SpaceX IPO Valuation: Am I Missing Something?

Disclaimer: All analysis here is speculation and speculation only. Written for fun and intellectual curiosity. Not financial advice. Not an accusation of anyone, any organization, or anything. Just someone doing math on a Sunday and asking questions when there is nothing better to do.

So, let's get started

The Approaching IPO

It is Sunday noon, and I am catching up on news from last week. The sky is blue and the sunshine is warm. I have done all my Sunday routines, and am enjoying my second cup of coffee. You know, life in the valley can be enjoyable once in a while, especially when all the news is reading: bright future, big IPO, stock will go up up up.

That is the general vibe this week. SpaceX is going public. The roadshow launched on June 4. Pricing is coming on June 11. Trading starts June 12. Morgan Stanley, one of the most respected names in finance, shared its model with top investors on the roadshow: SpaceX revenue could reach $3.4 trillion by 2040, with adjusted EBITDA topping $2.7 trillion in the same year. Goldman Sachs, the other lead underwriter, projects over $470 billion by 2030.

Both banks agree on the near term: approximately $160 billion by 2028, which would be roughly 8.5 times what SpaceX reported in 2025. I have talked about this IPO in the past. It will be the largest in the history, bigger than anything that has happened in the past. Unlike Saudi Aramco, which sort of locked out retail (0.5% allocation), this time, Elon is inviting retail to join the party with 30% allocation.

The party is staged and now two banks are saying, it will be marvelous.

I almost opened my E*TRADE account right then, thinking about 8.5 times in 2 years. Just imagine

Then I did the math. Well, can't help.

The Pause

Bottom line: From $18.7 billion in 2025 to $3.4 trillion in 2040. That is 181 times growth in 15 years, or 41.3% CAGR.

For fifteen consecutive years.

I checked the number twice before finishing my coffee. Now I have both hands back and I can think clearer. Don't challenge me on this logic, because if I am logical, I could have stopped right here.

Why?

No company in the modern era, operating in any market, has sustained 41% annual growth at this scale. Not Nvidia, whose best 15-year CAGR during the greatest semiconductor demand cycle in history came in around 35%. Not Apple, the most valuable consumer brand ever built, at roughly 20%. Not Amazon, which invented cloud computing and the infrastructure of modern e-commerce, at around 28%. Not Google, which held a monopoly on search for two decades, at around 25%.

None of them is even close to 41.5% for 15 years.

Well, maybe Alexander the Great and Genghis Khan could have done it, if they were smart enough to remember to incorporate.

So here is where I want to stop and be honest with you. If you are smart enough to stop listening to a nobody mumbling, read the Morgan Stanley number and have a good Sunday afternoon is what I suggest.

If you thought about Nvidia at $10 and Apple when it launched the iPhone and you missed both and you are not going to miss this one, and you want to trust Morgan Stanley, now is the time. If you have a Fidelity account with $2,000 or more sitting in it. Then stop reading right now. Genuinely. The IPO is open to retail investors with as little as $2,000, which is something unusual. Log in. Submit your indication of interest. The train is arriving at the station on June 12 under the ticker SPCX on Nasdaq.

I mean that. If you believe the number, go.

But if you are still here. Then hear me out: I am genuinely confused, and I think I must be missing something. Maybe you could tell me how I am wrong by finishing the article.

The Second Disclosure

At this point, I am showing my hesitation against Morgan Stanley. I think that requires some clarification before we move further.

Morgan Stanley is an indisputable giant in the industry. It is where I trade stock and hold my IRA, or at least I think so. The number is dwarfed by any measure, so as a result, I wasn't that confident either. For now, let's assume my wealth, or rather poor, is managed by them.

Morgan Stanley is also the lead underwriter for this IPO. They are also the stabilization agent, which means they manage price stability after the offering. They are also managing the SpaceX direct share program for employees. And they own Shareworks, the platform where SpaceX employees hold their stock and options. And they own E*TRADE, where I almost just logged in.

Every door leads to Morgan Stanley.

So basically I am saying, even a nobody like me, I trust them completely.

The Industry

When I looked at the 41% number, I'm like...So. Which industry is SpaceX actually in? It must be an industry that grows faster than 41% in next 15 years.

The roadshow deck calls it three things: Space, Connectivity, and AI. A $6 trillion total addressable market transitioned itself to $26.5 trillion once AI is in the picture. The presentation describes SpaceX as a "N of 1." The only company building integrated hardware and software infrastructure across all three.

That is a beautiful perspective, worth digging in a bit.

The AI business

SpaceX acquired xAI in February 2026, which brought the Colossus data center in Memphis under its umbrella. The data center has 220,000 Nvidia GPUs, and over 300 megawatts of power. This piece of capital spending is bringing in revenue.

Last month, Elon shook hands with Dario and signed a lease of $1.25 billion/month through 2029. I wrote about it.

Then, right after the roadshow, Google signed a deal on June 5 to pay $920 million per month for roughly 110,000 GPUs through June 2029. Together, those two contracts represent approximately $25 billion in annualized revenue, and also maybe close to 100% of their GPU usage (this part is guesswork, given Google's usage and the fact that Anthropic paid higher price).

So what happen to xAI? Does it still run?

When a company builds a massive AI training cluster, there is one reason you lease it to competitors: you cannot fill it yourself. xAI was burning $1 billion a month prior to the lease agreements. Elon Musk publicly admitted xAI was not built right the first time. The founding team had already left. Grok, xAI's flagship model, has not appeared at the top of a benchmark in some time. Its most memorable cultural moment involved generating images that removed people's clothing. And now, almost 100% of the compute is sent to make revenue.

