In a 2024 end-of-year memo, Gary Bolton of the Fiber Broadband Association said that FBA had partnered with the consulting firm Cartesian to look at the pros and cons of Starlink in the U.S. FBA says that report shows that Starlink currently has 1.4 million customers in the U.S., and with the current satellite constellation has the capacity to serve 1.7 million customers.
FBA is a pro-fiber trade association, and as such, it probably takes the most pessimistic look at satellite capacity. The short summary I’ve seen of the Cartesian report says that satellite broadband has some natural limitations on capacity. The implied conclusion of the report is that Starlink can’t serve everybody in rural America.
Starlink has never claimed that ability or goal. However, since there are those advocating that most of the $42 billion BEAD grant award should go to satellite broadband, it’s fair to assess Starlink’s capacity.
Starlink currently has 6,957 working broadband satellites with the stated goal is to grow to 30,000 satellites. That would be a 430% increase, and if the FBA claimed limit of 1.7 million U.S. customers is right, that implies a future capacity of more than 7 million U.S. customers when the constellation is completed – which could be even higher if new satellites have more capacity than older ones.
One of the more important FBA claim is that Starlink has a limitation on the number of people that can be served in any geographic area. That seems to be true today as evidenced by reports that Starlink has quietly implemented waiting lists for service in some parts of the country, presumably due to local capacity.
Just as with any ISP, Starlink also has potential limitations due to backhaul. Starlink currently shows 64 working ground stations, with plans underway to complete 99. A ground station is where broadband traffic passes back and forth between satellites and the terrestrial Internet. Starlink can obviously build more ground stations in the future as needed.
Perhaps the biggest constraint on Starlink is getting the needed spectrum to communicate between satellites and ground stations. Roger Entner of Recon Analytics was quoted recently as saying that Starlink doesn’t have enough spectrum today and new spectrum doesn’t seem to be likely over the next several years. Anybody who follows filings at the FCC has seen numerous filings made by Starlink and cellular companies over the last five years arguing about the allocation of spectrum. A lot of the spectrum that Starlink needs is also currently being used by the military and other parts of the government. It seems likely that Starlink will eventually get the spectrum it needs, but spectrum fights have never been resolved quickly, and this will be a slow struggle.
I have to agree with FBA that Starlink isn’t prepared to handle everybody in rural America today – something I’m sure Starlink would acknowledge. The question that FBA is raising is if satellite capacity can grow quickly enough to meet increasing demands from BEAD plus normal growth. The BEAD program gives ISPs four years to implement BEAD awards. Starlink’s first satellite launch was in May 2019. Who knows what the company can do in four more years?
Presumably the US is covered by certain Starlink orbits but not others. Is there any clue whether the 7000 / 30000 LEOs are evenly distributed across all its orbital planes, or concentrated on specific ones?
Something I don’t see mentioned much is the ongoing replacement of LEO birds. Happy to be corrected on these numbers but at a max of 30k birds in the sky and a lifespan of 5 years that is an average of 500 birds launched per month.
The Gen 3 satellite is reported to be around 4,188 lbs. With an estimated cost of $2.4M per ton delivered to LEO, that is $2.5 Billion per month to keep that 30k satellite network alive. Ongoing, forever, until it shuts down.
That’s my understanding as well. That’s a lot of rocket launches just to keep the size of the constellation.
My assumption is that Elon has a fond dream of it being a fully 100% automated process. Robotic plant churning out new birds, trucking them out to a launch site, loading them onto a rocket, launch, deploy, return and recover, re-outfit the rocket for the next launch. Wash, rinse, repeat. No human touchy at all. Possible? Yes. Affordable? Maybe.
At $125 / month it would take 20 Million subscribers to burn down that launch cost each month. I just don’t see how that is a workable situation.
absolutely, the inherent limitations of LEO. Starlink is simply not profitable on residential broadband and can’t be without a dramatic decrease in the cost of the satellites and launches. However, they have much more lucrative contacts than residential.
Residential services with Starlink are subsidizing other contracts and services out of the US. The US pays the highest rates (sound familiar?) while other nations are paying a fraction, even though we can do this math right here and show that even at $120/m it’s not profitable at 1.4M subscribers ($168M revenues on residential at $120/m, each satellite would need to cost less than about $24k for hardware AND launch….)
This is a heavily subsidized product and will continue to be forever because of these ongoing replacement costs and launches.
Fiber is absolutely the lowest cost in a short time period for long haul and middle mile. I’d say copper beats fiber by a bit in medium and high density when paired with fiber backhaul. Wireless is without a doubt the most economical way to deliver to low density, especial with modern tech and fiber backhaul. No where in this mix does LEO actually come out as economical except the very far flung places and <0.1% of otherwise unreachable homes or businesses.
Another item that is not discussed is Mean Time to Repair (MTTR). What happens when a solar flare damages a significant number of satellites? Or what if someone intentionally damages them? The Kessler syndrome is another concern with this technology. How long would it take to restore services? Fiber is not cheap in the short term, but it is the best long term solution.
The other network types that exist have less chance of a wide spread failure. There is almost no way to physically destroy a majority of the miles of fiber buried in the US, let alone the whole world. Certainly no natural disaster, it would have to be a man-made one, like a nuclear war.
Interestingly enough, DARPA created the Internet as a survivable network that would route around areas destroyed by nuclear detonations.
Yes, but was the “DARPA-net” created with a satellite constellation in mind? I bleieve itwas envisioned as a terrestrial network…
Yes. It was very early in satellite telecommunications. The first telecommunications satellite, Telstar, was put in orbit only a few years before.