> I do suspect we need to find commercial enterprises to justify the Starship economics and I'm just not sure what that looks like yet.
At $1 MM per launch with 100 t per launch puts Starship roughly at $10 per kg.
Asteroid mining: Will be a huge industry. Especially with growing environmental concerns all around Earth. Waste can basically be disposed of very cheaply into the atmosphere or sun.
Starlink, Kupier, etc constellations: will continually need maintenance. Including similar constellations across the Moon and also Mars.
Entertainment: Assuming 500 kgs per human (because of life support, seats, luggage, food, etc), that is $5000 per ticket to LEO. This is a very reasonable price point for luxury vacations, especially considering an "all inclusive" trip like a cruise or resort. In line with this, a space hotel/casino is very reasonable to get constructed.
> All of the things you've listed here though aren't really commercial enterprises. Or at least they're far from it. They're flagship missions by governments because they're so wildly expensive.
You're using past data to predict the future. Which is typically a good heuristic. However, when it comes to game changing technologies you can no longer use past data to predict future performance. SpaceX will continue to make margins off Starlink and regular F9-like flights. These profits are all going to be reinvested until a Moon and Mars colonies exist and regular trade begins.
Finally, other companies like Relativity Space are creating amazing 3D printers for rockets. These new technologies are only used in a small way in Starship, but could greatly reduce costs for Starship or the next ship. If the 2010s were about F9, and 2020s are about Starship, then the 2030s call for another SpaceX ship that gets even better tech advantages.