the computers by the end of next year. Then the satellites. Then the Moon. The data centers in space. In January: the cheapest place for AI, in two to three years. We'll get back to that. In June: as early as 2028. In August: next year. Next year. That's twelve.

It got closer the more they talked about it. Starship: at least one flight a day. A year from now. So that's the biggest market ever, from the prospectus. And from the same man on the same call: a trillion in revenue, data centers launching next year, and a rocket a day. That's your two trillion, right there. Let's go check how much of it exists yet.

3:44 Follow the money

Where does the money actually come from? Last quarter: 7.8 billion in revenue. Starlink is more than half. AI a third. The rockets, twelve percent. And profit? Starlink makes money. The rockets lose money. AI loses more. So the boring part pays for everything. Remember the board? Starlink was the only train actually boarding. The AI business spends about 174 million dollars a day. On building.

Starlink's profit, after its own bills, covers the first two days of the month. Then it's loans, and new shareholders. Every month. Now, the AI business itself. Most of the money comes from renting out computers to other AI companies. Anthropic. Google. The renting actually makes money. What eats it is the research. Two billion a quarter. Mainly, in their own words, training Grok.

Grok has 117 million users. One point nine million pay. That's one point six percent. So the part that works is renting out chips by the hour. The part that costs is the chatbot. And the renting is mostly one customer. Anthropic. Guaranteed three months at a time. Of 75 billion in deals, 14 is actually binding.

So: Starlink pays two days a month, somebody else pays the rest, and the income is one customer on three months' notice. That's your two trillion, today. You're paying for next year.

5:06 The claims, one at a time

So. Next year. One claim at a time. 'The largest market in human history.' Twenty-eight and a half trillion. TAM means: if everyone who could buy, bought everything. The ceiling. Not a forecast. And look how they built it. Space: one percent. Twenty-two trillion is, in their words, 'the digital economy'. Everything sold on the internet. Plus every human over the age of ten, paying twelve dollars a month.

By that math, this channel is on the internet too. So my TAM is twenty-two trillion. Congratulations to me. And that's the modest version. Asteroid mining isn't even in it. Big number good. Small number bad.

Claim two. A trillion a year by 2030. Today Anthropic pays about one and a quarter billion a month for 325 thousand chips. A trillion a year is about 22 million chips. Rented out. At today's price. Right now they rent out about 435 thousand. So where do 22 million chips go? Fifty times more. In four years. No problem. Claim three.

Okay, this next part is the one that matters. Give me three minutes. Because this is the bet. Your retirement, your house, your kid's college fund, or, if you were smart, a small slice of it. It's riding on this one. Claim three. Data centers in space. The idea: put the chips on satellites, and let the sun power them for free.

Google wants to do the same thing, and they published the math. Problem one: the ride. One chip costs thirty thousand dollars and lasts about five years. That's seventy cents an hour. On Earth you add the electricity. Ten cents. Eighty cents an hour.

In space the sun is free. The ride up is not. At today's rocket prices, spread over those five years, the ride is two dollars an hour. Almost three dollars, all in. You're launching a thirty-thousand-dollar chip into orbit to save ten cents an hour. Hahaha. For space to break even, the ride has to get eighteen times cheaper. Google's own estimate for when that happens: the mid-2030s.

And our 22 million chips? With their own satellite weights, about 300 thousand tons. Around 2,500 Starship flights. To have them up by 2030, that's two launches a day. Every day. Starting now.

So with today's rockets, the math doesn't work. SpaceX's bet is Starship. Most of a rocket's cost is developing it, and that money is already spent. Spread it over thousands of flights, reuse the same rocket a hundred times, and the ride gets cheap. Starship reached orbit for the first time in September. Three flights this year. The upper stage has never flown twice.

Problem two: heat, and sun. Each satellite is mostly panels. Solar panels to catch the sun, and radiators to glow the heat away, because in space there's no air to blow over the chips. The chips are the small part.

For one big data center, about 700 thousand chips, the panels cover two Central Parks. For our 22 million chips: four Manhattans. SpaceX's plan is a hundred data centers a year. Fourteen Manhattans. Every year. Hahaha.

And all of that weighs something. Panels, batteries, radiation shielding, lasers to send the answers back down. The NASA physicist who did this math gets three to six times the weight SpaceX assumes. So take those two launches a day, and multiply.

Add it up. One big data center on Earth: about 38 billion dollars. The same thing in space, by the people who've done the math: three to ten times that. Just the rocket rides, at today's prices, cost as much as the whole data center on the ground. Google's own project lead: not cheaper within five years.

Problem three: nothing lasts. Their own books: the chips are written off in five to six years. The satellites carrying them, three to five. And nobody can go up and fix them. A rocket a day. Right now it's about one every two months. And that's the step everything else stands on. It's a plan. A plan that lands somewhere in the 2030s. So when does this train actually leave?

9:10 The ticket: are we late?

So. Is the hype about something real? Rockets: yes. Satellites: yes. Chips in orbit: four. Google's. On a SpaceX rocket. It's just not arriving anytime soon. The one train leaving today is Starlink. Everything else is on the board for the 2030s. And the ticket price is today's. Ninety-five times the last twelve months of sales.

Ninety-five years of today's sales to pay back the price. If they spend nothing. Ever. Even at the hundred billion a year Elon says they'll hit in December: 2048. From profit? There isn't any yet. So that train doesn't have a time. So no, we're not late. We're early. Which sounds great. Early just means you wait longer. And pay for the waiting.

