It might come down with more satellites and more breaking up service into different levels.
Imagine a cheap emergency-text-message level of service (like apple has, or maybe the garmin inreach type service). Then a compete-with-cellular cheap version for limited cellular service. Then a car service (premium-connectivity+satellite). then on up to residential broadband.
also, with more satellites might come the ability to talk to each other laterally, more load balancing and fewer ground stations. might be drive prices down.
More satellites increases the cost, not the reach, and so, not the customer base. I've worked in internet for 20+ years. The problem with "cheap emergency-text-message level of service" is not that it's not possible. It already exists (really Iridium offers that with a tiny modem that plugs into your cell phone). It's like observing that doing internet for your village or neighborhood is pretty easy and cheap, especially with cooperation from city hall, which is often a possibility. It's easy to beat incumbent telcos locally at internet services.
The argument incumbents make is that once that happens, the government will be responsible, first for paying for the entirety of telecom infrastructure outside of well-populated areas. Second governments need infrastructure wherever there is a border, or even the tiniest of villages, and getting dsl service in a border post will often mean paying for 100km of cable laying (or at the very least microwave towers).
And so governments make these services impossible at any scale through legislation. So either SpaceX pushes these through with "help" from Trump, or this doesn't happen.
There is a reason everybody RAN out of the global satellite data business in 2010 (and government bail outs mean that those people are still in business at $0 valuation). Satellite launches do not pay for the income that the market can provide. Then you calculate with Elon's numbers and well, frankly you get the usual outcome with Elon. "Yes, you improved the economics ... but not to the point of profitability, can you do another factor of about 2? No? Sorry then". And, yes, I realize that the problem with that factor of 2 is that it's not physically possible (technically SpaceX could get another 30% or so cost reduction with an RDE engine). That is why SpaceX owns Starlink. It's value is negative (meaning it would be stuck running a loss without access to SpaceX investors)
Imagine a cheap emergency-text-message level of service (like apple has, or maybe the garmin inreach type service). Then a compete-with-cellular cheap version for limited cellular service. Then a car service (premium-connectivity+satellite). then on up to residential broadband.
also, with more satellites might come the ability to talk to each other laterally, more load balancing and fewer ground stations. might be drive prices down.