That would definitely save some money on launch costs. And if launches are cheap enough, then it becomes more practical to do in-orbit assembly and refueling of spacecraft.
Couple that with reuse of the Dragon V2 capsule and add to that a “full scale” Bigelow Aerospace LEO station and we could be witnessing the beginnings of a commercial space station program open to more than just government astronauts.
I found the presser very interesting. Elon has cleared up a lot of the uncertainty about F9 reusability with his answers and I think we’re getting a significant insight into his end game.
You amortize the rocket’s cost over the life of the rocket. If the rocket first stage cost 16 million and you fly it ten times your cost of the first stage is 1.6 million per flight…
Broadly correct although you’d also need to add about 50% to the construction cost for refurbishment and flight ops over the lifetime of the vehicle. Even if I’m underestimating that by a factor of two or three, we’re still looking at a core that only adds $5M to the launch cost per mission. No conventional ELV would be able to match that.
It could be either more or less than the cost you estimate. I suspect NASA would insist on a new first stage for ISS missions (more likely for crew transfer than cargo.) I also suspect they would not object it that cost more. On the other hand, if the rated lifetime is 20 launches, and that stage is on number 19, I suspect most customers (and their insurance companies) would need some encouragement. I so I could see a higher price for a new one and a lower price for a very used one.
Rockets only operate for 15 minutes or so. 15 minutes! And although I’ve pointed out that I am not a rocket scientist, I wonder if a rocket is more complex than, say, a 787?
It’s true that the short time rockets are alive is a bone-jarring and very high energy 15 minutes, but still.
The notion that rockets will somehow ‘wear out’ rapidly is, if you will forgive me, Old Think. Yes, airplanes do wear out; but rockets could be subject to the same rigid procedures governing modern airliners, giving them for all practical purposes an unlimited lifetime.
At some point they will simply be called ‘boosters’ and will be easily available.
I’d like to note that there is nothing fundamental about a liquid fueled rocket engine that makes it “single use”. In fact, the Merlin engines on a “new” stage are test fired more than once before a mission (once in Texas and once on the pad). So, the engines are already being “reused” on each and every flight.
Traditional rocket engineering is that if the rocket is not “worn out” after those 15 minutes, before you drown it in the atlantic, than it is overbuilt. That is, as you say, “Old Think”. Now we are in a new paradigm of “New Think” how do I have to engineer an engine to use it again and again, the tankage, the avionics.. Now we are moving to the over build it rather than under build it.
I’ve read several commentaries that question Economics of rocket re usability. There critique rests on two arguments. 1. The refurbishment costs of a used rocket are nearly as much as a new rocket. This seems to be based on the Space Shuttles main engines which had to be nearly rebuilt each time.
2. You have to use a much bigger rocket to put a payload into orbit to account for the fuel to return to earth. Why not spend that fuel on bigger payloads?
I think the answer to these critiques is the same. The Falcon 9 is much more rocket than you usually need for the payloads that it is putting into orbit. Traditionally, if you wanted to put 1000 pounds into orbit, you put it on a rocket that can put 1001 pounds into orbit. That makes sense if you have to throw the rocket away afterward. A 1001 lb payload rocket is supposedly cheaper than a 2000 lb payload rocket. But, a 2000 lb payload rocket is cheaper is you can use it 20 times and you only pay for the fuel. Look at all the variations of the Atlas series of rockets.
That fact that the Falcon 9 is over built compared to other rockets is also the answer to the refurbishment problem. Running a rocket at 99% of its capacity for 10 minutes is going to wear it out. If it doesn’t wear it out than you should make the components lighter which gives you increased payload and reduced costs which is what you want to do if you are going to throw the rocket away. When the goal is to reuse the rocket, the formula changes. Yes, you are giving up payload but as long as there are payloads out there that you can lift you have a viable product.
The “Old Space” people got stuck in circular thinking.
