[spa3thyb]

More terrestrial infrastructure - fiber and radios - is an interesting 'Tech New Deal', but I look at just Starlink (and I hope there are more competitors with customers soon!) and see this area is already covered:

https://sebsebmc.github.io/starlink-coverage/index.html

If we're gonna spend 65B, could we throw 100M at buying everyone in the delta a dish?

[posguy]

Starlink has a finite amount of bandwidth, orders of magnitude less bandwidth than fiber. This is due to a restricted range of spectrum that will penetrate the atmosphere to their clients and an even smaller set which Starlink is allowed to use.

It is questionable whether their network could immediately handle 100k additional households, and by the time they scale the network to handle this, it is likely their existing 90k users will have continued with the yearly double digit bandwidth usage increase, eating much of the added capacity.

https://www.lightreading.com/4g3gwifi/starlinks-network-face...

[smoldesu]

Starlink is quite good. I've been using them for a few months now after switching from a more traditional, smaller WISP (and before that, godforsaken Hughesnet). It's really hard to beat Starlink, especially at the price. Unless you have a local telecom willing to run CAT6 out to your house, you can't really find speeds like it. Latency is low enough to game on, bandwidth is zippy (and unlimited), and the uptime puts 4G hotspots to shame.

If I had one complaint with it, it's probably the hardware itself. For $500, the installation kit is fairly barebones. On top of that, the router is pretty obviously "beta hardware" too, as well as the software and even parts of the dish itself. I'd be a little disappointed if I opened a Comcast installation kit with the same trappings, but I'll give Starlink some credit for pulling together such an impressive infrastructure/consumer hardware product at a non-alienating price.

[jagger27]

> Unless you have a local telecom willing to run CAT6 out to your house, you can't really find speeds like it.

Is CAT6 all that common? 100m runs seems pretty limiting in a rural setting. I think even the fibre that was strung up to my house in a fairly dense inner city neighbourhood has a longer run than that to the hub.

Is it possible to provide power to signal repeaters with POE to get around the ~100 meter limit?

[tim--]

> Is it possible to provide power to signal repeaters with POE to get around the ~100 meter limit?

Yes, but at that point you may as well just run fiber. These days, the SFP modules are cheap (20km SFP for ~$80).

The loss of signal on a CAT6 cable will be much greater then on fiber. When too much noise is introduced on a CAT6 cable, speeds will drop considerably. Packet loss will be quite high.

[jagger27]

That's what I figured. CAT6 isn't exactly cheap.

[chollida1]

Starlink has been incredibly unreliable for us in our testing. VPN's dropping alot. I'm sure its great for basic web surfing or things that don't require low latency but it has it issues for sure.

[fallingknife]

Their constellation is < 10% deployed, though.

[wmf]

Starlink needs to be reliable with just one shell; they can't wait years to become good enough.

[mdasen]

We kinda are. Starlink has won $885.5M in FCC subsidies already. However, as others have pointed out, Starlink doesn't have infinite bandwidth.

Starlink works because they're building a constellation of thousands of satellites. Previous satellite internet companies had comparatively few satellites so they needed to restrict how much you could use the internet with low data caps. Starlink still has capacity constraints.

Terrestrial wireless infrastructure is likely to be cheaper for the amount of capacity you get - especially when you consider that most of the cost will be shared with existing mobile networks.

With terrestrial wireless, it's relatively easy to split cells when you need more capacity. Wireless cells can cover hundreds of square miles in rural areas.

That's not to say that Starlink doesn't have a place, but it isn't a cheap service. $100/mo and $550 sign-up cost isn't cheap internet. $100M wouldn't buy a Starlink Dishy for even 200,000 people, never mind the monthly cost and never mind the fact that we're already spending $885.5M on Starlink.

$65B looks like a big number, but if we're talking 25M people it's only $2,600 per person and that pays for less than 2 years of Starlink per person (including the $550 startup cost). Plus, Starlink won't have capacity for 25M people. Elon Musk has said that they'll probably be able to serve the 500,000 preorders and that things get more challenging in the several million user range - never mind 25M users.

If we're talking 43M people like Broadband Now estimates or the 120M that Microsoft estimates, it gets even more clear that we'll need more than Starlink.

Starlink is a good way to serve extremely rural customers who likely won't even get decent wireless signals and it's a good way for SpaceX to get large government subsidies for providing service to rural areas. Starlink isn't a private company solving a problem for the public. It's public money being offered for a solution to a problem and SpaceX wanting to go after some of that money and hopefully create a decent business out of it.

[vel0city]

> $100M wouldn't buy a Starlink Dishy for even 200,000 people

By a large measure when you look at the actual costs and not just the amount they charge the customer. The dish's actual cost is >$1,000. So $100M really buys dishes for less than 100,000 people.