After writing several pieces about SpaceX, I noticed something odd.
Every article was about the same company, yet the company looked different in each one. In the piece on revenue it behaved like a telecom. In the piece on the Texas water deal it behaved like a regional infrastructure operator. In the piece on engine mixture ratios it was simply a chemical plant.
Six views of one company, scattered and unconnected. This article puts those pieces in one place and gives them an order. Each topic already has its own in-depth article, so the focus here is on how the pieces lock together.
1. This company does not make its money from rockets
That is the first assumption to drop.
Between 61% and 69% of SpaceX revenue comes from Starlink subscriptions, not launch services. By segment share, launch is the minority business. That structure is why the market keeps comparing the company to Amazon — the pattern of building your own logistics layer and stacking a subscription business on top genuinely does overlap.
Hold onto that fact and the remaining pieces read much more easily. Launch is closer to a cost centre, and everything that drives that cost down — reuse, cadence, owning your own launch site — is ultimately protecting Starlink’s margin.
2. Which means telecom is the main front
If Starlink is the core business, the hard engineering problems cluster there too.
Direct-to-cell, where a satellite talks straight to an ordinary handset, is far harder on the uplink than the downlink. A phone in your hand has to push a signal up to a satellite 550km overhead on well under a watt of power. How you bridge that asymmetry decides the quality of the service.
3. Inside the engine, chemistry beats economics
Driving launch costs down means going down into the engine.
Here is one counterintuitive fact: rockets do not run on as much oxygen as they can carry. They deliberately run fuel-rich, past the ratio that would burn cleanly — not for thrust, but because of the average molecular weight of the exhaust and the temperature the turbine can survive. The definition of efficiency itself is not the one most people assume.
And where does that oxygen come from? From chilling air to minus 183°C, liquefying it and separating out the oxygen — a process more than a century old. Once you know how much liquid oxygen a single launch consumes, it becomes obvious why a launch site has to be a heavy industrial facility.
4. A launch site is a factory, and factories drink water
One level below the engine you hit land and water.
Starship floods the pad with an enormous volume of water at liftoff to protect it. That is the background to SpaceX committing roughly $220 million to water infrastructure around Brownsville. It is the moment municipal waterworks spending shows up on a rocket company’s books.
5. And then it becomes a town
After water comes land, and after land come people.
Housing and local commerce grow up around the launch site, and the interests of the local administration and the company start to overlap. This is not a pattern SpaceX invented — it is the latest version of the American company town. The article examining the timing of the Louisiana announcement digs into that structure.
6. When the same logic moves to semiconductors
What makes this interesting is that the pattern does not stay in space.
The same bottlenecks surface in the discussion of Musk’s proposed mega-fab, “Terafab”: the things concrete cannot buy — power, water, skilled labour, equipment lead times — are what actually set the pace. The lesson from the launch site repeats almost exactly.
Laying all the pieces out
Put in order, it reads like this.
Starlink earns the money → protecting that margin requires cheaper launches → cheaper launches require engine efficiency and cadence → higher cadence requires your own launch site → the launch site pulls in water, land and people → and the company ends up as an actor in a regional economy.
“Rocket company” does not describe that chain. Neither does telecom, infrastructure operator, or property developer. Each face depends on the others, so pulling any single one out always leaves you half right.
For a sense of how hard this chain is to build, look at the competition. The article on Blue Origin’s New Glenn explosion covers why rebuilding the same chain from scratch takes so long.
As more SpaceX pieces get written, they will be added here.