1. You have to use the smallest rocket you can because you can’t recover it. 2. You can’t recover your rocket because you would have to build it too big for the payload and that would violate the first rule.
Plus the critics may be assuming that the stage must be “refurbished” after each flight (similar to what the space shuttle orbiters or SRBs needed to have done). From Musk’s tweets above, it’s clear that he’s hoping to re-fly an individual stage 10 times before “refurbishing” it. That is resuability.
There is also a third critique to the reusability question, and that is the one of market elasticity. Various articles through the years have argued that you need approximately 50 flights a year for reusability to pay for itself. Elon’s statements on how many times he hopes to reuse a Falcon and SpaceX’s stated flight rate certainly point toward a number similar to that. BUT – will there be 50 payloads a year? How elastic is the market? How far do prices have to drop for demand to significantly increase?
If you recall, 20 years ago we were going to have massive constellations like Globalstar and Teledesic which in turn would result in amazingly low cost vehicles derived from Pegasus, Amroc, and Conestoga. The launchers would be cheap (and ultimately reusable) because there would be such high demand — but the demand would be so high because launches would be so cheap. This was the proverbial chicken and egg of the last wave of New Space in the 90s. Elon has jump started that process by injecting a lot of his own money. However, is that enough to really bring the demand up to the levels that make reusability sustainable?
The GEO market is quite inelastic to launch costs. About 2% of DirectTV’s operating costs are satellite launches. Intelsat’s launch costs may be a touch higher, but not significantly so. We cansafely say a decrease in launch costs of 100% leads to essentially 0% growth in demand. Giving away launches for free would not really impact their business plan.
NASA science missions are a bit more sensitive to launch costs. About 10-20% of their costs are in the launch, so significant reduction in launch costs makes a difference. However, NASA’s budget is fixed, so even if you give away launches, you only get one new mission for every five missions. You could argue that perhaps total spacecraft costs also come down as a result of willing to accept more risk in order to significantly lower price, but I still can’t see much more than one new mission for every three existing missions or so. So a decrease in launch costs of 100% leads to about 30% increase in demand. Not nearly enough to get you to the volumes where reusability is economically feasible.
For the large remote sensing systems, the numbers are probably similar to the Geo comm market. For smaller constellations such as PlanetLabs and Skybox, the elasticity does start to go up significantly. However, their satellites are a lot smaller and the Falcon 9 is significantly oversized. If it is cheap enough, that may not make a difference, but I’m not sure what “cheap enough ” is.
Maybe the elasticity is in CubeSats! Spaceflight just sold an entire Falcon 9 to CubeSats and other small nanosats. However, how often do 80+ satellites need to go to the same orbit? And how often can you even just identify 80+ satellites needing to get into space? I’d venture to say that twice a year is the most we can expect from massive aggregated launches.
In the end I think we are left with just two areas where the market might be elastic enough to feed the demand required to make reusability economically feasible:
1) Large Communications constellations (OneWeb, Google, Facebook). However, reference the 1990s. And how many constellations can we realistically have before competition, regulatory concerns, debris mitigation, and frequency availability become an issue
2) Very large space logistic efforts: moon base, Martian colony, asteroid mining. However, even with launch vehicle reusability these projects are so capital intensive that they are likely to require much more than just Elon’s riches.
A bit negative, but still; truthfully we’ve not really identified the WHY of space travel very well. I can’t verify the numbers: DirectTV’s costs at 2% for launch seems mighty low, but then again, the savings are still in the tens of millions.
But I do think that many grandiose space plans are based on sandy foundations- and I’m looking at you, Mission To Mars. Same goes for the Lunites, although the case for a research base on the moon is easy to make. Certainly a lot easier to make than a mission to Mars.
And this: the chicken and the egg are still in charge here. Real money will be made when we learn how to actually live in space; how to use the rich resources available in asteroids, how to mine, smelt, and manufacture; how to construct space habs; how to live in space without regular departures to deep gravity wells.
