I think Elon cash a lot about cost.
Totally agree. He absolutely does, both from information in public domain as well as beyond. What he doesn't care about is adhering to traditional methods of managing cashflow.
350 million people in the US, say 4 to a household, 87.5 million household, if Starlink is internet and phone for $50 a month and only serves 4% of the population (19% live in rural areas), that is $2.1 Billion a year.
I don't wish to quibble over math too much, but my math checks out differently when using your base assumptions:
350M people x 4% = 14M people served by starlink
14M people x $50 = $700M annual, and that's before down factoring raw population into aggregate cells, like
Regardless, what remains to be seen is whether or not Starlink will offer a product compelling enough for 4% of the population to actually sign on. While ~10% of the country doesn't have internet, 0.04% of the US country doesn't have
access to at least 10MB either terrestrially or through existing satellites, so I think its pretty fair to assume that anyone who wants and can afford internet in the US has it already. There's esentially nobody out there that's waiting for faster/better/cheaper internet to get internet service. They've already made the choice that shitty internet is better than no internet.
So then the exercise turns to understanding how many people will drop existing service in favor of Starlink. The issue I see is that pretty much the only significant population willing to do that will be existing satellite customers. There's ~10M or so people in that camp, many of whom will no doubt switch to starlink, so that's the good news. Of course not all of them will transition--some won't want the hassle of switching, some are probably bundled with other services, some probably hate on Elon/SpaceX/Tesla, and certainly the legacy operators will try to offer a better product when Starlink actually provides competition. As GEO customers naturally transition to Starlink, especially the customers at the top of the speed/cap bell curves, it will be easier for GEO sats to offer higher performing service. Ultimately though I suspect its going to come down to the bottom line for many of those 10M customers, and I think that bottom line is going to be pretty close across the options.
Now expand that to the rest of the world.
Where people generally have less money. (And thus will have a harder time affording any kind of service)
Where population densities are often more urban. (= the exact place satellite service doesn't work well)
Where there's already quite a bit of GEO/MEO internet service. (And so, existing competition, some of it state sponsored)
Where all kinds of companies are trying to get their piece of the connectivity pie. (like terrestrial services)
Where countries deny landing rights for satellite services. (Good luck getting open internet to China's 1B people)
To be clear, I believe Starlink will achieve the primary goal of being a profitable service provider.