Impact of SpaceX rideshare on small sat launchers market

[TrevorMonty]

I think its kinda funny how allot of small launch providers are complaining about transporter eating into their market share when a majority of these small launch companies are launching less rockets in a year than SpaceX launches transporter missions

Electron is booked solid for 2024. As for other small LVs, lack of execution has been more of an issue than lack of customers.

[TrevorMonty]

But that has to balance out against Starship lift being so cheap a sat could purchase an OTV tug service with ludicrous deltaV to get where they want and still be cheaper as long as they get dropped off at vaguely polar SSO orbit. Single lift constellations could become a thing.

OTVs aren't free, they may even cost more than satellite plus add another potential point of failure. The other alternative is more expensive satellite with extra DV. This becomes more attractive if in orbit refuelling is option. We aren't there yet but there is move to fit refuelling adaptors to new satellites to future proof them.

 

[Exastro]

But that has to balance out against Starship lift being so cheap a sat could purchase an OTV tug service with ludicrous deltaV to get where they want and still be cheaper as long as they get dropped off at vaguely polar SSO orbit. Single lift constellations could become a thing.

OTVs aren't free, they may even cost more than satellite plus add another potential point of failure. The other alternative is more expensive satellite with extra DV. This becomes more attractive if in orbit refuelling is option. We aren't there yet but there is move to fit refuelling adaptors to new satellites to future proof them.

Without having done the analysis, a couple of possible Starship scenarios seem like they might plausibly turn out to be efficient:

1) Starship launches many payloads that need to go to somewhat similar (meaning accessible with small to moderate delta-V) orbits, along with a set of methalox tugs that get fueled from the residuals in Starship's tanks.  These perform the 'last-mile' service for the payloads.

  Option a.  The tugs are expendable but mass-produced, so they're cheap
  Option b.  The tugs are reused by returning to the Starship for return to Earth
  Option c.  The tugs are reused by having them maneuver into clusters which can be retrieved by a Starship at some arbitrary later time.  These clusters will contain tugs from more than one Starship launch, so they are big enough to be worth retrieving.
  Option c2.  Instead of clustering in open space, the tugs maneuver into unpressurized shelters that protect them from MMOD, large temperature swings, and radiation while they wait for retrieval.  This might be a minimal version of the 'orbital garage' concept, in which the shelter is a long, open-ended cylinder gravity-stabilized so that its open end faces Earth.

2) Starship delivers a similar group of payloads to an orbital station, where it offloads excess propellant.  The station maintains a set of reusable methalox tugs that perform the last-mile service and return to the station.  These are fueled with the offloaded Starship propellant.

Option a.  The orbital station is one of many, which collectively serve a variety of orbits.  It's unmanned and highly automated, and stays specialized.
  Option b.  The orbital station grows to be come a large, manned research and/or manufacturing hub and interplanetary spaceport. It sells propellant to other users also, partly offsetting its costs. It may also collect upper stages from various launchers for return to Earth, enabling those vehicles to become fully reusable.

Crude economic case:  A future small LV might put 1 tonne into LEO for $5M, assuming some technological progress relative to what's expected to be available in a few years.  Starship is targeting $30/kg but surely won't get there for some time yet.  So guess it costs $100/kg, or $100K to get the same 1 tonne payload to LEO.

So that leaves $4.9M for the last-mile service.  To estimate the cost of the tug, start with the ~$10M cost of F9 S2.  Figure the mass and thrust of the tug will be around 1/10 of that, and suppose its delta-V is similar.  Make the wild guess that the cost of the stage goes like the square root of its mass.  So we're looking at maybe $3M for the tug.  The tug's propellant mass should scale roughly with the payload mass, so about 5% that of F9 S2, or 5 tonnes, and its cost (in LEO) is $0.5M.  So the total cost of the delivery to the desired orbit is (0.1M + 3M + 0.5M) = $3.6M, plus the cost of operations at the station (if used), minus the value of the returned tug (if reusable).

I think the case for the tug is iffy if it's expendable, and may be compelling if it's reusable.

[trimeta]

But that has to balance out against Starship lift being so cheap a sat could purchase an OTV tug service with ludicrous deltaV to get where they want and still be cheaper as long as they get dropped off at vaguely polar SSO orbit. Single lift constellations could become a thing.