And while we wait, somebody has to keep the lights on.

9:55 The AI bubble, simplified

You've probably heard the words 'AI bubble'. Here's what they actually mean. Simplified. Raise money. Buy chips. Build a data center. Rent it out. The chips cost a fortune, and they're written off in five years. So you can't wait for customers to pay for them. You borrow, or you sell shares. You buy more chips. Round and round.

Real money, from customers, is the thin line. For SpaceX it's about four cents of every dollar they spend on AI this year. The rest is loans, new shareholders, and profits from other businesses. Starlink. And it's not just SpaceX. Across the whole industry, for every dollar a customer pays for AI right now, about four to five dollars are being spent building for it.

That's fine, as long as everyone believes the customers are coming. The loans get renewed. The shares go up. More chips. If the customers don't come fast enough? The chips still lose value every year. The loans still come due. That's what a bubble is: a lot of building, paid for with borrowed money and belief, before the customers show up.

One more thing. The money goes in a circle. The chip maker invests in the companies that buy its chips. In SpaceX's case, Nvidia helps fund the fund that buys Nvidia chips. To be fair: the gap is closing fast. Last year it was fifteen to one. This year four or five. The customers are catching up.

The question is whether they catch up before the lenders want their money back. Because if they don't... who pays?

11:15 Who pays the bill

Say the customers never pay enough. Who's left holding the bill? First, the shareholders. They lose first. That's the deal. Then the people who lent them money. 25 billion in bonds. Then the lenders with collateral. They get the chips back. Chips that are worth less every year. And who are the lenders? Bond funds. Pension funds. Insurance companies. The bonds are investment grade, so they're allowed to buy them.

If you own the big Nasdaq fund, three percent of it is SpaceX. You didn't pick it. That's where the bill ends up. We've done this before. Telecom, 2002. Fiber everywhere, built on loans. The bonds that defaulted were worth about twelve cents on the dollar. The fiber is still in use today. A GPU is written off in five.

To be fair again: SpaceX has 100 billion in cash. Against 39 in debt. If this goes wrong today, the shareholders pay. Not the lenders. That changes if the cash runs out. They burned 25 billion in the first half of the year. At that pace, the cushion lasts about a year. Which is why they're asking for 40 billion more. So. Knowing all that... would I still buy it?

12:19 What's real vs what can go wrong

That's the roast, now the toast. Let's be fair to both sides. On the left: what's real. On the right: what can go wrong. Starlink works and keeps doubling. Nobody launches more rockets, and Starship just reached orbit. The demand for computing is real: the competition is sold out. They're sitting on a hundred billion in cash. And AI customers pay about five times what they did a year ago.

One customer on ninety days' notice. The AI business loses money, and Grok eats the rest. Ninety-five times sales, no profit. Burning twenty-five billion a half-year on chips they write off in five. Everything waits for Starship. More shares unlock after the next report. And one man with 82 percent of the votes. And if it fails? There's always next year. Hahaha.

The left side is why people are excited. The right side is what they're paying 95 times sales to ignore. So. What are you paying two trillion for? The left column is real, and the books say it's worth about 130 billion. The other two trillion is the claims from the lab. Delivered sometime in the 2030s. That's the whole lesson. Potential isn't the problem. The price of potential is.

So, would I still buy it? Yes. Nobody buying this cares about any of it. That's the hype wave, and I'm riding it. But first: who's making the wave?

13:38 The headline factory

Since the IPO, these sites put out over three thousand SpaceX headlines. One an hour. Day and night. Seven times. One of them came out on the day. Same page. The ad says SpaceX made their top ten. The article says sell. And they asked SpaceX's own chatbot which SpaceX funds to buy.

14:00 The finfluencers

And the finfluencers. Fifty-eight videos. Surely one of them covers the risks. Ah, finally. They do mention risk. The risk of not investing. Hahaha. And then, social media. All you wanted was a little doomscrolling before bed. So. What can we learn from the small investors online? The rule in here? Down, buy more. Up, you should have bought more.

If you get jealous, and you want to be a billionaire too, you can be one. Just in time for Christmas. It's easy. You buy, it dips, you buy more. Every dip you miss, you'll regret forever. Anyone who disagrees, you block. Hahaha.

14:48 My plan

So what's my plan? This stock doesn't move on rockets. Starship reached orbit and it went down two percent. It moves on promises. And almost every promise says 'next year'. That's 2027. The same year the cash runs low and they need more. So I'm not waiting for next year to show up. If I buy, I ride the hype, and I'm out sometime in 2027.

Okay. Let's go buy some.

15:11 The Hype Fund

Quick check on The Hype Fund. Last time, I bought Micron at 1,084. Down. In hindsight, I caught a local top. But the plan says 2027, not October. No panic. And today, SpaceX joins the fund.

15:27 A toast

One more thing. I stand by every roast in this video.

But every road, every wire, every light in the sky… started as an idea somebody laughed at. If the world were full of people like me, we'd still be sitting in a cave. Laughing at the guy with the fire. So here's to the dreamers. The Wright brothers. Jules Verne. Elon Musk.

They build the future. I check the deadlines. Because when the deadlines get pushed back, the stock gets pushed down. And I still bought a ticket.

Like, comment, subscribe. To the moon!