SpaceX being their own customer, and a lower priority one, lets them levellize their revenue, which also helps retain staff. And if anyone commoditizes launch systems, they have a hedge against that. However if demand for starlink ever does exceed capacity, the launch side could find their schedules being much more variable.
In software, you see this with priority queues all the time if you know where to look. Someone makes a “low priority” system that soaks up spare capacity. Internal customers become dependent on these systems and then become delay-averse. Meanwhile, other teams also start soaking up that spare capacity, by improved design or feature creep or both. That low priority system becomes starved of resources, and the internal customers start saber rattling. That system gets its own dedicated resources, and now stranded resources climb back to the original levels, if not worse.
Starlink will eventually become the same thing. And I wonder if they will spin it off when that situation arises.
You could say the same thing for terrestrial internet. Not trying to be contradictory, I just find this a very interesting ethical topic.
As a UIUC student used to be fond of reminding me, part of the University of Illinois’ contribution to the old Internet was in a building that, if it were to collapse, would fall outward and not crush the basement. Officially that building was designed so every office has natural light, like a collapsible water cup, but it’s a very interesting coincidence that it (the basement) is bomb proof.
https://en.wikipedia.org/wiki/Golden_Dome_(missile_defense_s...
It would be a pretty inefficient way to deploy offensive weapons.