OTVs aren't free, they may even cost more than satellite plus add another potential point of failure. The other alternative is more expensive satellite with extra DV. This becomes more attractive if in orbit refuelling is option. We aren't there yet but there is move to fit refuelling adaptors to new satellites to future proof them.

Don't forget also that OTVs have mass. So when thinking about "the only metric that matters is $/kg," if a massively-rideshared launch puts you in the wrong place and you need an OTV to correct your orbit, you're paying for that OTV's mass as well, not just the mass of the payload.

[DanClemmensen]

What is the projected marginal cost of a Starship launch?  I think it's less than $10 Million.

What is the largest payload that can be launched by any projected LV for less than $10 Million?

For any payload that is larger than this number, a dedicated Starship launch will be cheaper than the dedicated small launcher.

That is the most optimistic case. For any form of rideshare, the numbers get better for Starship.

Costs are not prices, and payload mass is not the only metric, but I think these two are the fundamentals.

[M.E.T.]

What is the projected marginal cost of a Starship launch?  I think it's less than $10 Million.

What is the largest payload that can be launched by any projected LV for less than $10 Million?

For any payload that is larger than this number, a dedicated Starship launch will be cheaper than the dedicated small launcher.

That is the most optimistic case. For any form of rideshare, the numbers get better for Starship.

Costs are not prices, and payload mass is not the only metric, but I think these two are the fundamentals.

I’m comfortable with $30M per SS launch, for the next 3 years or so until full and rapid upper stage reuse is perfected. So that means expending SS but recovering Super Heavy, with a payload capability of about 150-200t in that configuration.

I think this is highly realistically achievable within the next year - more or less when Neutron might optimistically have a first launch.

So, in this scenario we have $30M, 150t Starship competing with $50M, 13t Neutron by say end of 2025. Give Starship a $20M markup so we have a $50M launch price for both rockets, one launching 150t and one 13t, both recovering their booster stages.

In this context, Starship matches Neutron on price for any payload between 0-13t, and has no competition from Neutron for any payload above 13t. That’s without rideshare.

Add any rideshare options and Neutron gets obliterated for payloads under 13t as well.

Not much of a business case for Neutron when looking at it like that.

[TrevorMonty]

What is the projected marginal cost of a Starship launch?  I think it's less than $10 Million.

What is the largest payload that can be launched by any projected LV for less than $10 Million?

For any payload that is larger than this number, a dedicated Starship launch will be cheaper than the dedicated small launcher.

That is the most optimistic case. For any form of rideshare, the numbers get better for Starship.

Costs are not prices, and payload mass is not the only metric, but I think these two are the fundamentals.

I’m comfortable with $30M per SS launch, for the next 3 years or so until full and rapid upper stage reuse is perfected. So that means expending SS but recovering Super Heavy, with a payload capability of about 150-200t in that configuration.

I think this is highly realistically achievable within the next year - more or less when Neutron might optimistically have a first launch.

So, in this scenario we have $30M, 150t Starship competing with $50M, 13t Neutron by say end of 2025. Give Starship a $20M markup so we have a $50M launch price for both rockets, one launching 150t and one 13t, both recovering their booster stages.

In this context, Starship matches Neutron on price for any payload between 0-13t, and has no competition from Neutron for any payload above 13t. That’s without rideshare.

Add any rideshare options and Neutron gets obliterated for payloads under 13t as well.

Not much of a business case for Neutron when looking at it like that.

Based in your prices either RL staff are paid to much or SpaceX staff way to little.

Both vehicles are being built in USA ie same labour and suppliers market.
How can 30 engine booster and expendable 6 engine US be considerably cheaper to operate than 9 engine booster and single engine US.

Its like say boeing 737 is cheaper to operate than considerably smaller cessana skycourier.

[chopsticks]

People have cognitive dissonance with believing low cost figures from SpaceX but always assume high dollar for any competitor.

[M.E.T.]

What is the projected marginal cost of a Starship launch?  I think it's less than $10 Million.

What is the largest payload that can be launched by any projected LV for less than $10 Million?

