SRC-012

Space Based Solar Power.

Commissioned by NASA OTPS. Who paid for a study is recorded because it bears on how its conclusions should be read, not because it makes them wrong.

Published claims resting on this source

• CLM-0006 None of the five main studies mentions the international radio-frequency exposure limits, or any safety standard, or a restricted area around the receiving site.
• CLM-0123 NASA's own 2024 study puts space solar power at 12 to 80 times the cost it expects in 2050 from power stations on the ground, a set it calls renewable but which includes nuclear, with launch the dominant cost.
• CLM-0124 NASA also found that if several things improve at once, rather than just one, space solar power does become cost competitive.
• CLM-0125 On carbon, NASA found space solar power within the range of five renewable ways of making power on the ground (wind, solar, solar thermal, hydro and geothermal), and far better than coal or gas. NASA also counted nuclear power in that group, though nuclear is not renewable; it sits at the low end, level with wind.
• CLM-0144 A NASA study found that launching everything into space is by far the biggest cost of a space solar power system, and that most of those launches would just be carrying fuel to move earlier payloads from a low orbit up to the final, much higher orbit, not carrying useful hardware.
• CLM-0240 A from-scratch cost model of the "Innovative Heliostat Swarm" design gets almost the same electricity cost as NASA's own study: about 62 cents per kWh versus their 61 cents.
• CLM-0241 For the "Mature Planar Array" design, an independent cost model built the same way as for the other design gets a materially lower electricity cost than NASA's published figure, even though the underlying costs match closely.
• CLM-0242 Working the power output out from scratch, using only the physics of sunlight and the efficiency losses NASA itself lists, gives the same answer NASA published.
• CLM-0243 The huge number of rocket launches NASA calculated for each design can be rebuilt from scratch using just the mass of the hardware, how much a rocket can carry, and how many refuelling trips each delivery needs -- and it comes out almost exactly the same.
• CLM-0246 Adding up the costs from scratch -- rockets, factories, spare parts -- gives almost exactly the same total bill NASA calculated, for both designs.
• CLM-0247 NASA's published cost figure for the "Mature Planar Array" design may be counting the effect of its low duty cycle twice by mistake, which would make the true figure closer to 1.0 dollars per kWh rather than 1.59.
• CLM-0248 How fast the cost of building each solar-power-satellite module falls as more are made turns out to matter more to the final electricity price than the cost of getting them into orbit -- more than any other single thing tested in this model.
• CLM-0249 Simply choosing a different interest rate to value future costs and income changes the final electricity price about as much as most of the technical improvements NASA tested -- and NASA never actually tested this choice.
• CLM-0250 If the design that runs almost all the time had to operate at the same lower duty cycle as the other design, its electricity cost would rise by about two-thirds, all on its own.
• CLM-0256 If several other improvements happen together -- better propulsion, much better solar cells, faster manufacturing learning, and hardware that lasts 15 to 20 years -- then a launch price of roughly 15-30 million dollars per flight would be enough to bring this design's electricity down to about 5 US cents a kilowatt hour. For 2050 NASA projects about 2 to 5 cents for five renewable sources on the ground (wind and solar with storage about 2 cents; geothermal, hydro and solar thermal about 5) and about 5 cents for nuclear, which NASA counts with them though it is not renewable. So 5 cents matches the dearest of these, not the cheapest.
• CLM-0257 For every kilogram of hardware launched into orbit over the station's 30-year life, RD1 generates about 30,000 kilowatt-hours of electricity, and RD2 about 17,500.
• CLM-0258 Every dollar amount in this document is priced as if it were 2022, matching NASA's own report; nothing has been adjusted for inflation to today's prices.
• CLM-0259 Building the cost model completely independently still leads to the same overall conclusion NASA reached: today's space-based solar power concepts cost far more than the electricity NASA projects for 2050 from five renewable sources on the ground and from nuclear (about 2 to 5 cents a kilowatt hour), and no single fix closes that gap on its own.
• CLM-0696 NASA's own comparison table shows exactly why its headline "0.02 to 0.05 dollars per kilowatt-hour" figure needs care. The cheapest entries in that table are wind and solar without storage, which are not firmed: they generate only when the wind blows or the sun shines. The solar panel entry with storage barely changes how much of the time it can generate, only 30 percent instead of 29 percent, so it is not a round-the-clock supply either. The two entries that genuinely can run almost all the time, nuclear and geothermal, are priced at the top of the range, not the bottom, and so are solar-thermal power with heat storage and hydro.
