New SpaceX Starlink Terminals, Satellites and Factories

Sam Korus of Ark Invest analysis shows SpaceX Starlink will be able to serve 30 million customers with 12000 Starlink Satellites. This would be $6 billion of revenue per year for SpaceX.

The key inputs for the Ark SpaceX Starlink model are:
The number of satellites in the constellation.
The bandwidth of each satellite, measured in megabits per second — Mbps.
The oversubscription ratio (not everyone in the same area is online at the same time).
The acceptable cost of broadband, measured as a percent of monthly GDP.
The minimum bandwidth in Mbps Starlink will provide.

SpaceX has a new Starlink user terminal and dish factory in Texas. SpaceX has designed a rectangular flat terminal that will be lower cost and higher performance. SpaceX should be making millions of terminals by the end of 2021.

SpaceX should have its first orbital Super Heavy Starship flight this month or next month. If Super Heavy Starship can start delivering 400 Starlink Satellites at a time then this will speed up deployment of the constellation by six times over the same number of Falcon 9 flights. The 12000 Starlink will probably be fully deployed sometime in 2023.

SpaceX has new Starlink satellites with laser connections and higher power.

SpaceX has filed to launch another 30,000 satellites. The further increase in satellites will quadruple revenue and the number of customers. SpaceX would need about 80 Super Heavy Starship launches for the additional 30,000 satellites. This should happen in 2024-2025. SpaceX with over $24 billion per year in revenue for Starlink and $4 billion per year in launch revenue would surpass the entire budget of NASA.

SOURCES – Ark Invest, Hyperchange
Written by Brian Wang,

16 thoughts on “New SpaceX Starlink Terminals, Satellites and Factories”

  1. sure, but until they have starship these are the numbers. As it is now they mathematically cannot get to the number they want. I obviously expect the costs to rise if you have a significant amount of launches because you are consuming more fuel and oxygen and in the end this will push the prices up

  2. I doubt an analysis based on only having Falcon 9 available for launching the satellites is reasonable. It seems rather likely that they will be able to use SuperHeavy/Starship to launch Starlink satellites around a year from now, and possibly sooner. That would change the economics quite a bit in SpaceX's favor.

  3. I am not convinced, here are some costs breakdown (not all just some of the main costs)
    Assuming each satellite costs 500k,12000 satellites cost spaceX 6bn
    Until SyperHeavy starts service they will still use Falcon 9 and with 60 satellites per launch it will takes 200 launches and a cost of 200x50M=10bn so to have your infrastructure up and runing it will costs spaceX 16 bn. LEO satellites have a 5 years lifespan due to orbital decay (confirmed by Gwynne shotwell) so every year you have to recover at least 16/5= 3,2Bn of costs just to replace deorbiting satellites. If the failure rate is higher than the deorbiting rate (you have failed satellites still orbiting for few years) and you will have higher costs. If your launch schedule is not tight enough (you will need 40 launches per year or one every 9 days to maintain a fleet of 12000 satellites) you will lose satellites at a higher rate than your replacement.
    If you want a network of 42000 satellites you will lose 8400 satellites per year eqiuvalent to 140 falcon 9 launches or one every 2.5 days. You get your replacement frequency lower than that and your infrastructure degrades. Quickly: (you lose a full falcon 9 launch worth of satellites every 2.5 days).
    So at full capacity the maintenance of the satellite network alone will cost 7Bn in launches and 4.2 bn in satellite hardware. No salaries, no other, service provider fees that the different nations might require, no IT costs. just to have stuff flying.

  4. There is a possibility there may be a substantial similarity between Gen2 sats and the SDA tranche 0 heavy sats that SpaceX is also building.

  5. He seems to have skipped a closing double quote for the href attribute of the first link, causing it to eat up the whole first paragraph and the beginning of the second, up to where the second link would be. It should be easy enough to fix for Brian or his web team.

  6. Brian, sometimes your posts are just broken or incoherent, like the beginning of this one. What is that first paragraph with the list supposed to be saying? The link you have there is broken too.

  7. It’s not clear that SpaceX plans to keep the current 250kg Starlink sat and just add laser links. There is some indication V2 will be much larger, like 4X the size, which would make sense with fitting 4 flat pack stacks of the V1.

    It’s also not clear that they plan on just launching Starlink by itself. They’ve already experimented with ride share mixed in, reducing the stack to accommodate customers.

    Starship might be launched on regularly scheduled at least weekly flights that can also accommodate pretty much any other satellites that want to book. It would mean the first scheduled regular service to orbit. If there are no paying customers it’s all Starlink.

  8. I think it would be nice to add the new antenna to my existing setup, creating an array. But that would only be worthwhile if the two antennae were able to track satellites independent of each other, so that there is no signal loss.

  9. Cities in Iran often see authorities pick a neighbourhood and swarm through it taking tv satellite dishes and fining users. It looks good that these dishes are fairly small and the new flat ones may be easier to camouflage but it doesn’t look like a complete solution to the problem. It will, however be great for rural populations in repressive states.


Leave a Comment