Could be more like hundreds of billions in the very long-term, as they plan to just keep growing it in size, capability, and per-satellite mass. Enabled by BFR.

That’d put SpaceX in league with other telecommunications giants.

That’s, of course, not the only possible outcome, but that’s why the valuation is so high.

Maybe... but I see the system being bandwidth constrained. If each spacecraft is 20 Gbps and how many will be visible over the USA at a time? 50 or so? So 1 Tbps for the entire country? That's less than the capacity of a single neighborhood in Denver on gigabit fiber. Not to mention the capacity of 4G networks across the country. Some of this smells like the first round of Iridium to me.

That's just the initial system, and they'll be selling globally, not just in the US. It's still about 10 times the aggregate bandwidth of their nearest competitor (OneWeb) plus the later constellation is going to be much larger and more capable.

Additionally, there should be about 150-200 satellites above the US in the initial deployment.

They also plan to upgrade the satellites themselves (according to SpaceX's Mueller) so they're much larger. That should, in principle, allow them to exceed the per-satellite throughput of GSO birds like Viasat's which are over 1Tbps per bird.

They aren't competing with fiber in Denver. They're competing with high-latency and expensive satellite internet in rural areas. It's possible they may eventually dip into the cities a bit if they sell out their capacity and then build out their VLEO constellation (total of 12,000 or so satellites, but in much lower orbits enabling yet lower latency and higher bandwidth).

But Iridium's problem wasn't a lack of capacity, it was slow customer acquisition combined with very high interest rates and launch prices. If SpaceX fills up their network with customers to its capacity, they will be just fine.

Iridium's problem was they didn't know who they were marketing service to. It seemed like a good idea before cell service was common and before roaming agreements. Those things were in place before the constellation was operational. So they tried to target international business travelers and failed. The restart of Iridium succeeded by selling to the US government. I wonder if Starlink has the same problem, at least in the US.

For rural areas, there are wireless ISPs for rural areas here in Colorado where you put an antenna on your roof. You get line of sight to an antenna on a nearby mountain for internet service. I don't see how Starlink can possibly compete with that on a technological or cost basis. It seems like most of the western US could have service like this for many rural markets, if it's not already available.

Regular satellite internet already provides price-per-gigabyte that is competitive (or superior to, in the case of Viasat) with mobile internet services. It's the latency that's lacking. With a LEO (and the then VLEO) constellation, Starlink could actually do better than some mobile internet providers as far as latency. 25-35 milliseconds for Starlink (maybe more in some cases) versus like 80-100 milliseconds for LTE providers (much worse for 4G and terrible for 3G).

So I definitely think Starlink could compete in that market. But more as an equal than as the only option.

Latency? No, the problem is price and data caps.

If SpaceX is going to make huge returns on starlink, they will have to be price competitive with at least the DSL providers that are in many rural communities. Many of your remote rural areas in the US have at least 6 to 10mbps for ~$60 a month, and that's plenty for most casual users. Even for video streaming.

Keep in mind that 6-10mbps isn't going to be enough for casual users in a few years. Take gaming for example. Companies are regularly releasing 6-10 GB patches willy nilly because they assume their customers can download them in a matter of minutes/hours.

And that's now, when there's still large urban areas in a lot of countries on DSL. Once that DSL is phased out, rural areas are going to need better sattelite options or expensive long stretches of fibre to not fall behind the curve.

Fixed wireless plans also have data caps, and you can get satellite unlimited plans affordably now (with soft caps, but so does most everyone else), at speeds greater than DSL.

The main issue for rural satellite access is no longer bandwidth or even data caps but latency.

And if you have DSL in rural areas, count yourself lucky.

I'm not talking about competing with mobile networks. I'm talking about fixed wireless service.

What about places with no mountains?

What about mobile users?

Antennas on towers, like for cell networks.

Starlink doesn't work for mobile users, either.

The key point is that, if the satellites are cheap enough, they're a safety valve for surplus rocket supply. Or padding if schedules don't make for full payloads.

Essentially seeding their own demand in a market that didn't previously exist (cheaper launch on reused rockets).

As a side benefit, they get the ability to put a constellation in orbit (and continue adding to it) at a fraction of the cost of their telecom competitors.

Yeah, if Starlink merely breaks even (and continues to grow), they will have pushed the commercial launch market potentially one or two orders of magnitude larger (in terms of mass) than it is today. It'll help pay for BFR, their Mars rocket.

Therefore, if Starlink suffers the same fate as Iridium (but without damaging the rest of SpaceX), that would actually be a big net win. Iridium emerged from bankruptcy, turned a profit and launched replacement satellites and is now in the process of a huge constellation refresh, using SpaceX's reusable rockets. If that happens with Starlink, that would be a complete success for creating launch demand.

I was thinking recently.. once you have data uplink anywhere on the planet, whats stopping them from blanketing the planet in cell phone towers and taking over that industry as well? They won't be limited by where you can run fiber like the other guys.

In super high density areas, fibre is going to be cheaper and faster and more easily maintained and upgraded etc than satellites

I think "super high density" is probably hyperbolic. As density increases shared airwaves become more crowded and wired infrastructure becomes cheaper to operate. I can see wired internet being competitive for a long time. Even in places more like St. Louis than Manhattan.

I choose where to live because I can have two different fiber connections (for speed and a failsafe). How long will it be before satellite internet offers 2gbps symmetrical with low latency at reasonable prices?

The use case of someone who needs two gigabit fiber connections to their residence is exceedingly tiny and would have no bearing at all on spacex's market.

I can hardly imagine a situation where internet access is so critical at a home that you need two fiber lines anyway. I have probably had 2 hours of downtime in the past 3 years with my cable connection.

Starlink is intended to provide 1Gbps service to end users with low latency, especially if you’re accessing sites on the other side of the world.

Are you camping out at the datacenter again?

Providing backhaul for rural cell towers is expected to be a big part of Starlink's business, yes. Same for O3b and Oneweb.