SpaceX Isn't a Rocket Company Anymore
- Rich Washburn
- 20 hours ago
- 3 min read
Everyone still thinks SpaceX is a rocket company.
The first earnings report as a public company just revealed something different.
Revenue came in at $7.8 billion, up 92% year-over-year, beating Wall Street's expectations. Losses were smaller than expected too. Normally that's the whole story — a growth company blowing past estimates, stock should pop.
Instead the stock fell after hours. Because the number that actually matters isn't the top line. It's the breakdown underneath it.
Starlink, the connectivity business, brought in roughly $4.3 billion. AI brought in roughly $2.6 billion. Launch — the actual rocket business, the thing everyone associates with the SpaceX name — brought in just under $1 billion.
Read that again. The rocket business is no longer the primary economic engine. It's the smallest of the three.
WHAT'S ACTUALLY HAPPENING HERE
For twenty years, the SpaceX story has been about launch. Reusable boosters, falling launch costs, Starship, the whole mythology of "making humanity multiplanetary." That's the story that gets the headlines and the documentaries.
But look at where the money's actually coming from now. Starlink isn't a side project anymore — it's more than half the company's revenue. And AI, a category that didn't meaningfully exist in SpaceX's business a few years ago, is already bigger than the entire launch division that built the company's reputation.
That's not a rocket company that happens to have a satellite internet business and an AI bet on the side. That's an infrastructure company that happens to have started with rockets.
Rockets are becoming the transportation layer — the thing that gets hardware into orbit. Starlink is the connectivity layer — the thing that turns orbital hardware into a distribution network. AI is the monetization layer — the thing that turns connectivity and compute into revenue.
Elon isn't building three companies. He's building one vertically integrated infrastructure stack for the AI era, and the earnings report is the first time the market's gotten to see the stack's actual economics laid out in public.
WHY THE STOCK FELL ANYWAY
Here's the part that should make you pay attention rather than just nod along: despite crushing revenue expectations, the stock dropped after hours. Wall Street is worried about two things. The size of the AI capital expenditures, and a massive insider share unlock hitting this week.
The capex worry is the important one. Building AI infrastructure at the scale SpaceX is apparently building it — enough to generate $2.6 billion in AI revenue in a single quarter — requires enormous, sustained spending on compute, power, and physical buildout before that revenue shows up. Investors are pricing in the risk that the spending outpaces the payoff, at least in the near term.
That's the same tension playing out across every major AI infrastructure bet right now. The revenue growth is real. The capital intensity is also real. And the market hasn't fully figured out how to price a company that's simultaneously a connectivity provider, an AI infrastructure operator, and a rocket company — because there isn't really a clean comparable.
THE PART THAT CONNECTS TO EVERYTHING ELSE
This earnings report is a data point for something I've been arguing for months: AI isn't becoming a software industry. It's becoming an infrastructure industry.
Power. Data centers. Networking. Satellites. Compute. Memory. Cooling. Those are becoming the real competitive moats — not model architecture, not app-layer features. The companies that win the next decade of AI aren't necessarily the ones with the best model. They're the ones that control the physical stack the models run on.
SpaceX's earnings report is the clearest public illustration of that thesis I've seen yet, because it's not a story about a company pivoting into AI. It's a story about a company that built physical infrastructure for an entirely different reason — getting hardware into orbit — and discovered that the infrastructure itself, the satellites and the network and the compute sitting behind it, was worth more than the thing it was originally built to do.
That's the pattern to watch for elsewhere. Look at who already owns power generation, who already owns fiber and spectrum, who already owns data center real estate, who already owns the supply chain for chips and cooling systems. Those are the companies positioned to do what SpaceX just did — discover that their existing infrastructure is quietly becoming an AI company, whether that was the plan or not.
The rocket business made SpaceX famous. The infrastructure is what's going to make it dominant. And the market is still figuring out how to value a company that's already crossed that line.
Rich Washburn is a technologist and strategist working at the intersection of AI, infrastructure, and capital. He is Managing Partner and Chief AI Officer at Eliakim Capital.
Comments