After writing several articles about SpaceX, I discovered something strange.
Even though I was covering the same company, its identity seemed to change with every piece. In articles looking at its revenue structure, it looked like a telecom company; in pieces about its Texas water contracts, it appeared as a regional infrastructure operator; and in posts analyzing engine mixture ratios, it was simply a chemical plant.
I had viewed one company six different ways, and the pieces were scattered without being connected. This article is an attempt to lay those pieces out in one place and put them in order. Since each topic has already been covered in depth in individual posts, I will focus here on how these pieces lock together.
1. This company does not make money from rockets
This is the first premise that needs to be broken.
61~69% of SpaceX’s revenue comes from Starlink subscriptions, not rocket launches. Launch services are actually a minority in terms of revenue share. The market compares this company to Amazon precisely because of this structure; the pattern of “building the logistics yourself and layering a subscription business on top” actually overlaps.
If you grasp this fact first, the rest of the pieces become much easier to read. Launches are closer to a cost center, and every activity that lowers those costs (reusability, launch frequency, proprietary launch sites) is essentially about protecting Starlink’s margins.
2. So, telecommunications is the main base
If Starlink is the main base, then the technical challenges are also concentrated in telecommunications.
In particular, for Direct-to-Cell, where satellites receive signals directly from mobile phones, the uplink is much harder than the downlink. A handheld phone has to push a signal to a satellite 550km above, yet its output is less than 1 watt. How they overcome this asymmetry determines the service quality.
3. Inside the engine, chemistry beats economics
To lower launch costs, you have to go down to the engine level.
Here is one counterintuitive fact: Rockets do not use the maximum amount of oxygen. They intentionally inject more fuel than the chemically perfect stoichiometric ratio, not for thrust, but because of the average molecular weight of the exhaust and turbine temperatures. The very definition of efficiency is different from what we typically think.
And where does that oxygen come from? It is the product of a process over 100 years old, cooling air to -183 degrees Celsius to turn it into a liquid and then separating it. Once you know the amount of liquid oxygen that goes into a single rocket, you understand why a launch site must be a large-scale industrial facility.
4. The launch site is a factory, and the factory consumes water
If you go one step deeper from the engine, you find land and water.
Starship pours a massive amount of water during launch to protect the launch pad. This is the background behind SpaceX investing $220 million in Brownsville’s water resources. It is the moment when water supply investments appear on a rocket company’s financial statement.
5. What follows is a city
After water comes land, and after land comes people.
Employee housing and commercial zones emerge around the launch site, and the interests of local administration and the company begin to overlap. This isn’t a pattern SpaceX invented, but rather the latest version of the ‘company town’ that has repeatedly appeared in American industrial history. I covered this structure in an article analyzing the timing of the Louisiana announcement.
6. When the same logic moves to semiconductors
What is interesting is that this pattern does not stay in space.
The same bottlenecks appear in the discussions surrounding ‘Terafab,’ the ultra-large semiconductor factory Musk is envisioning. Things that cannot be bought with concrete—power, water, skilled labor, and equipment lead times—ultimately determine the speed. The lessons learned at the launch site are repeated exactly.
When you lay all the pieces out
If you place them in order, it reads like this:
Starlink makes money → To protect those margins, launch costs must be lowered → To do that, engine efficiency and launch frequency are required → To increase launch frequency, a dedicated launch site is needed → The launch site attracts water, land, and people → As a result, the company becomes an entity of the local economy.
This chain cannot be explained by the single word “rocket company.” Yet, it is not a telecom company, an infrastructure firm, or a real estate developer either. Because each face of the company requires the others, looking at just one part always leads to a half-truth.
For reference, you can see how difficult this structure is by looking at the competitors. In an article covering Blue Origin’s New Glenn explosion, I summarized why it takes so long to rebuild the same chain from scratch.
I plan to continue adding to this post if I write more about SpaceX in the future.