Value chain · CLM-0296
In plain terms
A 2025 peer-reviewed study from Caltech modelled its own space solar power design and calculated three costs of electricity: about 778 US cents per kilowatt-hour using today's technology, 9.4 cents after about ten years of development, and 3.8 cents in a long-run best case about twenty years out. The 9.4 cents figure that is often quoted is the middle case, not a present-day cost. The design is small, about 113 megawatts, and converts only about 4 percent of the sunlight it collects into electricity delivered on the ground.
The claim as recorded
A peer-reviewed Caltech technoeconomic analysis of its own "CSSPS" architecture (flexible phased-array sheet modules in GEO, 10 GHz, 1,600 m aperture diameter, 559 modules) reports LCOE of 777.8, 9.4 and 3.8 US cents/kWh for its "current", "middle" and "asymptotic" technology tiers respectively, at average power ratings of 76, 113 and 181 MW, overnight costs of 487.2, 8.4 and 4.0 USD/W, capacity factors of 69.9%, 81.7% and 81.7%, and end-to-end system efficiencies of 2.77%, 4.14% and 6.63%; the paper cites the NASA OTPS 2024 report in its introduction as one of the existing analyses that "discount certain technological advances with significant economic impact", but disputes none of NASA's cost figures, and its own headline figure is the middle tier, described as requiring about 10 years of technology development, maturation and scaling.
Scope and limits
Do NOT compare the 9.4 c/kWh figure directly against NASA's USD 0.61-1.59/kWh without carrying the three differences that drive it: an 18-fold smaller system (113 MW against about 2 GW), a projected rather than observed launch price, and a 2% discount rate. Do not quote 9.4 or 3.8 c/kWh without the words "middle tier" or "asymptotic", the 10-year or 20-year development horizon, and the fact that the same paper's own current-technology figure is 777.8 c/kWh. The paper's declaration of interests states "The authors have filed numerous patents with the United States Patent and Trademark Office relating to the technologies discussed in this work", so it is an advocate's analysis, peer-reviewed but not disinterested. It disputes none of NASA OTPS's figures, but its introduction does criticise that report's method, as one of the existing analyses that "discount certain technological advances with significant economic impact".
What it assumes
See CLM-0297 for the full assumption set, which is where the divergence from NASA lives. In summary: launch price of about USD 7,462/kg (current tier, Falcon 9 Heavy) falling to about USD 128/kg and USD 55/kg (middle and asymptotic tiers, Starship). These prices come from the paper's own launch cost model, which takes the Starship GEO transport format from SpaceX's 2020 Starship User's Guide, prices fuel and Falcon 9 upper-stage hardware, and then reduces the result by a logarithmic fit to historic launch cost data projected 10 and 20 years forward; GEO, reached by launch to LEO, refuel, transfer to GTO then GEO; discount rate 2%; rate-adjusted energy multipliers N2 of 7, 11 and 13 years; capacity factor 69.9% or 81.7%; 76 to 181 MW single system, first of a kind, no learning-curve or Nth-of-a-kind treatment found.
Where it comes from
Table 3 "Benchmark comparison for different energy sources": "CSSPS, current 487.2 [OC $/W] 69.9 [CF %] 777.8 [LCOE c/kWh]; CSSPS, middle 8.4 81.7 9.4; CSSPS, asymptotic 4.0 81.7 3.8", against "PV, utility 0.8 16.2 4.4", "Wind, onshore 1.2 36.0 3.3", "Wind, offshore 2.8 41.0 7.5". Table 2 row values: "Average power rating MW 76 113 181"; "Overnight cost $/W 487 8 4"; "LCOE c/kWh 777.8 9.4 3.8". System efficiency at f = 10 GHz: "system 2.77% 4.14% 6.63%". Summary: "A technoeconomic analysis of our proposal concludes with estimates of a LCOE of 9.4 c/kWh and a lifetime cost of $983 M for a 10 GHz, 1,600 m diameter, 113 MW SSP system." Tier definitions: "(3) Asymptotic: this is an estimate of what is technologically and economically possible with existing materials but new design and manufacturing advances in 20 years... We intend for the middle tier to serve as our primary estimate". NASA citation, reference 46: "Rodgers, E., Gertsen, E., Sotudeh, J., Mullins, C., Hernandez, A., Nguyen, H., Smith, P., and Jospeh, N. (2024). Space-based solar power, nasa otps report 20230018600."
Basis: USD, basis year not stated in the paper; submitted 2024-07-29 and published 2025-05-13, so treat as approximately USD2024
Related records
- Answers question
- QST-037 not published in this release
Cite this
energyaperture.com/c/clm-0296
CLM-0296, Energy Aperture, as of 2025-05-13.
Connected records
Other published records linked to this one. A list, not a judgement.
- Also from SRC-006
- CLM-0179This 2025 paper's own numbers show that with today's technology, their space solar power design would cost about $7.78 per kWh - far more...CLM-0180The low headline cost figure in this paper depends on using a very low 2% discount rate borrowed from a US home-energy-efficiency...CLM-0181The 9.4-cents and 3.8-cents per kWh figures both depend on launch prices that do not exist yet - they are extrapolated 10 and 20 years into...CLM-0297The low electricity costs in a 2025 Caltech study rest on two assumptions among others. First, it assumes that SpaceX's Starship will in...CLM-0670Caltech has two different space-solar-power efforts. One is a paper design that assumes a normal rocket-fuel-based trip to a high orbit...CLM-1201Caltech's design has no single large structure. Many separate square craft, each 60 metres across, fly side by side in geostationary orbit...CLM-1203Caltech's receiving station on the ground is a circle about 1.6 kilometres across, just large enough to catch the whole central spot of the...