This strikes me as a rather odd comparison to make.
4G serves an entire nation of cell phones, using generally similar wireless technology and limitations LEO satellite comms have. 4G towers are closer to the users than LEO satellites, meaning the power requirements are lower, this is a desirable thing for the people in terms of both battery life and radiation power levels.
5G takes it even further in this direction using more towers; shorter distances and lower power levels, AIUI.
Unless I'm utterly mistaken, LEO satellites are at a severe disadvantage when it comes to competing with the cell network for handheld mobile customers.
Satellite connectivity has its place, and it's basically where you won't have cell service, since that's obviously the preferable option if it's available.
Why not both. Have a 4G tower with a starlink on top. Give internet to a whole rural village. That might not be fast enough for the first world, but in other places that will be a much more viable solution compared to building physical infrastructure on the ground.
But those satellites cover everywhere, including places where no carrier would consider erecting a cell tower because it's not dense enough to be profitable or further out into the boonies than would be cost effective (or politically viable) to run cable.
It's an expensive upfront cost, but once the full constellation is up, they'll have basically full world coverage, and anywhere that's not a big city charges big money for internet. See how much internet in rural Canada costs.
Except their network is extremely limited in capacity. 4G/5G is a much better solution for 99.9% of use-cases. Starlink is better for the extremely rural locations where installing a tower just isn't feasible.
I have a pretty rural property and there are areas nearby that get 5G UW with speeds in excess of 1Gbps. Most of the area is covered by 4G with download speeds of over 40Mbps.
5G is already replacing wired home internet for many people. Give it a few more years to expand coverage and I bet 5G will make Starlink obsolete for most of its US users.
A few things to note are that Starlink is by far the best option for airplanes and boats, and the percentage of very rural customers is probably closer to 5-10% than .1%. Putting up towers is expensive, and space is pretty close. If you want the tower to be much closer than a leo satellite, it needs to <<200 miles away, and in non-flat areas, having the signal go up gives much better results than trying to send it through a mountain.
> It's an expensive upfront cost, but once the full constellation is up, they'll have basically full world coverage
This is wrong. Years before they have a full constellation they will already have started needing to replace previously launched satellites. It's not like they just launch them once and they're done and those satellites last forever. Just to maintain a constellation they'll need to be replacing something like 20% every year.
Are you counting how expensive it is to keep replenishing the sattelites?
Anyway, for spacex to make bank off of this they need customers. Extreme rural living areas are not that common, and when they do become common more cost effective methods such as cell towers make more sense than starlink.
Is it, though?
Can you share the pricing details of managing 4G towers vs. Starlink's small sats.
And yes, I fully expect that you have no idea what either of those costs.