Seven Years of the SpaceX Super Heavy Starship Future With Hundreds in Space and the Moon

66 thoughts on “Seven Years of the SpaceX Super Heavy Starship Future With Hundreds in Space and the Moon”

1. OK, I found an ars technica reference

   William Forchestein (spelling?) talked about this in one of his books. I forget the title. He called it fire jumping and the participants were given a piece of burnt toast to eat after their first jump. They were doing it from a 65 mile platform on a beanstalk.

2. I remember there’s some sort of trope/running gag joke that orbital skydivers eat burnt toast on landing.

3. A basic slinger could be a single lander payload, though you would want to bulldoze some protective berms around flinging radius, along with the counterweight catch pit…

4. I don’t think that that’s a huge problem. When you’re launching from Earth, launch costs dominate, and go straight into capital costs. Because launch costs decline slowly or not all (there are only so many Starship-like quantum leaps you can perform), you never get to amortize the launch costs if you’re deploying at high scale.

   For however powerpointy the R&D and deployment of lunar mining / manufacturing stuff is, at the scale we’re talking about, it’s almost a rounding error on the overnight costs of the actual SPS, because once it’s developed, launched, and deployed, the launch and maintenance costs go almost to zero, which means that you’re amortizing the costs over a massive amount of SPS capacity produced via ISRU.

5. That seems more like version 2. Any kind of mass driver requires actual construction. Version 1 seems like it’s vehicles collecting stuff and maybe transferring it to other vehicles for processing.

6. I worry we are falling for an old trap.

   Comparing real, actual, currently on the market launch costs and limitation with theoretical, projected, white-board and powerpoint level ISRU.

   Under those circumstances, the powerpoint stuff always looks much better.

   Just like that quote that Scaryspice keeps referring to about real reactors vs paper reactors.

   http://ecolo.org/documents/documents_in_english/Rickover.pdf

7. I’m not getting the burnt toast reference?

8. I don’t entirely disagree, but there’s a middle ground that doesn’t require us to wait for full-fledged lunar mining and satellite factories: launch components only to LEO, and get it to GEO with reusable space tugs fueled from the moon. At $10M/launch and 150 tonnes capacity, that’s $67/kg to LEO.

   The full-scale SPS-ALPHA design would deliver two gigawatts, and mass 34,814 metric tons, for a cost to LEO of $2.3B, or $1.2B per gigawatt.

   Orbital transfer would be required for lunar-manufactured sats as well, unless you intend to entirely use lunar mass drivers. But you’d likely need a more robust and heavier design if you did that. So essentially you’re looking at an extra $1.2B/GW to launch from Earth, which could be offset by lower manufacturing costs vs. a lunar factory.

   Here’s the NASA report on SPS-ALPHA:  https://www.nasa.gov/pdf/716070main_Mankins_2011_PhI_SPS_Alpha.pdf

   The mass estimate I used is from page 73.

9. I so desperately want to see some extreme sports adrenaline junkie doing this at some point. With the requiste eating of burnt toast on landing to complete the ritual.

10. Using a lunar surface sling and bagged water (with thin film solar lining the bags to provide power to keep it liquid) seems like the easier way to get the water up without using propellant/tankers.

11. I built a marginal cost model for this a while back. It’s 2-4x cheaper from the Moon, even with the operational unpleasantness of maintain mining and transport equipment with a small human crew. However, that didn’t include the initial capitalization, which is hefty.

    Short answer is that lunar water becomes a no-brainer at even moderate scale, but getting to that moderate scale is a dicier proposition. That’s why a government-sponsored base would be very helpful. Even if it charges a premium for power and hab facilities, it lowers initial capital costs and reduces risk enough to attract private investors.

12. What you’re saying is true for intra-exchange trading, but arbitrage is where the HFT money really is, and you have to see the trades on the distant exchanges, as quickly as possible, to be successful there.

