[fdaslkjlkjklj]

For those who don't know, capital expenses are depreciated over their useful lifetime (for StarLink sats, estimated at 5 years) and aren't accounted for by cashflow.

Cash flow positive tells you they're not going to go broke, at least until they need more capital expenses (5 years...), which is a milestone, sure.

But profit is a better measure of a business's value as a going concern. What happens in 5 years when they can finance with interest rates at 5% and they have competition? I don't know... If they were profitable and could give returns investors needed, that would tell you they can survive.

[JumpCrisscross]

> capital expenses are depreciated over their useful lifetime (for StarLink sats, estimated at 5 years) and aren't accounted for by cashflow

This is incorrect. Cash flow is cash flow. Net income smooths capital expenditures. Cash flow does not. What that means is if you have a ten-year capital asset, year 0 cash flow will suck while year 1 will be exaggerated. (Starlink is currently in year 0 as it’s building out its constellation. Presumably, capex will fall when it’s in maintenance.)

[ckdarby]

I get where this comment is coming from, but isn't it much more complicated?

They're cashflow breakeven and could run net income negative over the next five years as a subscription market grab with the capex amortization over the 5 years.

It is possible they could become incredibly profitable when the capex investment in 5 years is much less than the initial capex.

Equipment costed $5 or $1/yr. Replacement equipment could cost $2.50 or $0.50/yr.

I see this plenty with business capex for on-prem/datacenters. What costed $5M over 5 years on a refresh replacement for 1:1 replacement is usually more than half the cost.

[Zigurd]

Starlink also has capex and subsequent opex for ground stations. Even with inter-satellite communications links in the 1.5gen satellites, it needs ground stations and interconnection with backbone operators.

It is unclear if, as bad as many terrestrial ISPs are, Starlink has a TAM big enough. It's going to be slower and less reliable than terrestrial wireless, and expanding terrestrial wireless coverage is going to be cheaper for a very large percentage of potential customers. Starlink is not competitive with terrestrial fiber.

[fragmede]

> expanding terrestrial wireless coverage is going to be cheaper for a very large percentage of potential customers.

If that's the case, why has it taken until SpaceX for there to be viable Internet service to many of the rural areas they service? People are replacing 3Mbit DSL links, or even dial up (yes, in 2023, dial-up) service with starlink. terrestrial fiber beats Starlink handily, except where it hasn't been run. Which, even many urban areas still don't have gigabit fiber Internet service available, never mind when suburban or rural areas will get it (if ever). The other part of the addressable market is airplanes and boats/ships, and remote bases like McMurdo in Antarctica. Not sure how you're proposing we run fiber Internet to them.

[Zigurd]

Because ISPs suck and they defrauded the government of the money that was supposed to pay for expanded broadband coverage.

Elon has a nose for picking feckless, evil, competitors: People hate car dealers. People hate their ISP. Elon absolutely saw these opportunities.

[linuxftw]

You only need downlinks to backbones if the internet is terrestrial. Who will be the first to offer the LEO datacenter?

[ckdarby]

They might be able to sustain just simply via country contracts using it as a breakglass in the event of a telecommunications attack.

I'm also seeing more commercial entities looking at it simply as a backup.

[Zigurd]

There are a lot of terrestrial options for backup links: optical, microwave, fixed 5G. Generally any terrestrial link, wired or wireless, will be both higher-capacity and more reliable.