The Case for Space Stations

Part of my series on countering common misconceptions in space journalism.

Previously I’ve written about how practical space station size is limited due to non-obvious problems with modularity, and the economic challenges of O’Neill cylinders.

However a recent conversation prompted me to re-examine my assumptions and so, in this blog, I will be making the case for space stations.

The Day Skylab Crashed to Earth: Facts About the First U.S. Space Station's  Re-Entry - HISTORY

First, the challenges. The ISS took the better part of 30 years and $300b to build, including related programmatic costs for the Shuttle. For 20 of those years, the station has been occupied by between 3 and 6 astronauts, who have spent the vast bulk of their productive time building and maintaining the station. None of this is time wasted, as one of the primary aims of the program was developing operational practices for long duration station occupation, however it must be acknowledged that, to date, the ISS has permitted at most 10 person-years of unrelated scientific research, which works out to be about 20,000 hours.

In contrast, consider the South Pole Station, which has been permanently staffed since 1956 and supports around 100 scientists for 3 months of austral summer, and about 12 for 9 month winters, together with a crew of 30-60 support staff. This works out as 34 person-years of research per year, for a program cost of about $20m/year. The total Antarctic program is about $200m/year, but this also includes the much larger McMurdo Station among others.

In summary, all in costs for ISS science work out to be $15m/hour, though if you consider only operational expenses that comes down to $2m/hour. Bargain! In contrast, Antarctic science operational costs are about $300/hour, of which a non-trivial fraction is the salary of the researcher themself.

Okay, space is expensive. No-one ever expected to go to space to save money. But here the plot thickens. NASA would like to bring commercial partners into the ISS, just as they have with launch and service. CASIS is the organization set up to do this, recasting the ISS as the ISS National Lab. The hope is that commercial interests could help share some of NASA’s operational costs, enabling NASA to open budget wedges for what comes next, without a loss in continuity. Of course the ISS won’t last forever and will eventually have to be splashed (in whole or part) somewhere in the Pacific Ocean, but in the meantime the steep costs of keeping it going conflict with the need to free up cash to work on new programs.

CASIS has done all kinds of interesting things but it’s fair to say that they haven’t been swamped by demand for ISS human lab bench space. The ISS was never intended to be affordable by mere billion dollar corporations, and so the research benefits of space are financially hard to justify.

Is there a future for private human space research? I would hope so. If Antarctica is anything to go off, can we hope for a vibrant community of tens or hundreds of researchers doing 3-18 month stints on space stations at costs that the NSF could fund without having to add 8 zeros to the ledger?

Things have changed since 1990 in all kinds of positive ways. NASA routinely resupplies the space station at one tenth the cost of the Shuttle, with multiple competing commercial launch providers positioned to provide similar services.

It sure looks like SpaceX is going to focus on the Moon and Mars, but it hasn’t been lost that the internal volume of a Starship is about twice that of the ISS, and that’s not even including convertible fuel tanks. I don’t see a strong reason for SpaceX to make a space station version of the Starship, but when I look at the New Glenn I think it’s got to be used for something!

Taking round numbers, New Glenn can launch 45 T to LEO for maybe $100m. Let’s say Blue Origin develops a generic space station product that integrates with New Glenn, a thousand cubic meter Skylab-like cylinder with docking adapters, solar power, radiators, life support, and all the other basics, also for $100m. This sort of budget precludes a 20 year multilateral international development period – instead the emphasis would be on getting the MVP up and running and then adding bells and whistles later, if and when customers come along who can pay for them.

If the station can support 10 researchers for 5 years, on 6 month rotations, operating and capital expenditure costs would be comparable to resupply and crew transport costs, perhaps as low as $10,000/person/day. While still more expensive than an ice-encrusted Antarctican astronomer, it’s within the same order of magnitude and significantly cheaper than, say, staffing a nuclear submarine.

At this cost, the current NASA budget alone could support a hundred such stations with money to spare. But the real benefit is that commercial operators can fly researchers and even tourists for yacht money, not stealth bomber money. There isn’t a commercial product in the universe that can justify $2b for 6 months of researcher time. $2b is comparable to the entire budget of JPL, which employs about 10,000 people. There’s just no way that microgravity can make a researcher 10,000 times as productive, no matter what the question might be.

$2m is a different story. It’s more money than I have, but the productivity demand is now only 10x, not 10,000x. This is more than plausible across hundreds of industries.

Is it physically possible to develop a line of no-frills space stations and sell them for $100m while not losing money (too quickly)? SpaceX developed the Falcon 9 for $300m. A Boeing 737 costs about $90m new. Both of these systems are substantially more complex, given that they have the capability to generally land non-destructively. Space station tech is mature and well understood. There is plenty of room for innovation, but no miracles are required.

