It's supposed to be a constellation of ~700 satellites, that definitely suggests low orbit.
(Rationale: With high orbits like GEO, a single satellite covers half the planet, so there's no obvious reason to have hundreds of them. But LEO is much nicer for many reasons (latency, launch cost). If you could cover the whole planet from LEO, that'd be the logical choice; and planning ~700 satellites at once suggests that's the aim).
The problem with low orbit is increased drag. This is basically why Iridium went bankrupt trying to charge $7/minute for phone calls - the lower you are the more fuel you need to expend for station keeping. A satellite that would last decades in geostationary orbit may only last four or five years at 120km.
> The difference between their satellites and others in use would be the size of the fleet .. and the size of the satellites, which would be much smaller than those currently used for communications.
Pretty sure we're looking at satellites far far smaller than 250 pounds if this is going to work.
Realistically you can't make them much smaller than that. I know everyone's excited about "cubesats" that are shrunk down to ten cm on a side, but those things are always part of some "swarm" experiment that runs for a short period of time (usually hours) and ends. Here we're talking about something that needs to work for an extended period. And it's not some made up task like "Let's see if we can get ten tiny satellites to fly in formation!" You're actually going to be slinging RF energy around. So you're going to need:
1) Power. How much really depends on what's going to be on the ground. If your customers are going to be using twelve foot tracking dishes (they won't), you can get away with maybe a quarter watt or less. I'm not exactly sure what they'll need, but remember you're trying to communicate with something that's 100+ km away, minimum. Nobody's going to let you put up 700 plutonium RTGs, either, so that means solar cells and probably some sort of mechanical deployment system. And batteries - you still need power on the night side.
2) RF antenna(s). The more you skimp on your antenna, the more power you need. And vice versa. And you'll probably need at least two - one to point at your customers, and one to point at whatever you're using as a relay.
3) Thrust. Probably small, light thrusters and a comparatively large, heavy fuel tank. The lower your orbit, the more you need.
4) Stabilizers. Of some sort. Your antennas need to be pointed in specific directions, and your solar cells need to face the sun. So... gyros (and more electrical power). Plus more fuel, because you need to keep the gyros in their design range.
You're not going to be able to fit all this in a twenty pound satellite.
