To The Moon

Before I get to the SpaceX IPO, I’m going to be doing my regular State of the Portfolio series starting tomorrow. I hope this helps show what I do, and don’t, like about the stocks in the Asymmetric Portfolio today.

After a rough start to 2026, my regular reviews have me liking the long-term trajectory of most companies, and valuations are compelling, but that doesn’t mean the market agrees with me.

That will be going to premium subscribers and will be completed in a series of articles over the next few days.

Now, on with my thoughts on SpaceX’s IPO.

The SpaceX IPO

$SpaceX(SPCX)$ is going public on Friday with a stock price of $135 per share and a market cap of $1.77 trillion.

To be honest, this is one of the wildest IPOs I’ve ever seen. The debate around the stock is crazy. The valuation is crazy. And the comments from management are crazy!

But what did we expect? This is Elon Musk!

As I looked through the S-1, it’s one of the more surprising documents I’ve ever looked at.

What management has made clear is that space — aka launching rockets into space, SpaceX’s core competency — is a small business and a small part of the company’s future. Management sees connectivity as a bigger business than space, and AI stuff is an order of magnitude bigger than that.

In fact, enterprise applications — which don’t exist from SpaceX — are seen as a $22.7 trillion opportunity!!!

How could this possibly be the case?

This clip shows the kind of thinking investors have to go through for this to make sense. Elon Musk literally says they could launch 100 GW of AI compute per year into space in the near future.

For perspective, the entire U.S. uses about 466 GW of power at any given moment, and all AI data centers in the world today are about 30-40 GW, depending on the estimate.

So, Elon Musk thinks they can figure out how to build space-based compute, launch it at scale, get enterprises to pay for this compute to the tune of hundreds of billions of dollars per year (1 GW of data centers is about $60 billion on earth), and all of this will be done within a few years.

If you can’t sense my skepticism…I’M SKEPTICAL!

But this is the Elon Musk playbook. Promise a grandiose vision and get people to buy in. Like FSD, you never really have to deliver if another shiny object comes along that’s an even bigger prize.

SpaceX Isn’t Really a Space Company

So, management thinks the future is enterprise AI, and they’re not even pretending space is a big deal.

What’s wild about the space business is how small it is, even today, for SpaceX.

In 2025, the company launched 165 rockets, generating $4.1 billion in revenue. That’s a good business, but it’s only about $25 million per flight, and to grow, there needs to be more rockets launched into space, presumably to get the AI data centers up there.

So far this year, revenue is actually down, and the space segment lost $662 million.

Space could be a nice business, but it’s not a profitable business, and it’s not a big revenue generator, at least for a $1.77 trillion company!

Starlink is the money-maker for SpaceX. In the last three months alone, Starlink made $3.3 billion in revenue and $1.2 billion in operating income.

The business is profitable, and it’s growing. I have to wonder how durable it is given competition from $AST SpaceMobile, Inc.(ASTS)$ $Amazon.com(AMZN)$, and others, but right now, this is the financial bright spot.

And here’s where things get interesting.

AI, which includes the old Twitter and xAI, which Musk created and is known for its data centers and Grok, is barely generating any revenue and lost $2.5 billion in Q1 alone!

This will change shortly because Anthropic signed a deal to pay SpaceX $1.25 billion per month for compute from the Colosus 1 data center, and Google will pay $920 per month for compute starting in October of this year.

In other words, SpaceX is billing itself as a consumer and enterprise AI company, but operationally, it’s more like a neocloud, building compute and renting it out.

That’s a great business to be in today, but will it be forever? Will it be cost-effective (or possible) to launch data centers into space and rent them out?

We also need to consider that SpaceX is planning to build the Terafab foundry in Texas, and that will cost as much as $119 billion. Those chips will be used for AI and theoretically will be rented out as well.

Does SpaceX want to be $CoreWeave, Inc.(CRWV)$ ? Or $Taiwan Semiconductor Manufacturing(TSM)$ ? Or OpenAI? Or some kind of enterprise AI software layer?

It’s not really clear.

What we know is that SpaceX’s value relies more on AI than it does on space. That’s crazy to think about, given the company’s operations and losses in AI, but it’s true.

The Balance Sheet Is a Mess (for now)

The other surprising note is that SpaceX holds $29 billion in debt, primarily related to the xAI acquisition and the debt that came with buying Twitter.

On the right, you can see the pro forma cash expected after the IPO.

But that cash is going to be gone soon!

I mentioned the Terafab plans above, and that’s on top of the already expensive capex being put in the ground today. In the last three months alone, SpaceX had $1.0 billion in operating cash flow and $10.1 billion in capex, and capex is climbing fast!

This is not a company we’re going to see generating free cash flow anytime soon. In fact, I would be shocked if SpaceX isn’t back selling stock to the public in a matter of months.

It needs the money!

Competition Is Fierce

I don’t like SpaceX’s financials, but the competitive position is arguably worse.

None of the company’s business segments has a particularly compelling moat or the financials to suggest there’s great operating leverage on the horizon.

Space

I hear people say that SpaceX has a monopoly on space, which is just wrong. There’s no monopoly and no real moat on launching things into space. Once a company figures out how to send rockets to space repeatably, it can continue to do so.

In fact, Rocket Lab is growing extremely quickly, launching 22 rockets in the past year.

Jeff Bezos’ Blue Origin is only launching a few times a year, but it’s also in the game.

There’s another company in India launching satellites like AST’s.

Bottom line, there’s competition and no real moat around space.

Connectivity

There’s no doubt that connectivity is the biggest business today, but even when management is selling the business in presentations, you can see the limitation. Once every airplane, tractor, boat, and remote home is connected, growth will stall.

That’s because a satellite connection is fundamentally slower than a wired connection (fiber or cable at home/work) or local 5G on land.

The speed of light is working against satellites.

On top of that, competition is ramping up. AST Spacemobile is launching a full constellation and has telecom partners, and Amazon is getting into the business. A decade from now, what’s the argument that the economics of space connectivity is much better than terrestrial connectivity…which is a rough business to be in?

AI

I have an even harder time seeing SpaceX being anywhere near a leader in AI.

Grok is, at best, the 5th or 6th best AI model.

Twitter is, at best, the 4th or 5th most valuable social network for training data.

All of the founders of xAI have left, except Musk, and yet this is the future of the business.

I don’t see the differentiation or why the AI business will be extremely valuable outside of the willingness to skirt zoning and environmental laws more brazenly than competitors.

Count me skeptical that SpaceX will live up to its 100x sales valuation.

This isn’t going to go into the Asymmetric Portfolio. And I hope investors buying the IPO go in with eyes wide open about what they are, and aren’t, buying.

😍 Been eyeing Tiger merch but short on Tiger Coins? Now's your chance.

🎁 We’ve selected 4 high-demand items across practial, lifestyle, and learning, now with a lower redemption threshold!

Hot Merch Returns · Up to 43% Off