Can we enjoy a future with thousands of people working in Earth-orbiting space stations? Yes, but only with a strong emphasis on cost reduction and value creation.

17 thoughts on “The Case for Space Stations

  1. A starship that is able to sustain a large crew for the trip to mars and back is already a space station. It needs power, cooling, independent and redundant long term life support.

    So you would just have to repurpose the cargo bay from mars settlement equipment to science experiments.

    Liked by 1 person

      1. SpaceX is building a Starship factory with the goal of being able to crank them out at a rate of one a week. Selling a couple of them for use as space stations would not prevent SpaceX from sending six of them to Mars.


      1. Really? I’m thinking any mods will be minor. The current starship can be lifted from the top with cables, seems like that’s all that would be needed to allow 2 starships to be connected nose to nose, rotating around the common center, and the tether could be long to minimize coriolis effects. Some health problems of zero gee are too major to ignore: serious eye problems, bone/muscle loss, and more.
        BTW: love your blog!


  2. An interesting proposal or vision, but I wonder whether you have heard of Axiom Space. They have a contract and investments to build the follow on to ISS, first attaching one module to ISS then adding more and separating to be independent when ISS deorbits. It will be about 3x the internal cubic volume and roughly a few percent of the ISS’s cost to build. I don’t know its operating costs/hr, but I assume they are reduced by a roughly equal proportion. You are clearly looking ahead to “thousands of people” working in space, but my hunch is that Axiom’s station will be a good stepping stone to that future.


  3. Like other commenters, I don’t see why using some Starships would be hard or expensive enough to justify developing a completely different vehicle. If the Martian city gets going, there will be a version of Starship being mass produced that’s designed for people to live and work in for months in free fall. A space station has people arriving and departing routinely, whereas a Mars City Transport version of Starship would not. But I can’t imagine an MCT fleet forgoing the benefit of the backup options that come with having it be possible to move people and materials from one Starship to another en route.

    On the other hand, I’m not optimistic about microgravity providing big benefits that will justify the costs of working in orbit. The ISS ok, because it lets there be an International Something Something: a big expensive project for countries to cooperate on, that evokes the moon shot.


  4. I seem to remember an idea floating around to use the shuttle external tank as the core of a space station. Guess that wasn’t practical either.


    1. My guess, and this is absolutely nothing more than a guess, is that reprocessing shuttle external tanks would have been practical if we’d had a reason for large-scale human presence in low orbit. I just don’t think that there is any compelling purpose for it. Mars and the moon have materials that could hypothetically be used to make aluminum-oxygen rockets, oxygen for use with H2 or hydrocarbons lifted from Earth, basalt-fiber lunar space elevators, and metallic components for spacecraft. Asteroids have the whole periodic table. Low orbit has nothing to offer but free-fall, at least for human activity. As a place to put low-latency megaconstellations, it’s great, but that’s not human presence.


      1. I can see a case for a reasonable-sized presence in LEO, which is that I think the FAA is going to take many years to approve humans taking off and landing in Starship.

        Both take off (without a launch abort system) and propulsive landing are intrinsically more dangerous than the Crew Dragon style launch of a capsule with a launch abort and a non-propulsive landing. Starship will therefore need to demonstrate a very high level of reliability about the launch and landing system – hundreds of successes for every failure – and that will need them to reach a high automated launch cadence and sustain it for several years to convince them of that reliability. I’m confident that SpaceX can get there, but I think it will prove to be a barrier for a long time.

        This means that there will be lots of LEO launches for capsules and the crews will then be transferring to Starships for longer-distance travel – Mars, Moon, wherever. Having a station that can hold ten people permanently and twenty or so more short term would be really useful for gathering people together before transferring to a Starship – that way, four Dragon launches can put 24 into orbit, and 20 of those then leave on the Starship while four are the permanent crew rotation on the space station. The Dragons stay docked to the station until a Starship returns, at which point the crews shift to the Dragons to return to earth.

        If the FAA or other regulators are as intransigent as I think they may be about human-rating Starship, then I wonder if SpaceX will just keep Dragon going for a very long time, or if they will create a new capsule that is launched by a custom Starship. It could still be fully reusable: stock SuperHeavy, but a Starship with the payload bay removed, with a capsule attached via an interstage and a nosecone stored in the interstage to be attached after capsule separation to restore the aerodynamics for landing.

        Beyond-LEO capsules (ie Orion), on the other hand are a complete waste of money: build a simple capsule to get to and from space and then transfer to a Starship.


  5. and now in Oct 2021, Blue Origin etc propose Orbital Reef space station – perhaps hoping to have it funded by NASA as a Commercial LEO Destination.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s