    That said, I’m not sure how durable a market low-latency financial bandwidth is. When it first becomes available, there are first-mover advantages to having your own feed of the transactions, but it rapidly becomes table stakes for everybody, at which time some bright boy will discover that he can bring one feed into the exchange and sell it to hundreds of firms for a tenth what SpaceX would charge them. Since he only has to buy one SpaceX feed, his profit margin is huge.

    I’m guessing that finance is a non-trivial chunk of SpaceX’s early market, but I’ll bet the largest customer will be deploying 5G backhaul for places that have neither fiber infrastructure nor convenient microwave paths to a suitable terrestrial POP. Beyond that, you’ve got the early-adopter residential customers, planes-trains-n-automobiles, IoT instrumentation, and defense.

13. I chose that number because it seemed big enough to impress the rainbows-and-unicorns crowd that SBSP with no storage would scale better than piecemeal terrestrial renewables, which probably need something like 10-15 Wh/W of storage backing them in an all-renewable grid.

    See my second back-of-napkin to Dennis above. Even at 500 MW of nameplate, I get something quite a bit larger than nuke overnight $/kW if everything is launched from Earth.

    The crux of my argument to Dennis is that, even at SH/SS specific costs, SBSP is pretty iffy without lunar ISRU, manufacturing, and mass drivers. But those specific launch costs are low enough to enable the lunar surface stuff, which then returns you hundreds of times the mass to orbit per landed kilogram.

14. I’m more interested in how much mass you have to put in orbit to get a certain amount of power into the grid on the ground. So, for example, if you can do 175 W/kg, how much of that is lost via beam conversion losses, diffraction, atmospheric absorption, and rectenna losses?

    This isn’t super-important for full-up ISRU and lunar manufacturing, but it is if you plan to start out launching stuff from Earth.

15. There was a TV show about an alien from the moon who would fly his spaceship down to Earth and draw pictures.

    Fiction is not a good source of data.

    (Mr Squiggle. The world’s longest running science fiction TV show. https://en.wikipedia.org/wiki/Mr._Squiggle)

16. I disagree with your minimum size requirements.

    100 GW is needed to be significant on a world scale, but that’s not the scale that the decisions will be made on.

    You can run a small-medium industrial country, eg. Austria, on 8 GW. Now Austria is already 75% hydropower, so 2 GW will make the entire country “renewable”. So a 2 GW SSP is a significant, politically powerful milestone for Austria.

    I just chose Austria because it was the first place of the appropriate size that I could find the numbers for.

17. Wouldn’t a wrapping of a couple of lightweight sheets of reflective foil keep a decent sized chunk of ice from heating up? At least long enough to get it to Earth Orbit and make use of it?

    Actually getting it from the Asteroid belt to Earth orbit is left as an exercise for the reader.

18. Just say MOOSE

    https://en.wikipedia.org/wiki/MOOSE

19. To support BF rockets you are going to need BF payloads and BF money.

20. I do believe that flat-earthers are almost religious in their zeal, pseudo-scientific at best, just insane conspiracy theory at worst. However, I have worked in both scientific and religious institutes and have seen plenty of wacky conspiracy theories and even insane science theories in both.
    Part of the general idea that religion is associated with flat Earth is probably due to the anti-Catholic sentiment in England that always tried to paint Catholics as ignorant until the Renaissance and the age of discovery saved us. However, the idea of the earth being spherical was widely held in Catholic Middle Ages:

    https://en.wikipedia.org/wiki/Flat_Earth#Europe:_Late_Middle_Ages

    Hermannus Contractus (1013–1054) was among the earliest Christian scholars to estimate the circumference of Earth with Eratosthenes’ method. St. Thomas Aquinas (1225–1274), the most widely taught theologian of the Middle Ages, believed in a spherical Earth; and he even took for granted his readers also knew the Earth is round.[121] Lectures in the medieval universities commonly advanced evidence in favor of the idea that the Earth was a sphere.[122]

    Just to be clear.