I agree that elasticity of the market is a big deal, but I’m not sure I agree about the details.
I’m surprised by the 2% number for DirecTV’s spending on launches. It seems to be either too low or too high. I don’t know about their business model, but how deeply are they involved in owning and operating satellites? I thought a common practice was for service providers to rent transponders on a satellite launched and operated by someone else. (I need to learn more about this, since I’ve got an idea for an instrument which would fit well as a hosted payload on an SSL communication satellite, and I’d like to know how they can offer that service if they don’t retain ownership of the satellite.)
In any case, the same point you mentioned for scientific spacecraft applies to commercial ones. If the launch costs go down, the need for long lifetimes and high reliability also goes down, and the cost of building commercial satellites with it. If replacement of assets is economical the demand for launches will increase.
As far as NASA’s scientific missions go, my guess would be a factor of five to ten, rather than the factor of three you mentioned. The lower launch costs would allow lower cost missions in three ways (that I can think of right now.) First is less emphasis on reliability, which you mentioned. But I once discussed something similar with a colleague. I said we’d be better off with a 50% failure rate, if it would cut the cost of scientific spacecraft by a factor of ten. He agreed that might be possible, but wouldn’t want to take it past 10% and a factor of two. Specifically because, at that point, the overall cost would be dominated by the launch.
The second and third reasons are specialized missions and frequent reflight. For planetary missions, at least, a huge amount of time and effort goes in to making the spacecraft as capable and versatile as possible. Much of the motivation for that is the low number of opportunities. If opportunities happen once a decade, or less, there is tremendous pressure to make it do everything, perfectly. Lower launch costs would mitigate that. It would also permit frequent reflight, which would be based on the discoveries of previous ones. If low launch costs get us away from the idea of rare, highly optimized and multipurpose missions, I could see ten times as many, more specialized and less optimized missions, under a fixed budget.
Finally, we should not forget the rest of the world. The US and NASA are not the only ones interested in space science. There is a world full of countries interested in a scientific space program, if only they could afford the launch costs.
Hey Tania, good post. But you probably won’t be shocked to find I don’t entirely agree.
First, I think the market for GEO comsat launch services is a bit more elastic than you think. If the big GEO operators were honestly as indifferent to launch costs as you maintain, it would be hard to see why SES – the biggest in the business – has moved so much of its launch business to SpaceX and why it so aggressively pursues SpaceX’s discounted used boosters to do that business with.
One market for more discounted launches is very likely to be crew rotation to one or more commercial space stations in LEO and other cis-lunar locations. Bob Bigelow and Tory Bruno just announced a mostly detail-free “partnership” to put a twin-B330 space station into LEO by 2020. Bigelow already has memos of understanding with at least a half-dozen foreign space agencies. I’m guessing he won’t have any trouble fully leasing this first station by the time it flies.
Bigelow long ago indicated that his preferred crew changeout interval is 90 days, though he would be okay with 120. A twin-B330 station holds 12 crew. Swapping them out three times a year would require 6 missions, in some combination, of Falcon 9-Dragon 2 and/or Atlas V/Vulcan-Starliner. Figure that for every B330 anywhere in cis-lunar space, there would need to be 3 or 4 crew rotation missions per year. That would quickly dwarf NASA crew rotation demand for ISS and would leave its loss, when ISS is eventually de-orbited, barely a blip on the manifests.
Oh yeah, the same thing applies to strictly freight-hauling resupply missions too. The multiplication factor might not be quite as large vis-s-vis current and projected NASA demand for ISS as would be the case for crew rotation – especially if Bigelow figures out how to put a laundry room on the B330 – but it would still be a very significant bump in mission rate.
That would definitely save some money on launch costs. And if launches are cheap enough, then it becomes more practical to do in-orbit assembly and refueling of spacecraft.
Couple that with reuse of the Dragon V2 capsule and add to that a “full scale” Bigelow Aerospace LEO station and we could be witnessing the beginnings of a commercial space station program open to more than just government astronauts.
