If double latency is okay, it may be possible to do M-shaped ones where an isolated ground station (with only power) relays solely between two different satellites to reach further, for example in the middle of oceans (with a ground station on an island or ship). I am not sure if they plan to implement this interim step or not, but it is technically possible.
Their eventual goal is to allow communication directly between the satellites (via laser is the plan, but who knows if that will reach production). It's in testing/R&D and is quite a difficult challenge.
Yeah they've launched, but do they work? Teslas can be ordered with "autopilot" or "full self driving", too.
Nobody's ever done it before, and they've never said if it's working or not. I know they intend to do this, but it's important to draw a big fat line between what is planned to be accomplished and what is actually possible today. (To be clear, I am not casting doubt on their ability to accomplish it - the smartest people on Earth work for SpaceX. It's just not known publicly today if/when they will.)
Perhaps EM's best and most important skill is blurring the lines between today, tomorrow, next quarter, and next year as much as possible. Has someone named his Reality Distortion Field yet?
I guess they'll have to though if they want to make it more scale-able? They satellites will always have to be up in space, but with communication between them they won't have to build ground stations everywhere. Plus ground stations themselves have to have good internet connectivity