21. “Embodied energy” is all the energy required to make a product, from mining to delivery and installation. It varies by product, but a typical number is 20 MJ/kg. A modern space solar panel can produce 175 W/kg, so it can theoretically produce the power to make its own mass in products in 1.3 days. They typically last 15 years, so they can produce 4000 times their mass.

    Energy isn’t the limiting factor in space industry. It will be other things, like how long a given manufacturing step takes.

22. Yes, I think you are the one. And I agree some capacity will be needed on the moon.
    BTW, I like your user name.

23. The fly in your argumentative ointment here is that there’s a huge amount of path dependence built into getting to almost all of those industries. Modern finance being what it is, private investors get squirrelly when they have to risk multiple billions at >30% risk of failure and time horizons over ten years, unless they’re getting projected IRRs in the 20% range.

    Those IRRs make the actual products/services expensive, which in turn limits demand.

    That’s why government infrastructure is so important. If private enterprise can buy services like power and propellant from government-initiated public/private facilities on the Moon and in cis-lunar, and piggyback their projects on still other government public/private partnerships, it defrays risk and lowers the initial capitalization. Suddenly, investors are only demanding IRRs of 10%, and the lowered costs of services makes demand explode.

24. I’m gonna do another back-of-napkin here for a single 500 MW capacity SBSP plant, using only terrestrial equipment:

    If we’re doing a pilot plant, it needs to be soon, so let’s start with about 150 W/kg specific power for just the PV, structure, wiring, and controllers. Double the mass for beam forming, radiators, and spacecraft control: 75 W/kg. Assume 30% diffraction losses: 53 W/kg. Atmospheric losses are heavily dependent on wavelength and rectenna design (I’m going to assume that the aerostat-based stratospheric rectennas are off the table), but I’d be surprised if you got away with less than another 25% loss factor: 39 W/kg.

    500 MW is 12,800 tonnes to GEO.

    I get $200/kg to GEO with SH/SS, assuming that you need 3 launches @ $10M for each 150 tonnes of payload, so launch costs are $2.6B for about 260 launches. Figure that the actual modules will cost $200/kg, and that’s another $2.6B. Throw in $1B for the rectenna and grid connect, and finally $1B in R&D. Total capital cost: $7.2B.

    That’s an overnight cost of $14,400/kW. Nukes in the US are running about $4500/kW, so this isn’t great. I’m willing to believe that I’ve been a factor of 3 too pessimistic above, but you’re still looking at nuke-like capitalization, likely with somewhat lower ops costs, but I’d be surprised if you had an LCOE below $80/MWh.

    You might get away with that as a pilot to prove out tech you planned to manufacture on the Moon, but ISRU needs to be part of the plan from the git-go.

25. That was exactly my point.

    Perhaps I misunderstood Dennis’s comment, but it seemed like he was arguing that we could do viable SBSP with direct launch of all components from Earth. Even at purported SH/SS prices, that simply doesn’t work.

    I’m starting to get excited about lunar ISRU-based SBSP as a renewable energy source that can be deployed quicker than any terrestrial renewable solution. There is obviously a lot of R&D to do and hefty capitalization costs to getting things going, but I suspect that the exponential time constant for build-out is quite small once you starting flinging stuff off the Moon.

    If we’re even slightly serious about de-carbonizing by 2040 or so, SBSP has to be one of the leading contenders.

    Question for you: What did you wind up with as your specific power (W/kg) for your satellite, in terms of power delivered to the terrestrial grid?

26. All they need is 10 million internet customers at $70/month. You can find that many in rural US, which has sucky bandwith today. In less developed parts of the world, a 1 Gbps satellite link mounted on a cell tower and split 100 ways can feed a village worth of cellphones with decent enough broadband at low cost. That’s millions more customers.

27. Maybe you aren’t aware of this, but space industry is worth $350 billion a year. Most of it is commercial stuff like communications satellites.

28. I’m not too worried about the beliefs of religious nuts. They believe all kinds of crazy things. But that hasn’t stopped the engineers and scientists of the world from building things that work.

