It's a SpaceX World (Everyone Else Is Playing Catch-Up) | Hacker News Reader

It's a SpaceX World (Everyone Else Is Playing Catch-Up)

astralytical.com

29 pointstontonius44 comments

They launch more satellites than everyone else combined, but a big caveat here:

"Over 99% of SpaceX’s 1,986 spacecraft deployed in 2023 were Starlink satellites."

EDIT: oh, I see the confusion. The 99% figure means "of the satellites operated by SpaceX, 99% are Starlink". It's 99% of the ones they operate, not 99% of the ones they launch. (Not sure if I'm using the right nomenclature)

On the other hand, launching anything at all at that cadence is something others aren't doing either.

I guess the question is, what happens after the Starlink constellation is complete?

They send replacements. Starlink satellites are in an orbit so low that trace atmosphere drag quickly plucks them out of the sky once they are out of fuel.

I wonder if it might be cheaper to send up a manoeuvrable "refuelling drone" to rendezvous with each one and top-up the fuel supplies. (The satellites would need to be designed to allow this, of course)

My KSP knowledge suggests that moving between orbits that aren't aligned costs a lot of fuel. The extra fuel you'd need to go from orbit to orbit might end up requiring more rocket capacity than sending up more duplicate dry mass of complete replacements. And that's before even considering updated satellites that are improvements in capability over the predecessors they replace.

When at one point I the future they stop expecting any further hardware improvements, they might of course change the satellites to a different balance between fuel and dry mass, sending less satellites per launch that last longer.

they have lifetime of 4-5 years - decaying orbit. It will never be 'complete'

Because they don't need to, not necessarily because they can't.

34 out of 96 launches involved 3rd parties though.

So there were at most 19 spacecraft deployed that were NOT Starlink satellites, although 34 launches involved third parties. How? How can I distribute 19 spacecraft on 34 launches?

> How can I distribute 19 spacecraft on 34 launches?

19 customers deployed spacecraft across 33 launches. The 99% figure is wrong.

I had misread it, it's 99% of SpaceX satellites are Starlink. As in, 99% of satellites launched/operated by SpaceX (e.g. the Tesla roadster being another).

Swarm Technologies

SpaceX has had 10 "Transporter" smallsat missions over the last four years that have each launched at least 40 non Starlink satellites and as many as 143 to sun synchronous orbit.

Note that some Starlink satellites may actually be Starshield satellites. SpaceX and the Department of Defense are being very secretive about Starshield.

I would not be surprised if there was no difference between the two. That is, at a moments notice, every single recent Starlink satellite has the capability to become a Starshield node with a single software deployment. It makes a lot of sense to just modify their existing design to be able to support both networks.

It's a good theory. I think Starshield has SAR, and SAR is a pretty big piece to leave unactivated in all Starlinks. But they could be doing that.

Good point! But I still wouldn't be surprised if they did it anyways. They may even get paid to do so.

How about free sats?

I believe there's SAR hardware on all v2mini sats period thanks to the thug shaker links

My suspicion is the Internet is a cover for SAR. Like the recently completed cable from Oman to Australia that's just a cover for getting Internet to Diego Garcia

All the Starlink launches make SpaceX the only company in the world that has a massive amount of spare launch capacity on a moment's notice. When a customer needs their payload to launch ASAP to start generating revenue in orbit, instead of sitting on Earth waiting for years for an available launcher, SpaceX are the only option. Payload integration would still take weeks to months, but it would be much faster than any competitor. Take, for example, OneWeb, who have been forced to seek another option after launching from Russia become impossible in 2022. They had to go to SpaceX, their formal competitor. Same for the EU space missions after their recent launcher troubles. Same for Amazon's Kuiper satellites, who bought every possible bit for spare launch capacity from every Western company for years to come, and yet ultimately had to turn to SpaceX, because no one else could offer the launch cadence they needed for their constellation deployment schedule.

> But the lack might also be because no customer wants their service associated with launches in which rocket bodies drop on the heads of China’s residents.

Nah its ITAR / (insert western country export control law here). Sending spacecraft hardware to Chinese launch provider will definitely violate all kinds of export control. Even if its a cubesat made from Raspberry Pis and assembled PCBs from JLCPCB

Nice data !

I'm a bit surprised by the % of flights / mass to orbit that's taken by Starlink actually - 2/3 of the launches and 80% of mass. If I read this correctly, if we exclude Starlink related launches, SpaceX is still big, but not that much bigger than the rest actually.

The title should almost be 'It's a Starlink World' !

It’s the 21st century’s railroad-tycoon advantage. Except we’re still in the first inning where the engines are getting exponentially better.

I've become very skeptical of Musk's companies frankly, so I'm not sure there is a real advantage here, from a business point of view. Would rather say it's a slightly red flag actually (to be very dependent on a single customer, and this customer being part of the same company).

> not sure there is a real advantage here, from a business point of view

You don’t see the advantage of replicating AT&T and Verizon’s non-urban tower advantage from space [1]? In-flight wi-fi alone is a $15 going on $40bn market [2].

To say nothing of the military advantages, given a world in which anti-satellite weapons are a given in the opening salvos of a war, of being able to rapidly replenish lost orbits.

[1] https://arstechnica.com/tech-policy/2024/01/spacex-launches-...

