SpaceX: Is It Worth the Hype?

I’ve been gone a few months. I was heads-down building something, and I’ll have more to say about it soon. Thanks for still being here.

Now to the stock everyone wants to argue about.

SpaceX went public in June in the largest IPO in history, raising $85.7 billion in net proceeds at $135 a share. Retail piled in. Institutions built positions fast. The coverage was impossible to avoid.

Then it broke. The stock ran above $225 within days of listing, and it has been falling ever since. A failed Starship engine ignition in July, a first earnings report that spooked everyone on spending, and it now trades near $108. That is down roughly 52% from the high and about 20% below where it priced, in under two months.

That round trip is the whole debate in one chart.

On one side, the most ambitious technology company in public markets, operating across launch, satellite broadband, and AI.

On the other, a business that spent $18.4 billion in a single quarter and is asking investors to pay a premium for returns nobody can date.

SpaceX reported its first public quarter on August 4th. I have been building my view of this one for months. Here it is.

Help us become the fastest growth stock newsletter:
Share HappyStocks with Friends

Most people still file SpaceX under “the rocket company.” That stopped being accurate a while ago.

There are three segments now.

•      Space. Falcon launches, Dragon missions, Starship development, and related launch services. This is the foundation of the business, and it powers government, commercial, and defense missions worldwide.

•      Connectivity. Starlink consumer broadband plus a fast-growing enterprise and government business, delivered through thousands of satellites already in orbit and serving households, airlines, maritime customers, and militaries.

•      AI. The newest segment. It bundles AI infrastructure, Grok, and X into one reporting line covering compute, models, enterprise software, and digital advertising.

All three grew last quarter. That has not happened before.

Three businesses, three completely different ways of earning a dollar.

Space gets paid per launch.

A customer has something to put in orbit. SpaceX quotes a price, flies the mission, and books the revenue. Customers are NASA, the Space Force, commercial satellite operators, and foreign governments. There are also development contracts, where a customer pays SpaceX to build or adapt something for them.

The economics come down to one idea: reuse the rocket. A booster that flies twenty times spreads its cost over twenty missions instead of one. That is why SpaceX can charge less than anyone else and still make money on the flight.

Worth knowing that most SpaceX launches are not for customers at all. It is mostly flying its own Starlink satellites. That does not show up as revenue. It shows up as a cost nobody else can match.

Connectivity gets paid every month.

This is Starlink, and it is a subscription business. You buy the dish once, then you pay a monthly fee for internet. Roughly $66 a month on average.

It splits in two.

•      Consumer. Households, mostly rural or remote, where cable and fiber never showed up. One dish, one monthly bill, 12 million subscribers and counting.

•      Enterprise and government. Airlines buying wifi for their fleets, shipping companies, cruise lines, telecom carriers, and militaries. Bigger contracts, multi-year, and far more revenue per customer than a household. This side is now growing faster than consumer.

The beauty of it is what happens after the satellites are already up. The constellation costs the same whether it serves nine million people or twelve million. Every subscriber added on top of existing capacity is close to pure profit. That is why this segment funds the rest of the company.

AI gets paid three ways.

•      Selling compute. Companies rent access to SpaceX’s data centers to train and run their own models, usually on multi-year contracts.

•      Subscriptions. Paid tiers for Grok and for X.

•      Advertising. Ads sold against X.

Compute is the piece that matters. It is the largest, the fastest growing, and the reason the company is spending so heavily.

Put it together and the model is simple to describe. Starlink produces steady recurring cash. Launch produces lumpy contract revenue and a cost advantage nobody can copy. AI produces contracts that are enormous, new, and still unproven.

Three separate growth markets, and they no longer move together.

Yes.

Building rockets is hard. Building reusable rockets at scale is much harder.

Then layer on launch infrastructure, manufacturing, satellite networks, spectrum licenses, government certifications, decades of engineering knowledge, and regulatory approvals. Very few companies can put that together.

The real moat is vertical integration. Launch capability feeds satellite deployment. Deployment supports the broadband business. Broadband generates the cash and the network that support AI infrastructure. What would be a supplier invoice at any other company is an internal cost line here.

A competitor would need years and an enormous amount of capital to build the same loop.

Market Opportunity:

By SpaceX’s own math, the total addressable market is $28.5 trillion.