SLS feels more and more like yesterday’s news.
Between Falcon Heavy and the MCT architecture SLS is looking like it will be the 21st century equivalent of Howard Hughes’ Spruce Goose…
I found the presser very interesting. Elon has cleared up a lot of the uncertainty about F9 reusability with his answers and I think we’re getting a significant insight into his end game.
keith At what $$ would you think a core would for for/
..say on the 97th or 98th launch?
You amortize the rocket’s cost over the life of the rocket. If the rocket first stage cost 16 million and you fly it ten times your cost of the first stage is 1.6 million per flight…
Broadly correct although you’d also need to add about 50% to the construction cost for refurbishment and flight ops over the lifetime of the vehicle. Even if I’m underestimating that by a factor of two or three, we’re still looking at a core that only adds $5M to the launch cost per mission. No conventional ELV would be able to match that.
It could be either more or less than the cost you estimate. I suspect NASA would insist on a new first stage for ISS missions (more likely for crew transfer than cargo.) I also suspect they would not object it that cost more. On the other hand, if the rated lifetime is 20 launches, and that stage is on number 19, I suspect most customers (and their insurance companies) would need some encouragement. I so I could see a higher price for a new one and a lower price for a very used one.
Rockets only operate for 15 minutes or so. 15 minutes! And although I’ve pointed out that I am not a rocket scientist, I wonder if a rocket is more complex than, say, a 787?
It’s true that the short time rockets are alive is a bone-jarring and very high energy 15 minutes, but still.
The notion that rockets will somehow ‘wear out’ rapidly is, if you will forgive me, Old Think. Yes, airplanes do wear out; but rockets could be subject to the same rigid procedures governing modern airliners, giving them for all practical purposes an unlimited lifetime.
At some point they will simply be called ‘boosters’ and will be easily available.
I’d like to note that there is nothing fundamental about a liquid fueled rocket engine that makes it “single use”. In fact, the Merlin engines on a “new” stage are test fired more than once before a mission (once in Texas and once on the pad). So, the engines are already being “reused” on each and every flight.
Traditional rocket engineering is that if the rocket is not “worn out” after those 15 minutes, before you drown it in the atlantic, than it is overbuilt. That is, as you say, “Old Think”. Now we are in a new paradigm of “New Think” how do I have to engineer an engine to use it again and again, the tankage, the avionics.. Now we are moving to the over build it rather than under build it.
I’ve read several commentaries that question Economics of rocket re usability. There critique rests on two arguments.
1. The refurbishment costs of a used rocket are nearly as much as a new rocket. This seems to be based on the Space Shuttles main engines which had to be nearly rebuilt each time.
2. You have to use a much bigger rocket to put a payload into orbit to account for the fuel to return to earth. Why not spend that fuel on bigger payloads?
I think the answer to these critiques is the same. The Falcon 9 is much more rocket than you usually need for the payloads that it is putting into orbit. Traditionally, if you wanted to put 1000 pounds into orbit, you put it on a rocket that can put 1001 pounds into orbit. That makes sense if you have to throw the rocket away afterward. A 1001 lb payload rocket is supposedly cheaper than a 2000 lb payload rocket. But, a 2000 lb payload rocket is cheaper is you can use it 20 times and you only pay for the fuel. Look at all the variations of the Atlas series of rockets.
That fact that the Falcon 9 is over built compared to other rockets is also the answer to the refurbishment problem. Running a rocket at 99% of its capacity for 10 minutes is going to wear it out. If it doesn’t wear it out than you should make the components lighter which gives you increased payload and reduced costs which is what you want to do if you are going to throw the rocket away. When the goal is to reuse the rocket, the formula changes. Yes, you are giving up payload but as long as there are payloads out there that you can lift you have a viable product.
The “Old Space” people got stuck in circular thinking.
1. You have to use the smallest rocket you can because you can’t recover it.
2. You can’t recover your rocket because you would have to build it too big for the payload and that would violate the first rule.