For any payload that is larger than this number, a dedicated Starship launch will be cheaper than the dedicated small launcher.

That is the most optimistic case. For any form of rideshare, the numbers get better for Starship.

Costs are not prices, and payload mass is not the only metric, but I think these two are the fundamentals.

I’m comfortable with $30M per SS launch, for the next 3 years or so until full and rapid upper stage reuse is perfected. So that means expending SS but recovering Super Heavy, with a payload capability of about 150-200t in that configuration.

I think this is highly realistically achievable within the next year - more or less when Neutron might optimistically have a first launch.

So, in this scenario we have $30M, 150t Starship competing with $50M, 13t Neutron by say end of 2025. Give Starship a $20M markup so we have a $50M launch price for both rockets, one launching 150t and one 13t, both recovering their booster stages.

In this context, Starship matches Neutron on price for any payload between 0-13t, and has no competition from Neutron for any payload above 13t. That’s without rideshare.

Add any rideshare options and Neutron gets obliterated for payloads under 13t as well.

Not much of a business case for Neutron when looking at it like that.

Based in your prices either RL staff are paid to much or SpaceX staff way to little.

Both vehicles are being built in USA ie same labour and suppliers market.
How can 30 engine booster and expendable 6 engine US be considerably cheaper to operate than 9 engine booster and single engine US.

Its like say boeing 737 is cheaper to operate than considerably smaller cessana skycourier.

I’m quoting RL’s own price target for Neutron, which is $50m. Their markup is unconfirmed but I believe they have referenced a 50% profit margin in one of their presentations. So I guess they are targeting ~$25M internal cost per Neutron rocket. But they need that profit margin desperately, as they don’t have economies of scale and are still building out their capacity.

As for Starship - Grok states that a Raptor engine costs $250k-300k. So that’s less than $2M for all 6 engines on the Starship.

After that you have cheap stainless steel and a mass production line which minimizes labour time in popping out successive Starships like a sausage machine.

I think $30M is a reasonable cost for a Starship flight with booster recovery. It might even be lower.

[JayWee]

People have cognitive dissonance with believing low cost figures from SpaceX but always assume high dollar for any competitor.

Economies of scale.
A lot of the SpaceX infrastructure is paid for by the Starlink launches. Keep in mind that Starlink V2 (the full 40k 1.25t constellation) requires like 300-600 launches of SH/SS.
That a) puts extreme pressure on cost control and b) makes the one additional sold launch marginal cost very small.

[DanClemmensen]

Based in your prices either RL staff are paid to much or SpaceX staff way to little.
Both vehicles are being built in USA ie same labour and suppliers market.

Build cost contributes little to marginal cost if the reuse is high enough. Build cost/unit goes down as production rate goes up. SpaceX is trying very hard to reach a very high reuse number, and the have a large captive customer that keeps the production rate up.

How can 30 engine booster and expendable 6 engine US be considerably cheaper to operate than 9 engine booster and single engine US.

Launch cadence affects marginal per/launch cost, as you amortize fixed costs. Again, a captive high-volume customer drives SpaceX' high cadence.

SpaceX has designed Starship for low cost of operations, with RTLS for both stages and rapid reuse and minimal refurbishment costs.

You are correct: with an expendable SS, the cost of the expended SS is likely higher than the cost of an expended F9 US, but there won't be all that many expended SS. Expended SS are basically SS reuse tests.

[DanClemmensen]

People have cognitive dissonance with believing low cost figures from SpaceX but always assume high dollar for any competitor.

I think one difference is that we can observe the actual Starship manufacturing process and we observed how it evolved. The SpaceX numbers seem to be consistent with these observations, and we see the actual flight results. Most of the non-SpaceX numbers are about the future, not the present.

[greybeardengineer]

People have cognitive dissonance with believing low cost figures from SpaceX but always assume high dollar for any competitor.

Sounds like those people are well acquainted with
1) economy of scale
2) design for manufacturing
3) process learning curve
4) continuous improvement
5) vertical integration

Too bad SpaceX's major competitors aren't.

[trimeta]

People have cognitive dissonance with believing low cost figures from SpaceX but always assume high dollar for any competitor.