•      $370 billion from space-enabled applications
•      $1.6 trillion from connectivity
•      $26.5 trillion from AI

I would not put much weight on a company’s own TAM slide. The number I care about is backlog, which stood at $47.5 billion at quarter end. That is contracted, not aspirational.

The useful takeaway is that SpaceX is aiming at industries in their early innings rather than mature markets with a ceiling already in view.

Revenue Growth

Growth is accelerating, which is rare at this size.

Revenue went from $10.4 billion in 2023 to $18.7 billion in 2025. The first half of 2026 alone did $12.5 billion against $8.1 billion a year earlier. Q2 grew 92% year over year.

Annualize the June quarter and you are already above $31 billion. Management is guiding to a $100 billion annualized run rate by the end of 2026, which I do not believe, and to $1 trillion in revenue by 2030, which I believe even less. Discount both.EPS Growth

EPS has gone from deeply negative to positive, and that happened fast. The company hit its first GAAP profitable year in 2024 with $0.03/share, then grew dramatically to $0.56 in 2025. The loss years reflect heavy investment in growth.

EPS Growth

Losses are shrinking fast.

Q2 net loss was $541 million, or 9 cents a share, against a $1.0 billion loss a year earlier. Analysts were modeling a loss closer to $1.9 billion.

The first half looks worse at $4.82 billion because Q1 carried a $4.28 billion loss. That number is misleading. Q1 operating loss was only $1.94 billion. The rest came from roughly $1.9 billion in non-operating charges below the line plus about $450 million in net interest. The business did not burn $4.3 billion in operations.

Cash Flow

Operating cash flow was $3.47 billion over the first half, up from $351 million a year earlier.

Capital spending is the whole story. SpaceX spent $18.37 billion in Q2 alone and $28.48 billion over the first half. AI took $15.83 billion of the second quarter figure. Analysts were modeling around $13 billion. The stock fell roughly 8% after hours on that line item, not on revenue.

Free cash flow is deeply negative. That is a decision.

Margins

Consolidated operating loss was $143 million on $7.81 billion of revenue. That is a rounding error from breakeven in the heaviest spending quarter in company history. Adjusted EBITDA was $3.54 billion, up 191%.

The segments look nothing alike underneath.

•      Space: $962 million revenue, $542 million operating loss. Falcon economics are healthy. Starship R&D consumes them.

•      Connectivity: $4.29 billion revenue, $1.66 billion operating income, $2.60 billion adjusted EBITDA. This funds everything else.

•      AI: $2.56 billion revenue, $1.26 billion operating loss, and positive adjusted EBITDA of $1.15 billion. Still the largest drag on GAAP results, but the trajectory reversed this quarter.

Read that as one sentence: Starlink is printing money, Space is spending it on Starship, and AI just stopped being a pure cash incinerator.

Balance Sheet Strength

The IPO changed this company’s risk profile more than any operating result did.

SpaceX ended June with roughly $100 billion in cash, cash equivalents, and marketable securities, against $24.7 billion at the end of 2025. Total assets are $192.8 billion and shareholders’ equity is $127.2 billion.

Debt and finance leases total about $39.4 billion, including a $25 billion inaugural bond issued in June across five tranches at a weighted average rate of 5.855%.

That’s real leverage. It is also fixed-rate, long-dated, and sitting next to $100 billion of cash. This company can fund its own capex plan for years without asking the market for anything.

Institutional Ownership

Ownership is unusual, and it matters more here than at most companies.

There are two share classes: 7.61 billion Class A shares and 5.57 billion Class B. Insiders hold roughly 41% of shares outstanding, with Musk somewhere near 36%.

That is strong alignment. It also means one person decides, and a lockup expiration freeing more than $100 billion in shares is now in front of the stock. Know both before you own it.

Launch is a monopoly in everything but name, and it is the cheapest ride to orbit that has ever existed. That is the foundation. Here is what sits on top of it.

If management executes across all three segments, the company five or ten years from now looks nothing like the one you are buying today.

Every one of these is a reason the stock has already fallen 52%.

Over the next five years, this comes down to one question.

Does the AI capex earn its return?

Starlink can fund Starship indefinitely. It cannot fund $16 billion a quarter of compute. That gap gets closed by the Cloud Services Agreements converting into durable, repeatable revenue, or it does not get closed at all.

Ten years out, the upside gets larger and so does the uncertainty. Regulation, technical setbacks, geopolitics, launch failures, competition, and AI economics can all rewrite the story.