Plus the critics may be assuming that the stage must be “refurbished” after each flight (similar to what the space shuttle orbiters or SRBs needed to have done). From Musk’s tweets above, it’s clear that he’s hoping to re-fly an individual stage 10 times before “refurbishing” it. That is resuability.
There is also a third critique to the reusability question, and that is the one of market elasticity. Various articles through
the years have argued that you need approximately 50 flights a year for reusability to pay for itself. Elon’s statements on how many times he hopes to reuse a Falcon and SpaceX’s stated flight rate certainly point toward a number similar to that. BUT – will there be 50 payloads a year? How elastic is the market? How far do prices have to drop for demand to significantly increase?
If you recall, 20 years ago we were going to have massive constellations like Globalstar and Teledesic which in turn would result in amazingly low cost vehicles derived from Pegasus, Amroc, and Conestoga. The launchers would be cheap (and ultimately reusable) because there would be such high demand — but the demand would be so high because launches would be so cheap. This was the proverbial
chicken and egg of the last wave of New Space in the 90s. Elon has jump started that process by injecting a lot of his own money. However, is that enough to really bring the demand up to the levels that make reusability sustainable?
The GEO market is quite inelastic to launch costs. About 2% of DirectTV’s operating costs are satellite launches. Intelsat’s launch costs may be a touch higher, but not significantly so. We cansafely say a decrease in launch costs of 100% leads to essentially 0% growth in demand. Giving away launches for free would not really impact their business plan.
NASA science missions are a bit more sensitive to launch costs. About 10-20% of their costs are in the launch, so significant reduction in launch costs makes a difference. However, NASA’s budget is fixed, so even if you give away launches, you only get one new mission for every five missions. You could argue that perhaps total spacecraft costs also come down as a result of willing to accept more risk in order to significantly lower price, but I still can’t see much more than one new mission for every three existing missions or so. So a decrease in launch costs of 100% leads to about 30% increase in demand. Not nearly enough to get you to the volumes where reusability is economically feasible.
For the large remote sensing systems, the numbers are
probably similar to the Geo comm market. For smaller constellations such as PlanetLabs and Skybox, the elasticity does start to go up significantly. However, their satellites are a lot smaller and the Falcon 9 is significantly oversized. If it is cheap enough, that may not make a difference, but I’m not sure what “cheap enough ” is.
Maybe the elasticity is in CubeSats! Spaceflight just sold an entire Falcon 9 to CubeSats and other small nanosats.
However, how often do 80+ satellites need to go to the same orbit? And how often can you even just identify 80+ satellites needing to get into space? I’d venture to say that twice a year is the most we can expect from massive aggregated launches.
In the end I think we are left with just two areas where the market might be elastic enough to feed the demand required to make reusability economically feasible:
1) Large Communications constellations (OneWeb, Google, Facebook). However, reference the 1990s. And how many constellations can we realistically have before
competition, regulatory concerns, debris mitigation, and frequency availability become an issue
2) Very large space logistic efforts: moon base, Martian colony, asteroid mining. However, even with launch vehicle reusability these projects are so capital intensive that they are likely to require much more than just Elon’s riches.
human cargo to LEO and the support cargo to keep them alive.
A bit negative, but still; truthfully we’ve not really identified the WHY of space travel very well. I can’t verify the numbers: DirectTV’s costs at 2% for launch seems mighty low, but then again, the savings are still in the tens of millions.
But I do think that many grandiose space plans are based on sandy foundations- and I’m looking at you, Mission To Mars. Same goes for the Lunites, although the case for a research base on the moon is easy to make. Certainly a lot easier to make than a mission to Mars.
And this: the chicken and the egg are still in charge here. Real money will be made when we learn how to actually live in space; how to use the rich resources available in asteroids, how to mine, smelt, and manufacture; how to construct space habs; how to live in space without regular departures to deep gravity wells.
