Technology Briefing · Space Nuclear Power · March 2026

Nuclear Power
for the Moon

From RTGs powering a microwave to 100 kW fission reactors sustaining a lunar base — a technical briefing for policymakers on America's nuclear space advantage.

🚀 2030 TARGET LAUNCH ⚛️ KRUSTY TESTED 2018 🇺🇸 NASA + DOE + INL

PROGRAM INVESTMENT

$500M

Per year from FY2027

$18M

KRUSTY entire program cost

1

Understanding the Numbers: Thermal vs. Electric Power

Wt 🔥

Watts Thermal

Raw heat energy released by the reactor through fission or radioactive decay. This is what the fuel produces — but it cannot directly power electronics or life support.

🔥 Example

A 40 Wt reactor produces heat equivalent to 40 candles burning — constantly, for years.

Conversion Efficiency

RTG Thermoelectric 5%
Thermoelectric conversion
Semiconductor pairs (thermocouples) generate voltage from a temperature difference. Extremely reliable — zero moving parts — but thermodynamically inefficient. RTGs on Voyager still work after 47 years.
Stirling Engine 25%
Free-piston Stirling engine
Heat expands sealed gas, pushing a piston that drives a linear alternator. No lubricants, hermetically sealed. Kilopower uses this. 5× better than thermoelectric. Proven in 2018 KRUSTY test.
Brayton Turbine 35%
Closed Brayton cycle gas turbine
Hot helium-xenon gas spins a turbine connected to a generator. Same principle as a jet engine — but closed loop, no combustion. Scales cleanly to 100 kWe+. Mandated for NASA's FSP 100 kWe program.
Stirling — Lunar Night Boost 🌕 ~40–45%
Daytime: 25% Night bonus: +15–20%
Lunar night efficiency gain
Carnot: η = 1 − T_cold/T_hot. As ambient drops from +127°C (day) to −173°C (equatorial night), the cold-end temperature falls from ~423K to a regulated ~213K, pushing actual efficiency from ~25% toward ~40–45%. A 10 kWe unit delivers ~14–16 kWe at night — a 50–60% power bonus. Radiator water-pipe freeze risk is managed with variable-conductance heat pipes throttled to hold cold end above −60°C.

Hover any bar for details

We ⚡

Watts Electric

Usable electricity after heat conversion. This is what powers lights, life support, computers, drills, ISRU equipment, and everything else in a lunar base.

⚡ Key Insight

A 10 kWe reactor delivers the same electricity as a typical American home at peak — from a system the size of a large desk.

The Kilopower Math: Wt → We
40 Wt thermal
The reactor core produces 40 kilowatts of raw heat (40 kWt). This is the fission output — not yet usable electricity. Think of it as 40 kW of fire that must be converted.
 → 25% Stirling → 10 We electric + 30 Wt waste heat
The 30 kWt not converted to electricity is radiated away — but on the Moon this "waste" heat is redirected to warm electronics, drive ISRU chemical reactions, and prevent water ice from refreezing. Nothing is truly wasted.
40 kWe: The Original Target

NASA's 2022 Phase 1 FSP program spec'd all three contractors (Lockheed, Westinghouse, IX) to design a 40 kWe system at ≤6 metric tons — achievable with four modular 10 kWe Kilopower units in an array. That covers life support, basic science, and minimal ISRU for a small crew. The August 2025 directive upgraded the requirement to 100 kWe at ≤15 tonnes to enable full-scale ISRU propellant manufacturing and match the strategic stakes of the China/Russia race. Think of 40 kWe as "keep the lights on" and 100 kWe as "build the base."

2

Power Scale: From a Microwave to a Moon Base

Power Output Comparison (We)

🍕

1 kWe

KRUSTY Demo Reactor (2018)

≈ 9 MMRTGs

= 1 microwave oven · 1 space heater · 10 LED bulbs · Curiosity rover × 9

KRUSTY 2018
First US fission reactor tested in 50+ years. 28-hour full-power nuclear test at Nevada National Security Site. Cost: $18M total. Proved the heat-pipe + Stirling architecture works in nuclear conditions.
🏠

10 kWe

Kilopower Flight Unit — Crew Life Support

DESIGN READY

= 1 average American home at peak · crew of 4 in lunar habitat · 6 EV chargers

10 kWe Flight Unit
Scales directly from KRUSTY demo. 43.7 kg of U-235 fuel. 8 sodium heat pipes. 8 Stirling converters. ~2,000 kg total mass. Provides adequate power for life support, science, and communications for a 4-person crew.
🏭

40 kWe

Original FSP Target — Minimum Viable Base

PHASE 1 DESIGN

= 33 avg U.S. homes · 4× 10 kWe Kilopower units · small ISRU plant · crew + science ops

Why 40 kWe was the original target
NASA's 2022 Phase 1 contracts (Lockheed, Westinghouse, IX) were all spec'd to 40 kWe at ≤6 tonne mass. This covers life support, science, and modest ISRU for a minimal crewed outpost. It was achievable with 4 modular Kilopower units in array. Upgraded to 100 kWe in August 2025 to enable full ISRU and sustained base operations — and to beat China.
🏢

100 kWe

NASA FSP Target — Lunar Base + Full ISRU

2030 GOAL

= 30 Teslas charging · medium office building · 10 homes · full ISRU propellant plant

Why 100 kWe?
ISRU requires 31–38 kWe continuously for 12–15 months before crew arrives. Add habitat systems, rover charging, science, and redundancy margin. 100 kWe is the minimum viable power for a sustained presence capable of manufacturing its own return fuel — and staking a territorial position ahead of China.
3

Three Nuclear Technologies for Space

☢️

Radioisotope Thermoelectric Generator

FLIGHT PROVEN · NO FISSION · NO MOVING PARTS

FuelPlutonium-238 (not weapons-grade)
Power Output110 – 300 We
Conversion~5% thermoelectric
Moving PartsZERO
Lifespan47+ years (Voyager)
ScalabilityCANNOT SCALE
Cost per kg of Pu-238~$8–10 million/kg · Annual US production: ~1.5 kg/year
Used on: Voyager 1 & 2, Cassini, Curiosity, Perseverance, New Horizons, Viking. To get 100 kWe from RTGs would require ~900 MMRTGs and ~30 years of full US Pu-238 production. Not viable for a base.