29. 98-99% of a solar power satellite could be built with materials from the Moon and asteroids. I know that because I was part of a study that worked out the numbers. In turn, a large percentage of your space factory to build the satellites can also be built from local materials.

    This reduces the launch problem to the bootstrap factory needed to build out the rest, plus the percentage of the remaining factory and powersats that still come from Earth.

30. Space industry worldwide is worth $350B per year. What has limited its growth is the high cost of launch and lack of off-planet industry. NASA is a small fraction of that total. Most of it is communications & earth observation.

    Off-planet industry would include mining, manufacturing, tourism, science, etc. Today, nearly everything is pre-built on Earth, but raw materials and energy are abundant in space. Once we start using them, new activities become viable because we don’t have to launch everything from Earth.

    The first products everyone expects are propellants and radiation shielding. Most anything you do in space uses propellants, and shielding is needed for people and equipment beyond low orbit. Satellite repair and refueling are likely the next businesses.

31. Anywhere closer than the middle of the Asteroid Belt (i.e orbit of Ceres) is too close to the Sun to retain water ice as ice unassisted. A gravity well (Earth and Mars) or a shaded crater (Moon and Mercury) can provide that assistance.

    What we do have in the inner Solar System is “hydrated minerals”, where water is chemically bound. The water of hydration can be extracted by heating to 200-300C (kitchen oven temperatures) via solar concentrators. Carbonaceous type asteroids also contain carbon compounds, which can also be extracted at these temperatures.

    Carbon + water yields methane + oxygen, which can fuel Starship. If you do the processing in open space, away from planet shadow/night, the sunlight is 24 hours to power things.

32. I agree it makes a lot of sense to manufacture them on the Moon. Not only would it be cheaper, but it’d avoid a significant amount of CO2 emitted from rocket launches. But it’s not likely to happen before a proof of concept.

    The excitement for a bit carbon impact can come from knowing it’s just the first step. But even that’s not strictly necessary. At 4 cents/kWh with no need for storage, people can get excited about making money.

33. And getting aboard them wold be like trying to board a plane standing on its nose, via a door in the tail…

    This needs plain ballistic capsules as lifeboats. Screw cross range and horizontal landings, just get me down.

34. Please direct your imagination at more productive targets.

    Seriously. There’s a huge social cost to ridiculing idiots. You can help make the world a better place by not contributing to it. Beyond that, engaging in ridicule diminishes your soul.

    As my mom used to say, just ignore ’em and they’ll go away.

35. I haven’t read the book, but that won’t prevent me from doing a quick back-of-napkin:

    1) The state of the art specific power is about 100 W/kg, and that’s with no beam forming. Figure that with a quadrupling in raw specific power, plus the microwave generation and radiators, plus diffraction and atmospheric losses, we’d be at maybe 75 W/kg in 10 years.

    2) I can’t see a viable system for anything under 100 GW of delivered capacity, if for no other reason than any capacity below that doesn’t put a big enough dent in carbon emissions to get people excited.

    3) Figure that the build-out time of a major project has to be under 10 years to attract investment at any ROI. Like the 100 GW threshold, this has a lot to do with investor excitement being balanced by risk tolerance.

    So we’re looking at 1.333 million tonnes of material launched. With a Starship capable of getting 150 t to LEO, you need 2 tanker launches per payload launch to get your payload to GEO, so we’re looking at 26,660 SH/SS launches. If you do that in 5 years, that’s 7 launches a day.

    Not happening.

    On the other hand, 50,000 tonnes of ISRU and manufacturing equipment landed on the Moon seems like it would handily provide the power, mining, metal reduction, PV, electronics, and structural manufacturing, along with a mass driver to launch the finished product. It takes about 8 SH/SS launches per 150 payload to the lunar surface, assuming that there’s LOX there to refuel. That’s 2670 total launches.

36. As usual, if the government spends on infrastructure, the infrastructure lowers the capital cost of spinning up real commercial businesses. Then you’ll see private companies start to offload the government investment with cheaper alternatives and, more exciting, start to do things with the infrastructure that nobody thought of.