The GPUs that were supposed to power the world's most truth-seeking AI are now rented to Anthropic and Google. Two companies that are actively building models to compete with Grok. On hardware that Nvidia will likely make look slow within three years. Under contracts that both customers can exit after December 2026 with 90 days notice.

That start to sound like a data center landlord with a short-term lease and depreciating assets.

Are we missing something here?

The connectivity business

Starlink is real. I want to be clear about that. It is the first genuinely global low-latency satellite internet network. This is an actual business, built by actual engineers doing genuinely difficult things.

But is it a business that grows at 41.5% for 15 years?

I wrote about this before, and let me quote myself:

Starlink is the only real cash machine. Genuine, impressive, $12 billion in 2025 revenue. But Amazon's Kuiper just launched. OneWeb is scaling. Chinese constellations are coming. France, Japan, India, and many other countries are also building their own versions.

And Japan's SoftBank isn't just building another constellation. They are launching stratospheric airships this year, floating base stations at 20km altitude with 0.3ms latency, connecting directly to existing smartphones with no dish required. With a very different (much lower) cost structure, it actually might bring new usage cases for 5G connections. Where Starlink struggles with high-density urban capacity, HAPS thrives. These technologies may not compete for the same users today, but when they meet in the middle, HAPS might form Starlink's growth ceiling.

First-mover advantage is real, but permanent moat? That is a much harder question than it was two years ago.

Since then, my friend Jeff did his best to make sure I was right. Again, let me quote myself:

About a month ago, this column noted that Starlink's moat was already narrowing. I called out Amazon's Kuiper as the key competitor it will have to face. Today, Delta confirmed I wasn't hallucinating. Delta Air Lines chose Amazon's Kuiper over Starlink for its fleet. Musk criticized the decision publicly, which is quite rare. Usually CEOs will pretend they are civilized and may even congratulate their competitors, stressing it is for the greater good of the entire industry. Let's consider it just a little bit more. Aviation is Starlink's highest-ARPU segment. It generates around $300,000 per aircraft per year. Delta is one of the world's largest airlines. The default choice of wi-fi provider is no longer default. I'd say this was always going to happen, because the moat was never as wide as the margin suggested.

And then there is the geopolitical reality that nobody is saying loudly enough. All serious governments, in Europe, in Japan, in any country with a functioning sovereign interest in its own digital infrastructure, are in love with Elon. Hence they are going to allow critical national connectivity to depend permanently on a network controlled by this one American citizen.

As we sit in our backyards wondering with calculators, Elon is running a political movement, feuding with heads of state, and has a clause in at least one of his contracts reserving the right to cut off access if he decides it threatens humanity.

That is basic sovereignty arithmetic. Even without it, 41% for 15 years is never realistic given most of the comms infra spendings would be authorized on annual basis anyways. We haven't seen a 41% increase in that category of spending for any country in a while. With it, the chances can only narrow.

On commercial side, we are seeing strong competition: Jeff's winning of Delta and Japan's experimenting with HAPS. The industry will need to grow much faster than 41% to allow one company to grow 41% annually. I doubt connectivity market is that hot. If so, Ericsson and Nokia would have a higher stock price.

The rockets and space business

This is where I have the most respect and the most confusion at the same time.

SpaceX has executed over 620 Falcon 9 flights with a 99% success rate. They have put over 80% of global mass to orbit since 2023. Starship flew its twelfth test flight. The engineering is, without exaggeration, the most impressive rocketry achievement in the history of the species. I genuinely mean that.

But defense contracts, which are the primary revenue source for the launch business, do not compound at 41% annually. Defense spending is political, cyclical, and contested. Blue Origin is operational. They are here, they are funded, and governments do not like single-source dependency.

And Mars? The first Starship missions to Mars are projected for the end of 2026, uncrewed, and human landings are estimated for 2029 at the earliest, more likely 2031. I doubt any billionaire would want to spend years going there and then coming back. So why would they want to buy any tickets? If tourism is not something intriguing, then revenue from Mars is not in any model that makes sense for at least a decade. It is also unclear who would pay for it, at a 41% compounding growth.

So here is the question I keep returning to: which market can sustain a 41% growth for 15 years? If it is SpaceX monopoly, the market only needs to grow at 41% for 15 years. If SpaceX is a strong competitor, say duopoly or something like that, then the market would need to grow much faster.

The Whale

Sure, AI is growing fast. but SpaceX is not an AI company. It has leased out almost all its computation already.

Connectivity is not a 41% CAGR market.

Rocket launching is not.

Even EV is not.

The Morgan Stanley model requires all three of SpaceX's businesses to simultaneously outperform every comparable company in history, in markets that are either contested, nascent, or not yet invented, for fifteen consecutive years without a recession, a regulatory event, a competitive disruption, or a bad quarter.

Elon is the new ironman. He can do magic. But he can't transform an industry to a high speed growth for that long. Look at EV market.

I am just saying: man, I must have missed something.

And if I am wrong, which is entirely possible, which is actually probable, which is what I keep telling myself, then $2,000 is not very much to pay to ride the greatest growth story since the invention of the internet.

But that's still $2000, roughly half month's rent in San Francisco. Maybe I should keep watching.

The pacific is deep and beautiful, full of things I don't understand. As we speak, a 2 year old whale is trapped in Monterey bay. I am more interested in helping it, if it can still be helped.

Bad news. I just checked. It passed.

Previous in this series:

Cursor Has No Moat. SpaceX Has No Cash. Perfect Match.

The Landlord, The Exit, and The Ghost of Houston

SpaceX's S1 Is Out