Sounds like those people are well acquainted with
1) economy of scale
2) design for manufacturing
3) process learning curve
4) continuous improvement
5) vertical integration

Too bad SpaceX's major competitors aren't.

Well, one company which aspires to build a medium-lift launch vehicle (I wouldn't argue that they're a "competitor" of SpaceX, in the sense of forcing SpaceX to make changes to their plans or practices, because I'm not sure one could claim that any company truly competes with SpaceX in that way) is familiar with all of that. For a small-lift vehicle.

Whether those lessons in general scale up is another question.

[darkenfast]

It seems to me that there's two types of small-sats: Those that just need to get into LEO, do their job and decay, and those that need a specific customizable orbit that can't be met that way. Let's say you are a university that wants to fly a research sat to monitor unicorn propagation across the globe. What if, instead of building a small-sat, you only had to supply the unicorn-detector? SpaceX (or contractor), books an Earth facing slot on a temporary, unmanned "rack". The university supplies the instrument and the "rack" supplies the mount and electricity. It has solar cells, an emergency de-orbit system, and a variety of angles for mounting instruments (based on market). It's launched on a Starship, and then, after a stated period, is retrieved by another Ship. Both Ships may have other missions during those flights. The University gets its instrument back, perhaps to use again next year.

Even if the launch cost was the same as for a Transporter, there would be a reduction in cost compared to having to build and qualify the satellite, and having to deal with any bureaucracy for that. The instruments can be larger, perhaps. 

[trimeta]

It seems to me that there's two types of small-sats: Those that just need to get into LEO, do their job and decay, and those that need a specific customizable orbit that can't be met that way. Let's say you are a university that wants to fly a research sat to monitor unicorn propagation across the globe. What if, instead of building a small-sat, you only had to supply the unicorn-detector? SpaceX (or contractor), books an Earth facing slot on a temporary, unmanned "rack". The university supplies the instrument and the "rack" supplies the mount and electricity. It has solar cells, an emergency de-orbit system, and a variety of angles for mounting instruments (based on market). It's launched on a Starship, and then, after a stated period, is retrieved by another Ship. Both Ships may have other missions during those flights. The University gets its instrument back, perhaps to use again next year.

Even if the launch cost was the same as for a Transporter, there would be a reduction in cost compared to having to build and qualify the satellite, and having to deal with any bureaucracy for that. The instruments can be larger, perhaps.

I'm a bit skeptical of the "retrieve and return" side of that, but the concept of moving from rideshares and OTVs to having multiple payloads permanently affixed to a single satellite bus (which provides all power/comms/propulsion) for the full duration of the mission seems like an obvious direction which the industry will move in. So long as there are enough payloads which don't care where they are in space, beyond "500km SSO seems fine."

[darkenfast]

It seems to me that there's two types of small-sats: Those that just need to get into LEO, do their job and decay, and those that need a specific customizable orbit that can't be met that way. Let's say you are a university that wants to fly a research sat to monitor unicorn propagation across the globe. What if, instead of building a small-sat, you only had to supply the unicorn-detector? SpaceX (or contractor), books an Earth facing slot on a temporary, unmanned "rack". The university supplies the instrument and the "rack" supplies the mount and electricity. It has solar cells, an emergency de-orbit system, and a variety of angles for mounting instruments (based on market). It's launched on a Starship, and then, after a stated period, is retrieved by another Ship. Both Ships may have other missions during those flights. The University gets its instrument back, perhaps to use again next year.

Even if the launch cost was the same as for a Transporter, there would be a reduction in cost compared to having to build and qualify the satellite, and having to deal with any bureaucracy for that. The instruments can be larger, perhaps.

I'm a bit skeptical of the "retrieve and return" side of that, but the concept of moving from rideshares and OTVs to having multiple payloads permanently affixed to a single satellite bus (which provides all power/comms/propulsion) for the full duration of the mission seems like an obvious direction which the industry will move in. So long as there are enough payloads which don't care where they are in space, beyond "500km SSO seems fine."

Your skepticism is warranted, and my brief statement "is retrieved..." doesn't cover how much development is involved in that process. However, sooner or later the process of bringing back free-flying payloads will need to be worked out, and I think SpaceX and Starship will have the best shot at it.