[2] https://www.marketsandmarkets.com/Market-Reports/global-wi-f...

Ah yes, the military business, it's pretty clear that it can be huge.

And wow ! Had no idea that in-flight wi-fi is such a huge market.

Before reusability, SpaceX was already the cheapest per KG

This is... interesting but not really the answer to the more interesting question. SpaceX undoubtedly dominates launch capacity today. It's basically the sole success in an industry that seems fundamentally uneconomical and the way it's acheived that is through government subsidy and Elon Musk's unparalleled fund raising. But I don't think that there's much of a moat here, hardware companies find this time after time - by building an incredible product what you're partly doing is training a team of engineers how to build that amazing product and those engineers can and will walk out the door and find a job down the street. Will SpaceX face strong competition within the next 18 months? Nah. 5 years? Probably. 10? Definitely. It's like Tesla, Tesla pioneered electric vehicles, but today it's not alone in the market, it's not even arguably the biggest player anymore.

What I'm far more interested in is... is there an actual market here? So far SpaceX makes money out of government subsidy, and Starlink. So to a large extent it is it's own biggest customer and because it's private it's difficult to really reason about it.

Like yeah, it looks really good for your rocket company if you have a customer that has to fire 1,000 new satelittes into orbit each year, that's going to make a tonne of volume. But it does kind of raise the question: Does that look like a great business from the internet service provider viewpoint?

How much capacity can you provide with these satellites? How often do they need replacing? And whose going to pay for satellite internet service?

The global space launch services market size was valued at $8.07 Bn in 2022 and is projected to grow from $9.15 Bn in 2023 to $20.54 Bn by 2030.

https://www.fortunebusinessinsights.com/industry-reports/spa...

> government subsidy

Government as customer != "government subsidy"

no. they get government grants and contracts.

The Commercial Resupply and Commercial Crew programs were intended to help develop unmanned and manned flight capability to the ISS, but they were contingent on a) demonstrating the technical capabilities to deliver and b) actually delivering. (Commercial Crew has become a gigantic money sink for Boeing, because unlike SpaceX the Starliner is many, many years late and the contract doesn't pay a cent more for delays/overruns.)

contracts = "government is a customer." the government wants a job done and you (the contractor) are paid to do it.

grants depend. grants can be "government wants someone to develop a technology/engineering capability and is paying you to do it if you can show that you're the best candidate". definitely a subsidy but not necessarily a handout.

There's a public subsidy tracker

The interesting thing is that they are still pushing the limits in new ways. Will there be someone else in 10 years that can do what they are doing now? Sure, but SpaceX will then also be ahead of current SpaceX by 10 years. I don't know for how long they'll be able to keep that up, but it's going to be interesting to see.

It does seem different to Tesla. Tesla's big innovation was building an electric car people actually wanted to buy. Everything afterwards looks like incremental improvement to me.

This take completely ignores that SpaceX launches at a fraction of their competitors costs. There are no other reusable rockets out there (or even coming soon).

Don't dismiss the flywheel effect of them moving so fast. They have a team of A+ players that will be difficult to replicate at other companies by just a handful of poached employees. I belive its the primary reason why Musk is able to work his employees so hard, pay them so little and still be one of the top desired companies to work for new engineering grads. As Steve Jobs once supposedly said, 'A' players want to only work with other 'A' players.

And speaking of Tesla, it remains to be seen whether any other automaker can match Tesla's engineering lead in EVs. There have been multiple false starts and failures. The only people I can see are the Chinese but that is a non-option for many western employees so Tesla remains on top.

>next 18 months? Nah. 5 years? Probably. 10? Definitely. It's like Tesla, Tesla pioneered electric vehicles, but today it's not alone in the market, it's not even arguably the biggest player anymore.

PRC has multiple private space companies with reusable vehicles on the way, a few look like they can do Falcon9 capabilities in 3-5 years. And if they decide to hit the overcapacity brrrt button, I imagine all the current wanking over SpaceX launch capacity will look very premature. There's ~40 falcon 9s doing concurrent missions, that's really not much in grand scheme of things. About "American" scale. I imagine once PRC tech is passable they'll scale to point to eclipse aggregate spaceX payloads rapidly fast with their own mega constellations, and magnitude more (like ship building) if there's ACTUAL case for sustaining that level of space launch demand. Which I hope there is... outside of militarization efforts.

Seems like mostly PRC is playing catchup, and if past performance any indicator, it'll take half the time for them to throw up X times more capacity to the point where current SpaceX lead won't mean much at all except politics behind orbit rights. IMO SpaceX more like EVs / shipbuilding, PRC already pretty decent at rocketry from MIC and communications tech - putting up stupid space launch capacity to throw their own mega constellations is a much easier/quicker catchup than commercial aviation or semiconductors that's mired in global geopolitics.

Are there estimates to how much it costs to refurb a falcon 9? It's gotta be at least 2 orders of magnitude cheaper right?

I don't think it's 2 orders of magnitude, my understanding (based on general spectating for many years) is that it's 0-1, i.e. the ~$80m rocket costs maybe just under $10m, maybe closer to $20m, somewhere in that ballpark. They have reduced the time taken to regularly under 1 month though now so it's possible this is now comfortably the lower end of that.

28 million in 2021 for refurb and launch

Reply on news.ycombinator.com