Don't forget the US DOD who spent $300 million per satellite for 10 geostationary communication satellites that each have 2.4 Gbit/s of capacity [0]. That's $3 billion dollars on just the satellites and Starlink is being designed to have orders of magnitude more capacity (21 Gbit/s per satellite * 12,000 satellites vs 2.4 Gbit/s * 10 satellites), lower latency (LEO vs GEO), and way more redundancy.

They've also already awarded spacex a $28.7 million dollar contract for "experimentation in the areas of establishing connectivity, operational experimentation, and special purpose experimentation. Experimentation will include connectivity demonstrations to Air Force ground sites and aircraft for experimental purposes" [1] so they are definitely eyeing it.

[0]: https://en.wikipedia.org/wiki/Wideband_Global_SATCOM

[1]: http://www.parabolicarc.com/2018/12/27/spacex-receives-287-m...

If all the constellation + GEO systems proposals end up being launched, there will be an oversupply of capacity from space.

Rural areas are a viable market only in the US and Canada, where people can afford to pay $50+/month for connectivity. Elsewhere in the world, they are not going to sell to any end-users.

The 4,425 and the 7,500 satellite constellations proposed by SpaceX are way oversized and simply put, a waste of money and efforts.

The mesh of satellites favor the northern hemispheres, you have a much lower density in lower regions, so you could charge less for them.

And they could stop launching them if they don't see a supply, basically just matching their registration numbers and launching more when the bandwidth/users require it.

Those numbers seem to just be a target for an ideal completed system, not for it to become operational.

There are essentially no variable costs for the satellites, and since they will be in LEO, a satellite to supply rural US also supplies everywhere in India. So SpaceX can sell their services for $5/month (or less) to people in India and $50/month to people in the US. Everyone (within certain longitude bounds) is a possible customer.

At $10 billion to deploy and lets assume 5 years satellite lifetime and $1 billion a year to operate, this is $3 billion a year or $250 million / month. You would only need 5 million subscribers at $50 / month to break even.

Seriously. I’ll switch just so I don’t have to pay Comcast.

Now I just need some Tesla solar panels and a couple powerwalls, and betweeen that, a well, my septic system, I can cut every cable coming into my house.

Well, I do use natural gas for heating and I love my gas stove, but nearly there!

If only the receiver and the power requirements were small enough to fit in my cell phone!