• CLM-0192 The full list of assumptions behind the headline result of NASA's 2024 study, that space solar power would cost far more than ground alternatives, includes a launch cost of $850-1,000 per kilogram, only a 10-year satellite lifetime, and a solar cell efficiency of 35% - each of these, not just launch cost, is doing some of the work in producing the high headline cost figure.
• CLM-0193 To get to a competitive cost, NASA's own 2024 analysis needs launch costs to halve, solar cells to get much more efficient, satellites to last half as long again, and manufacturing to get cheaper faster, all at the same time, not just one improvement on its own; and there is a small but real inconsistency inside the NASA report itself about exactly how low the resulting cost estimate for the smaller design (RD1) is - Table 3 says 4 cents per kWh, the summary text says 3 cents.
• CLM-0407 Cutting the price of a rocket launch brings the electricity cost down to roughly the same low point NASA itself found, using an independently built model.
• CLM-0408 If the hardware lasted the full 30-year mission without needing to be replaced twice, the electricity cost would fall by almost half, on its own.
• CLM-0410 Switching from chemical rockets that need refuelling to electric thrusters that do not would cut the electricity cost by about two-thirds, matching what NASA itself found.
• CLM-0315 Even though NASA's own figures say space solar power's carbon footprint is about as good as five renewable sources of power on the ground (NASA also counts nuclear with them, though nuclear is not renewable), that similarity means paying extra for space solar power buys almost no additional carbon saving compared with those low-carbon sources - and even compared with gas, the implied price per tonne of carbon saved is extremely high, about 1,250 to 1,400 dollars.
• CLM-0466 NASA's report prints the computer code for its cost-learning formula, and running that code gives back the report's own module and servicer costs. The one-line formula written in words just above the code is mistyped, so anyone working from the words rather than the code would get a badly wrong answer.
• CLM-0467 Using much more efficient solar cells would let the whole station be built with far less mass and far fewer launches, and that combined effect cuts the electricity cost by more than a launch-price cut alone would.
• CLM-0468 In this model, free launch alone, with every other input unchanged, still leaves the electricity cost at about 0.10 dollars per kWh or more, because building and replacing the hardware itself costs too much.
• CLM-0600 NASA's launch price was optimistic, not pessimistic. The expensive part is its assumption that each payload needs twelve refuelling flights, which multiplies the cost thirteen times.
• CLM-0601 In one model, swapping refuelling flights for electric-propulsion tugs cuts the electricity cost by about two thirds. NASA's own analysis reports 62 per cent total cost savings for its electric-propulsion variant, and a critic said that using tugs with electric propulsion would be a far more efficient approach. The three are not fully independent, because the model uses NASA's own inputs.
• CLM-0547 Building about 1.46 to 2 million identical satellite modules means the production line has to double its total output about 20 to 21 times over, starting from the first one built.
• CLM-0528 NASA's assumptions imply a high cost per kilogram delivered to the power station's actual orbit, mainly because NASA assumes every launch carrying hardware needs twelve extra fuel-only launches to refuel it for the climb from low orbit.
• CLM-0531 Neither the UK nor the US figure for what it costs to get a kilogram of hardware into the power station's real orbit is based on an actual, completed transaction. Both are assumptions about future launches, and NASA says it makes no claim that its assumed launch price is reliable. For the same rocket and the same starting price per launch, the two figures differ mainly in how hardware is assumed to get from low orbit to its final orbit, and that is what changes the number of launches needed.
• CLM-0534 SpaceX's published Falcon 9 price is lower per launch than the 100 million dollars both the US and UK studies assumed, but higher per kilogram, because Falcon 9 carries far less than the rocket those studies assumed. A real government order in late 2025 for five missions averaged more than 100 million dollars each. Neither figure allows for getting from low orbit up to the power station's real orbit.
• CLM-0522 The UK study used a 20% return requirement, based on what institutional investors are expected to want. The US study used 3%, on a one-line assumption of government funding, and did not test what a different rate would do.