37. I think I might have been that commenter. However, you do need some hydrolox capability on the lunar surface, because you have to launch the water. I looked at whether it might be cheaper to land hydrolox from orbit to fuel the water tankers, and it’s not even close.

    So the proper sequence is:

    1) Land a big power system.
    2) Land your ISRU system and start making hydrolox.
    3) Scale surface hydrolox up enough to fuel water tankers.
    4) Build a big water storage and just-in-time hydrolox production facility in orbit (or orbits).
    5) Use the stuff you built in the previous steps to do commercially viable propellant sales in orbit.

38. Find ways to make money by going to space and it will happen by magic.

39. The solution to that is a well trodden path, which is actively being pursued not by NASA alone but by a dozen space agencies world wide. They have agreements, and have committed funds, to create and foster this CISLUNAR market as a commercial market by buying services. In essence, there is now an attractive market with buyers who pay (more than) premium prices to create that necessary volume. Once R&D costs are recouped through this scheme, cost drops dramatically and predictably. Case in point, there are literally hundreds of markets that were created by government taking initiative or subisidizing the market by buying services and products (ocean travel, air travel by promoting the most expensive mail possible (airmail), mining in remote areas, Alaskan villages, etc.) ..until the market could support itself. Now, in the case of Lunar Mining, the only problem is the cost of getting off of Earth cheaply. That problem has been solved by SpaceX with the Falcon 9 and is looking rosier still when the BFR and New Glenn comes online. With their reusable capability, refueling in orbit, landing on the Moon, roving, staying and returning isn’t difficult ( the ill-fated Indian and Israëli effort were only first tries by new players who have already committed to try again. in contrast with SpaceX, they have not yet had the ability to test in a relevant environment). The market is already there and is already buying goods and services, some haven’t just caught up with company news.

40. Yes, but why would you buy moon water products when you can buy cheaper earth water products.

    Companies who develop expensive technology/capabilities -while hoping volume will eventually make it cheaper- usually die from a lack of cash flow.

    For private financing, nothing can exists unless there is a large and easy profit pathway. Most people consider all public financing illegitimate unless it’s military related.
    NASA will be the only customer if it doesn’t have a viable business model.

41. “doesn’t that make launching water into space cheaper($) than sourcing it from the moon”
    No.

42. There was a movie about mining in the moon and the moon fractured. The movie? Time machine…

43. Brian – to clarify your link on latency and trading systems. Rules have changed. It isn’t “algo trading” as much as it’s high-frequency trading (10’s of thousands of trades/sec). In the good old days microwaves could handle ultra-low latency. Nowadays if you want to do HFT you need to reside INSIDE the exchange. In the server hall. HFT was abusive (and still is to some extent) where traders could “front run” the market (see what orders came in before anyone else did and make the trade accordingly).

    No one will buy trading latency “capacity” in space. And I am not sure how realistic the rest of Starlink’s business model is.

44. I really appreciate from SX the need to establish first a relation with te Moon and his resources (similar in this to the Bezos’s approach) and AFTER, with all the experience gained on the Moon, make focus on Mars..

45. While it’s beaming sunlight down it might be increasing the total energy received by the Earth, but depending on the orbits it will probably spend more time shielding the Earth from sunlight than increasing the load.

    The problem with this idea is that it only gives your solar farm a couple more hours per day. It still doesn’t help during the night.

    https://www.nextbigfuture.com/2018/10/chengdu-launching-satellite-to-create-8x-full-moon-lighting.html

46. If they need dozens of lifeboats there’s probably a better choice than Dream Chaser. That seems rather like old fashioned NASA style spread the pork politics. Using existing designs isn’t necessarily cheaper as SLS proved beyond the possibility of doubt.

47. But would that not heat the earth, since the efficiency of the solar panels at earth is of the order of 20-30%? So what is best, burn carbon and get energy and AGW, or use space reflected solar and get AGW?

