Your revenue figures are consumer only. And while you're generous on utilization factor, we capitalised the TCO up front while amortising revenue, and then reduced asset tenure to worst case observed during development.
Flex up to 4 years, let $1mm TCO be paid up front and the rest amortised, and reduce utilisation to 80% ($80k/month revenue) and IRR shoots up to 73%. Take TCO to $3mm ($1mm up front, $2mm amortised), reduce utilisation to 75% and we're still over 20%.
> while "7 million subscribers" might sound impressive on first sight: This is the number of DSL connections subscribed to in the tiny country of Belgium. But for magical reasons Starlink is valuated at a price higher than if you would buy all of Belgium
Well, yes. Starlink connections are more profitable and you can't scale selling internet to Belgium into a Starshield defence contract. Or selling to airlines and cruise ships and yachts and mining operations, all of which pay more than a Belgian.
> some countries in the EU for example years ago created laws that say that whoever opens the street for any reasons has to put in empty tubes for someone to later put in fiber before closing the street again
Starlink doesn't sense in densely-populated areas of the EU or Asia. (And the equivalent for SpaceX would be ridesharing Starlink on someone else's flight.)
> Fiber is dirt cheap if you are allowed to use existing power poles for example
If you have the scale. You're underestimating the risk that comes from having to place infrastructure up front.
Your analysis is pretty solid. But I don't think it's taking into account the fact that you can build multibillion-dollar telecoms business on a few tens of millions of high-paying customers.