• CLM-0538 The US and UK studies are not just using different toughness levels for the same test -- they are answering different questions. The US study asks "what would this cost the government to build and pay for entirely by itself," which is why its rate is so low. The UK study asks "what return would a private investor need to put money in," which is why its rate is so high. Whether the same would hold for a real UK project like the new nuclear power station has not been checked against a held source.
• CLM-0662 NASA's own report lays out a timeline: build in the 2030s, launch and move to the final high orbit in the 2040s, then run the system from 2050 to 2080.
• CLM-0666 NASA also modelled a third option: a rocket that flies itself all the way to the final orbit without refuelling, at the cost of carrying so much of its own fuel that it can only deliver about a fifth of its normal cargo. This cuts cost a lot but barely cuts emissions, because so many more rockets have to be built.
• CLM-0683 If NASA's own numbers for switching to electric propulsion are applied to the existing side-by-side comparison of the big institutional cost studies, the gap between the cheapest and most expensive study shrinks from about twenty-to-thirty-five times down to about seven-to-ten times. That is a large improvement, but the studies still do not agree.
• CLM-0697 The Chinese 6.6 cent figure ignores the cost of money. Once you charge interest on the capital, the same numbers give about 10 cents at a low interest rate and 21 cents at a high one. Even so, NASA's own figures stay between about 2 and 16 times more expensive, depending which of NASA's cases you pick.
• CLM-0702 Electric trips from a low orbit to a high one took four to seven months on real flights and in most models held here, and NASA's four-month figure is at the fast end of that. One European study expects at least 20 months for its modules, because the time depends on how much thrust each tonne gets.
• CLM-0715 The 2021 UK study assumed one extra fuel launch for every payload launch; NASA's 2024 study assumed twelve. The reason, which neither study states, is not that they disagree about the price of a rocket: the UK plan used chemical rockets only part of the way up, to about 5,900 km, and then switched to electric, while NASA's used chemical rockets all the way to the top. The two studies also assumed different amounts of cargo reach the top per launch.
• CLM-0845 NASA states in its own report that its carbon figures come from a method that converts money spent, by industry sector, into emissions, combined with terms based on mass and area.
• CLM-0846 NASA's report says its carbon figures leave out the effects of rocket exhaust released high in the atmosphere, and in several places says those effects are assumed to be worse than releasing the same emissions at ground level, but it never says how much worse.
• CLM-0849 The range of 13 to 43 grams that NASA compares its satellites against comes from one US laboratory's dataset, and the two ends are the middle values for two different technologies, wind and solar, not the spread of possible values for either.
• CLM-0850 NASA does test how its carbon figures move if assumptions change, and under its most favourable combination they fall by roughly 86 to 89 per cent, to around 4 grams per kilowatt-hour; but the report prints that best-case number differently in three places.
• CLM-0851 NASA itself warns that its method assumes cost and emissions move together, which may not match what is actually measured, and that costs can fall without emissions falling; it describes the whole study as a first approximation and never states a carbon payback time.
• CLM-0878 By late 2022 the Starlink production line had roughly halved the distance to the scale a solar power satellite programme would need, about nine more doublings of output still to go. A doubling means building twice as many as every one built before, so nine of them is a very long way.
• CLM-1349 NASA's report does print how much electricity each design is credited with, in its cost tables, and its carbon figures follow from those amounts exactly. An earlier reading that it had left out one design's part-time operation was wrong.
• CLM-1450 (low confidence) SpaceX's price sheet lists a Falcon 9 launch at 74 million dollars, and 22,000 kilograms as the most it can carry to low orbit if the rocket is not recovered. Using those listed figures with the US study's own way of counting the extra launches needed to reach the power station's orbit gives a cost per kilogram about four times what the US study's own assumptions give, and the UK study's lowest input would need launches about ninety times cheaper than that listed figure. Both studies turn to electric propulsion for part of the climb, NASA only in the case where it finds a competitive cost, and this has not been shown at power-station scale.
• CLM-1473 A 2020 UK government study expects early test satellites to be left to wear out slowly. For full-size satellites, it assumes robots would swap out faulty modules, using the same robots that built the satellite. Those robots do not exist yet. NASA's 2024 study instead plans to replace all of a satellite's hardware every 10 years.

Cite this