A permanent autonomous lunar facility should treat power as a multi-layered survival system: high-reliability baseload from fission, daytime bulk generation from solar, shortfall coverage from storage, and contingency power from radioisotope systems. For a 1000-year civilizational archive, the winning architecture is not one source but a portfolio designed around degradation, redundancy, and in-situ replacement.
1) Solar arrays: abundant, but not the primary long-life backbone
Solar is attractive on the Moon because sunlight is intense and predictable at the poles, but long-term reliability is constrained by dust, radiation, thermal cycling, micrometeoroids, and aging. A commonly cited estimate for future lunar missions is 2–3% annual solar-cell degradation, which compounds severely over decades. At 2% per year, output falls to about 55% after 30 years; at 3% per year, to about 40% after 30 years. Over 100 years, the remaining output is roughly 13% at 2% annual loss and 5% at 3% annual loss, before accounting for catastrophic damage.
Operational consequences:
- Solar should be treated as replaceable infrastructure, not permanent heritage hardware.
- Arrays need modular replacement cycles on the order of 10–25 years, depending on dust exposure and repair capability.
- Designs should assume progressive derating, not fixed nameplate capacity.
- Surface cleaning, tilting, and elevation above regolith help, but do not eliminate the long-term decline.
For a 1000-year facility, solar is best used for:
- Primary daytime generation.
- Battery/fuel-cell charging.
- Noncritical loads.
- Supplementing fission during peak demand.
2) Nuclear fission: the correct baseload for autonomy
For a permanent lunar base, fission is the only mature option that directly solves the 14-day night problem without massive storage burdens.
### Kilopower
Kilopower was NASA’s small-reactor concept for early lunar and deep-space power. It was widely described as a 10 kWe-class modular reactor for near-term surface missions. The advantage is simple: compact, continuous, independent of sunlight, and scalable by replication.
### Fission Surface Power (FSP)
NASA’s Fission Surface Power work has moved toward a practical lunar deployment architecture. A recent NASA FSP project document describes a system with 40 kWe output at 120 Vdc, sized to fit on a lander and capable of operating on the lander or being transported. Industry coverage also describes near-term reactor sizing in the 5–10 kW range for small habitats and 25–50 kW for multi-habitat and industrial nodes[5].
Strategic implication:
- 10 kWe class reactors are suitable for an initial outpost.
- 40 kWe class systems are the minimum credible base unit for a permanently inhabited node with maintenance margin.
- Multiple reactors are preferred over one large unit to provide n+1 redundancy.
For civilizational continuity, the reactor fleet should be designed for:
- Hot-swappable modules
- Shielding that can be incrementally improved with regolith berms
- Ground-side maintenance and spare core components
- A full chain for fuel handling, radiological monitoring, and remote servicing
3) RTGs: excellent for longevity, inadequate for settlements
Radioisotope thermoelectric generators provide extraordinary endurance, but they are not settlement power systems. They convert decay heat to electricity at low efficiency; one NASA source notes up to 7% thermal-to-electric conversion efficiency[4]. That means RTGs are power-dense in reliability, not in output.
What RTGs are good for:
- Emergency beacons
- Instrument packages
- Autonomous monitoring nodes
- Survival heaters
- Backup electronics
What RTGs are not good for:
- Habitat-scale electrical loads
- Mining
- Oxygen production
- Cryogenic processing
- Large storage charging
A settlement needs kilowatts to tens of kilowatts; RTGs are typically suited to watts to low hundreds of watts, not base operations. They remain valuable as a fail-safe last line, especially for long-duration unattended assets.
4) Lunar night storage: 14 days is the central design constraint
The lunar night is roughly 14 Earth days long, and it dominates energy architecture. A lunar ISRU storage study states that night power requirements in an early stage concept are on the order of 10 kWe[2]. Another lunar power study notes that storing enough energy for a rover running at 3 kW over 14+ days would require at least 5 tons of batteries, and cites battery specific energy around 200 Wh/kg as unattractive for this duty[3].
That mass penalty is decisive.
### Batteries
Batteries are useful for:
- Seconds-to-hours buffering
- Peak shaving
- Rover and EVA support
- Short emergency reserve
They are poor for:
- Whole-base 14-day autonomy at meaningful loads
### Regenerative fuel cells
Regenerative fuel cell systems are the best known non-nuclear storage option for long lunar nights. A lunar power review describes RFCS energy density in the range of 200–300 Wh/kg up to 1000 Wh/kg depending on design and power level[3]. Another study reported RFC specific energy of just over 830 Wh/kg for higher nighttime load operation, and 456 Wh/kg for a lower-power keep-alive case.
This is a major improvement over batteries, but the system still requires:
- Water
- Electrolysis hardware
- Gas storage tanks
- Fuel cell stacks
- Thermal management
- Long-life seals and valves
Operationally:
- Use RFCS for night bridging when solar is the main source.
- Use nuclear to reduce required storage depth.
- Use batteries only as local fast-response buffers.
### Thermal storage and other ISRU concepts
ISRU-derived storage concepts are promising for long-lived autonomy:
- Heat-storage systems using regolith as thermal mass can store daytime energy and run