[TrevorMonty]

It seems to me that there's two types of small-sats: Those that just need to get into LEO, do their job and decay, and those that need a specific customizable orbit that can't be met that way. Let's say you are a university that wants to fly a research sat to monitor unicorn propagation across the globe. What if, instead of building a small-sat, you only had to supply the unicorn-detector? SpaceX (or contractor), books an Earth facing slot on a temporary, unmanned "rack". The university supplies the instrument and the "rack" supplies the mount and electricity. It has solar cells, an emergency de-orbit system, and a variety of angles for mounting instruments (based on market). It's launched on a Starship, and then, after a stated period, is retrieved by another Ship. Both Ships may have other missions during those flights. The University gets its instrument back, perhaps to use again next year.

Even if the launch cost was the same as for a Transporter, there would be a reduction in cost compared to having to build and qualify the satellite, and having to deal with any bureaucracy for that. The instruments can be larger, perhaps.

I'm a bit skeptical of the "retrieve and return" side of that, but the concept of moving from rideshares and OTVs to having multiple payloads permanently affixed to a single satellite bus (which provides all power/comms/propulsion) for the full duration of the mission seems like an obvious direction which the industry will move in. So long as there are enough payloads which don't care where they are in space, beyond "500km SSO seems fine."

Having multiple payloads on single platform is called persistent platform.

https://www.nasa.gov/nexis/isam/

There was talk of developing GEO ones but need robotics and space tug to take playload from LV to platform and install it.

[imprezive]

It seems to me that there's two types of small-sats: Those that just need to get into LEO, do their job and decay, and those that need a specific customizable orbit that can't be met that way. Let's say you are a university that wants to fly a research sat to monitor unicorn propagation across the globe. What if, instead of building a small-sat, you only had to supply the unicorn-detector? SpaceX (or contractor), books an Earth facing slot on a temporary, unmanned "rack". The university supplies the instrument and the "rack" supplies the mount and electricity. It has solar cells, an emergency de-orbit system, and a variety of angles for mounting instruments (based on market). It's launched on a Starship, and then, after a stated period, is retrieved by another Ship. Both Ships may have other missions during those flights. The University gets its instrument back, perhaps to use again next year.

Even if the launch cost was the same as for a Transporter, there would be a reduction in cost compared to having to build and qualify the satellite, and having to deal with any bureaucracy for that. The instruments can be larger, perhaps.

Most small sats go to SSO in LEO. It’s a unique orbit that gives both continuous coverage and continuous solar power on a satellite (so small panels, smaller batteries, and less cost). It’s the reason rideshare works so well. It’s not just cost is the physics of orbits pushing most satellites to the same orbit.

[DanClemmensen]

It seems to me that there's two types of small-sats: Those that just need to get into LEO, do their job and decay, and those that need a specific customizable orbit that can't be met that way. Let's say you are a university that wants to fly a research sat to monitor unicorn propagation across the globe. What if, instead of building a small-sat, you only had to supply the unicorn-detector? SpaceX (or contractor), books an Earth facing slot on a temporary, unmanned "rack". The university supplies the instrument and the "rack" supplies the mount and electricity. It has solar cells, an emergency de-orbit system, and a variety of angles for mounting instruments (based on market). It's launched on a Starship, and then, after a stated period, is retrieved by another Ship. Both Ships may have other missions during those flights. The University gets its instrument back, perhaps to use again next year.

Even if the launch cost was the same as for a Transporter, there would be a reduction in cost compared to having to build and qualify the satellite, and having to deal with any bureaucracy for that. The instruments can be larger, perhaps.

Most small sats go to SSO in LEO. It’s a unique orbit that gives both continuous coverage and continuous solar power on a satellite (so small panels, smaller batteries, and less cost). It’s the reason rideshare works so well. It’s not just cost is the physics of orbits pushing most satellites to the same orbit.

Those SSO planes could also be useful for Starlink: the planes, not the SSO altitude. With Raptor restart, it should be possible to deploy Starlinks at their lower altitude and then boost to SSO and deploy the SSO satellites. Plane changes are expensive. Altitude changes are much less expensive.