This is not a predictable consumer software company. It is a capital-intensive platform trying to reshape several industries at the same time.

I would not lean on analyst models here. Below is one projection of revenue and net income. Treat it as a scenario, not a forecast. Nobody sees 15 years out with any accuracy.

Here is where I land.

I think SpaceX has the potential to be the biggest company in the world.

The first reason is structural.

Almost no company owns its entire stack. SpaceX builds the rockets, builds the satellites, owns the spectrum, owns the constellation, builds the compute, and trains the models that run on it. Every layer that would be a margin paid to a supplier somewhere else is an internal cost line here, and internal cost lines fall over time.

You can already see it working. Consolidated operating loss of $143 million on $7.8 billion of revenue, in the quarter they spent $18.4 billion. Connectivity operating income up 79%. AI EBITDA positive for the first time. That is operating leverage showing up early, at the exact moment the spending peaked.

The second reason is Musk, and I mean that in a way most analysts will not write down.

He is the best product engineer of his generation. He is also the most effective recruiter of belief in modern markets. Those are two different skills and he happens to have both. The first one builds the rockets. The second one funds them.

Look at Tesla for the pattern. Sixteen years, repeated drawdowns of 50% and worse, and a shareholder base that treated every one of them as an entry point. The cheapest that stock ever got was roughly 30 times earnings, in January 2023, after a 70% decline and a genuine margin scare. Thirty times. Ford trades at six. That was the bargain, and the people waiting for Tesla to price like a normal automaker have now waited sixteen years and never owned it.

I expect that same base to form around SpaceX. It has not formed yet.

Eight weeks public, down 52% from the high, sitting at its lowest price since it listed, and both the July engine failure and the capex print stuck. No floor showed up. That is not an argument against the thesis. It is one of the reasons I have not bought yet.

The third reason is the one I have not seen anyone else say plainly.

I do not think this is the final shape of the company. I expect SpaceX and Tesla to become one business before the decade is out.

The logic is not sentimental. Both companies are solving the same three problems: manufacturing at scale, autonomy, and power. Data centers need energy, and Tesla builds grid storage. Robotaxis and Optimus need models and compute, and SpaceX now runs a 1.4 gigawatt AI platform. Vehicles need connectivity, and Starlink is the only global network that can deliver it. The same person controls both, and the dual-class structure at SpaceX already concentrates that control.

Two companies solving one problem eventually stop being two companies. Put them together and you are looking at the largest enterprise ever assembled.

There is no announced deal, and a merger would need two boards, two shareholder votes, and antitrust review. Price it as a forecast, not a fact.

So am I buying?

Yes. Just not today.

If you follow me on Instagram at @invest, you already know how I treat IPOs. I do not buy on day one. Nor do I buy in their first quarter or more as a public company.

Every high-profile IPO comes in hot. Most IPO’s float only about 10-25% of the company, and insiders are locked up and cannot sell. Price discovery has not happened yet, and you are also bidding against many people who do not have a full understanding of the overall business.

SpaceX is the textbook case. Priced at $135, opened at $150, ran past $225 inside a week. Then one engine failure and one capex print took it to $108. That is a 52% round trip in under two months, and none of the business changed. Only the number of people willing to pay up did.

So, yes, I am waiting for a better price. But better is not the same as cheap, and this is where many people will get it wrong.

I expect SpaceX to trade lower than it does today. What I do not expect is for it to ever look cheap by the standards most investors were trained on. If the shareholder base I described above forms the way I think it will, the multiple gets a floor well above anything a spreadsheet will call reasonable. Holding out for 20 times earnings on this name is how you never own it.

You cannot run an Elon company through the same comps as everything else. The tools are not broken. They are built for companies whose shareholders behave differently.

Right now, I am watching two things: a better entry and a real float.

The lockup expiration can deliver both. When it lifts, more than $100 billion in shares become tradable, insiders can finally sell, and the float multiplies. More supply usually means a softer price, which is exactly the entry I want.  You can also think of it as the moment the queue ends and an actual market begins.

That window is weeks away, not years. I should start building my position soon.  Not because it will be cheap by the old rules, it will not be, but because I want one of the biggest companies in the world in my portfolio, and I feel much more confident now that I’ve seen their potential this early in the game.

High potential and high uncertainty are the same thing here. You do not get to own one without the other.

Invest with caution.

Happy Investing,

Ralph