48. Yes, it is cheaper to launch H2O and LOX from Earth in the next decade to come. But the point is to develop the technology /capability on the moon and eventually, if enough activity arises in space creating a demand, it will become cheaper.

49. ISRU == $$$

50. Please see proposals of the late Paul Spudis. Very insightful, IMHO.

51. Easy/Hard is irrelevant, $$ is all that matters.

52. Like it! Then use Criswell redirectors to send it around for balance. Only build big stuff in Space when scale requires.

53. Huh? Launch form Moon easy, Earth hard!

54. O’Neill sez space easier than Earth, even tho we are already here. Way easier than Moon! All moist lunar dirt can be used in Space. Just send it there directly.

55. http://www.searchanddiscovery.com/pdfz/documents/2009/70070criswell/ndx_criswell.pdf.html
    Enjoy the ride!

56. Dr. Lewis Fraas has a Space Mirror Concept that beams extra sunlight down to terrestrial Solar Farms around dusk and dawn. It’s 100 times cheaper than traditional microwave transmitted SPS. Most earth based Solar Plants are based in places that have no cloud cover 95% of the time. But for areas above 45 degrees latitude, it is good to hear that the low cost of SpaceX Super Heavy 2.0 will make even microwave beaming SPS economically viable.

57. I would really like to start seeing actual ISRU hardware for the moon.
    The mining equipment. Maybe bakers to get off volatiles from the regolith. Processors. Stereolithography or 3D printers.
    You could use the iron and carbon to make steel structures or just get the Aluminum out.

58. Some people just can’t be helped.

59. Given 100% reusable LH2/LOX rockets, doesn’t that make launching water into space cheaper($) than sourcing it from the moon for a launch provider?

60. My guess at order:
    NASA public-private programs
    International Lunar Exploration Phase (ILEP) – about 4x NASA.
    LEO to GEO boosting.
    Private lunar settlers with revenue from their savings.

    There would be a number of smaller sources of revenue. (e.g. tourism, cislunar propellant sales, PGMs, lunar jewelry – don’t laugh).

    Private savings is not a small business model yet is often overlooked in favor of space resources. But this is the heart of Elon’s business case (intersection of those who can go with those who want to go).

    Space resources will primarily play the role of reducing transportation and habitation costs rather than direct sales to others.

61. Can you just imagine Flattards stuffing their chests to say that astronauts being able to walk on the Von Braun station is proof it’s all a lie and that they are in a Hollywood studio??

    They already do it with astronauts floating 100% of the time in the ISS!

62. Lots of ice is in near space, may be better to direct it to a depot in a near point in space and process it there.

63. At BFR prices, SPS is viable without asteroid mining.

    Old SPS designs from the ’70s were expensive monolithic beasts. Current designs use lots of small identical parts that self-assemble in orbit. Manufacturing cost is much lower that way.

    The book *The Case for Space Solar Power* does a detailed cost breakdown. It estimated a total cost of $0.15/kWh with pre-SpaceX launch prices. I plugged in BFR’s $50/lb (which it should get to pretty quickly with a customer of this scale) and got $0.04/kWh, which is pretty good for low-carbon baseload.

    At that point the bulk of the cost is manufacturing, but we could drop the cost a bit further with a LEO-geosych tug fueled from the moon or asteroids.

    (Edit: I’m aware that Musk scoffs at SPS, famously asking “what’s the conversion rate?” The answer was 40% when that book was written, which isn’t bad considering a panel collects 5.4X more sunlight in 24 hours than one on the ground, and you barely need storage.)

64. As one commentor here mentioned some time back, you don’t need much electrolysis equipment on the Moon. Just get the water off the surface and to the fuel depot and then do the electrolysis in space where it is cheaper to bring equipment including solar panels. Water is also easier to handle than liquid hydrogen and oxygen. Store it as water and do the electrolysis as needed.

65. Asteroid mining could start being viable at these sorts of cost to orbit. That in turn could make SPS viable.

66. Are there any viable beyond LEO business models floating around or will everyone just be competing for limited NASA bucks?

Comments are closed.