LOE will always provide faster latency than Geosynchronous orbit.

I suspect that total bandwidth will go to the person with the most satellites, and LEO has much cheaper launches than GeoSync.

Geosynchronous orbit is more than 15 times farther away. That is why the communication will be quicker, and why the launches are cheaper.

I specifically said high bandwidth since the latency is obviously better with leo. I don't know what you are implying by much cheaper launches. They need many, many more launches to hit the same capacity/coverage as a geo satellite. One launch may be cheaper, but SpaceX will need dozens.

You're correct about coverage -- 3 GEO sats can cover most of the planet -- but not about capacity, which kinda scales with satellite mass, and benefits from being closer (less power needed per bit.)

That is not correct. You have satellites of roughly the same mass over the last decade that are adding more and more capacity. You also do not benefit from being closer.

Why yes, I'd noticed high throughput satellites. And being closer means you can use less electricity on the satellite for the same signal-to-noise ratio, so yes, it is a benefit that comes from physics.

I kind of assume bandwidth is a function of satellite count. So a network with 1000 satellites will be faster than a network with 3.

User per satellite would also he important.

There's capacity and there's usable capacity. Geosynchronous, especially the newer ones, can put the capacity exactly where they need it. Leo can't. You end up with 70-90% of your capacity stranded unless you can sell it over water, even where planes don't go. So while spacex may have 3Tbps after 1000 satellites, the useable capacity will probably be less than a viasat-3 or any of the newer class satellites being built.