Their own internet megaconstellation, called Project Kuiper until earlier today when they renamed it to Project Leo.
It's actually the current biggest commercial launch customer, Starlink is internal to SpaceX, but Kuiper/Leo has bought many launches with ULA, SpaceX and Arianespace (and Blue Origin, of course).
Even if you exclude Starlink SpaceX is overwhelmingly dominant. The stuff they don't launch is mostly China and Russia and Europe they can't compete for.
Starlink prints money, SpaceX is absolutely launch mass constrained right now. They literally spent $17 billion on spectrum to make Starlink better and more efficient because of how constrained they are on the launch side until Starship is fully operational and reusable, which may not be until 2027 even for Starlink launches.
It doesn't seem that atypical when extremely high capex and proprietary R&D are moats. Off the top of my head, the semiconductor industry looks broadly similar right now and the fusion industry might end up looking similar for a while.
Only small parts of the semiconductor industry at the very cutting edge even remotely resemble that. And that’s technology with outcomes (I.e. process nodes) that are genuinely new and have never been done before. What’s being accomplished now in space are outcomes that were accomplished before PCs existed, so the idea of it being insurmountable R&D doesn’t hold. It’s very telling that the only “commercially viable” launch providers are billionaire trophy assets with induced demand from a heavy slice of government sponsorship and self dealing.
Saying there's no market demand for cheaper launchers, when the company with the cheapest large launchers has cornered the market makes no sense. That was my only point.
Meta point: why does that matter? They launch 90% of the demand for payload to orbit. Some of that demand is from a vertically integrated part of the company. It is still part of industrial demand, given that Starlink is profitable already.
The launch count of SpaceX per year compared to the rest of the world is quite large.
SpaceX in 2025 has launched 134 times. Everyone else in the entire world has launched 115 times combined, including other US companies. SpaceX launches a lot of stuff very often.
EDIT: Originally meant to do 2024 but accidentally read the wrong bar. Regardless, this holds for most years.
They're booked out years in advance only in the sense that bookings are sorted out years before the payload is ready to fly. SpaceX has emphasized that they're capable of swapping out Starlink launches with a commercial payload if needed on short notice.
F9 launches are available anytime a customer wants them. SpaceX will bump down a Starlink launch to accommodate a paying customer, All they would really need would be the payload assembly time?