I agree that elasticity of the market is a big deal, but I’m not sure I agree about the details.
I’m surprised by the 2% number for DirecTV’s spending on launches. It seems to be either too low or too high. I don’t know about their business model, but how deeply are they involved in owning and operating satellites? I thought a common practice was for service providers to rent transponders on a satellite launched and operated by someone else. (I need to learn more about this, since I’ve got an idea for an instrument which would fit well as a hosted payload on an SSL communication satellite, and I’d like to know how they can offer that service if they don’t retain ownership of the satellite.)
In any case, the same point you mentioned for scientific spacecraft applies to commercial ones. If the launch costs go down, the need for long lifetimes and high reliability also goes down, and the cost of building commercial satellites with it. If replacement of assets is economical the demand for launches will increase.
As far as NASA’s scientific missions go, my guess would be a factor of five to ten, rather than the factor of three you mentioned. The lower launch costs would allow lower cost missions in three ways (that I can think of right now.) First is less emphasis on reliability, which you mentioned. But I once discussed something similar with a colleague. I said we’d be better off with a 50% failure rate, if it would cut the cost of scientific spacecraft by a factor of ten. He agreed that might be possible, but wouldn’t want to take it past 10% and a factor of two. Specifically because, at that point, the overall cost would be dominated by the launch.
The second and third reasons are specialized missions and frequent reflight. For planetary missions, at least, a huge amount of time and effort goes in to making the spacecraft as capable and versatile as possible. Much of the motivation for that is the low number of opportunities. If opportunities happen once a decade, or less, there is tremendous pressure to make it do everything, perfectly. Lower launch costs would mitigate that. It would also permit frequent reflight, which would be based on the discoveries of previous ones. If low launch costs get us away from the idea of rare, highly optimized and multipurpose missions, I could see ten times as many, more specialized and less optimized missions, under a fixed budget.
Finally, we should not forget the rest of the world. The US and NASA are not the only ones interested in space science. There is a world full of countries interested in a scientific space program, if only they could afford the launch costs.
Two thumbs and both big toes waaaay up!
Hey Tania, good post. But you probably won’t be shocked to find I don’t entirely agree.
First, I think the market for GEO comsat launch services is a bit more elastic than you think. If the big GEO operators were honestly as indifferent to launch costs as you maintain, it would be hard to see why SES – the biggest in the business – has moved so much of its launch business to SpaceX and why it so aggressively pursues SpaceX’s discounted used boosters to do that business with.
One market for more discounted launches is very likely to be crew rotation to one or more commercial space stations in LEO and other cis-lunar locations. Bob Bigelow and Tory Bruno just announced a mostly detail-free “partnership” to put a twin-B330 space station into LEO by 2020. Bigelow already has memos of understanding with at least a half-dozen foreign space agencies. I’m guessing he won’t have any trouble fully leasing this first station by the time it flies.
Bigelow long ago indicated that his preferred crew changeout interval is 90 days, though he would be okay with 120. A twin-B330 station holds 12 crew. Swapping them out three times a year would require 6 missions, in some combination, of Falcon 9-Dragon 2 and/or Atlas V/Vulcan-Starliner. Figure that for every B330 anywhere in cis-lunar space, there would need to be 3 or 4 crew rotation missions per year. That would quickly dwarf NASA crew rotation demand for ISS and would leave its loss, when ISS is eventually de-orbited, barely a blip on the manifests.
Oh yeah, the same thing applies to strictly freight-hauling resupply missions too. The multiplication factor might not be quite as large vis-s-vis current and projected NASA demand for ISS as would be the case for crew rotation – especially if Bigelow figures out how to put a laundry room on the B330 – but it would still be a very significant bump in mission rate.
The 10-20 times estimate is probably based on real test results. The 100 times sounds a bit arbitrary.
Gwen Shotwell stated that engines have been static test fired 20 full burns before.
That is Ms. Shotwell warming up her audience.