Best Use Cases

✅Outer solar system (Jupiter+) where solar fails
✅Long-duration rovers (12+ years continuous)
✅Deep space probes with zero maintenance
✅Sensors in permanently shadowed craters
❌Cannot scale to base power needs
❌Plutonium supply is severely constrained

Power Comparison

9×

MMRTGs needed to match a single 1 kWe Kilopower unit

⚛️

Kilopower / KRUSTY

HEAT PIPE FISSION · STIRLING ENGINE · TESTED 2018

FuelUranium-235 (93% HEU)
Power Output1 – 10 kWe
Conversion~25–30% Stirling
Total Mass (10 kWe)~2,000 kg
Program Cost~$18M total
ScalabilityMODULAR ARRAYS
Core DesignU-Mo core + BeO reflector + sodium heat pipes + Stirling engines. Single B₄C control rod. Passively self-regulating — reactor power adjusts automatically via thermal expansion physics.
KRUSTY Test Result: 28-hour full-power nuclear run, March 20–21, 2018, Nevada National Security Site. Every failure scenario survived. Poston: "We threw everything we could at this reactor and it passed with flying colors."

Why It Matters

✅Only U.S. space fission system tested in 50 years
✅Cost 1/50th of prior failed programs
✅Self-regulating — no operator needed
✅Lunar night efficiency bonus 50–80%
✅Not radioactive until first turned on — safe to launch

$18M

vs $1B+ for SP-100 which produced no hardware

50 yrs

gap since SNAP-10A — KRUSTY ended the drought

🚀

Fission Surface Power 100 kWe

HALEU FUEL · BRAYTON CYCLE · LAUNCH TARGET Q1 FY2030

FuelHALEU (19.75% U-235)
Power Output≥ 100 kWe
Conversion~35–40% Brayton cycle
Mass Budget≤ 15,000 kg
Program Budget$350M FY26 → $500M/yr
ContractorsLockheed, Westinghouse, IX
Radiator System~128 m² deployable titanium-water heat pipe radiator panels. 1 km power cable separates reactor from habitat. 5 rem/year dose limit at 1 km.
Strategic Context: China and Russia announced joint lunar reactor program targeting 2035. NASA August 2025 directive warns: first nation to deploy "could declare a keep-out zone which would significantly inhibit U.S. Artemis presence."

Key Risk: HALEU Supply

HALEU fuel supply is the hidden critical path. Centrus Energy is essentially the only U.S. enrichment source. Insufficient domestic production could delay the entire program regardless of reactor design readiness.

Contractor Teams

Lockheed Martin + BWX + Creare
Westinghouse (AstroVinci eVinci)
IX (Intuitive Machines + X-energy)

$3B

Estimated 5-year cost to hit 2030 launch (Lal, former NASA Assoc. Admin.)

4

Power Conversion: Stirling vs. Brayton

Stirling Engine

Free-Piston · No Lubricant · Hermetically Sealed

🔩
Heat expands sealed gas in a cylinder, pushing a free piston that drives a linear alternator. No combustion, no valves, no lubricant. Proven by 11+ continuous years at NASA Glenn. The engine used in the 2018 KRUSTY test.
✓Zero moving external parts — hermetically sealed for life
✓25–30% efficiency (5× better than RTG thermocouples)
✓Proven in KRUSTY 2018 — exceeded all performance targets
✓Lunar night efficiency bonus: cold ambient → higher ΔT
✗Doesn't scale efficiently above ~50 kWe total output
✗Cold end may freeze in deep lunar night without thermal controls
Best for: 1–40 kWe · Modular arrays · Robotic missions · Early outposts

Brayton Turbine

Closed Cycle · Gas Turbine · High Temperature

⚙️
Hot helium-xenon gas expands through a turbine spinning a generator. Gas is cooled through a radiator and recompressed. Closed loop — no fuel combustion, no intake, no exhaust. Same thermodynamic cycle as a jet engine, adapted for space.
✓35–40% efficiency — best available at high power
✓Scales cleanly: 100 kWe → 1 MWe and beyond
✓Operates at higher hot-end temperatures (1,000°C+)
✓Mandated by NASA Aug 2025 directive for FSP 100 kWe
✗Turbine bearings require careful engineering in vacuum
✗Less proven at space scale (GE, Rolls-Royce under contract now)
Best for: 40–1,000+ kWe · Permanent bases · ISRU plants · Mars power grid
🏛️

Bottom Line

Kilopower proved the Stirling approach works and produced real results for $18M. The FSP 100 kWe program switches to Brayton because scaling to 100 kWe with Stirling would require ~10 separate units vs. one integrated Brayton system at lower total mass. Think of it as going from 10 portable generators to one commercial power plant — same nuclear fuel, smarter architecture at scale.

5

Nuclear Reactors: Earth vs. Moon

🌍

Earth Nuclear Reactors

Water-cooled · Atmosphere · Gravity · Grid-scale

💧

Pressurized Water Reactor (PWR)

Most common U.S. type (Navy subs, commercial plants). Pressurized water as coolant and moderator. Output: 1,000–1,600 MWe. Cannot work on Moon — no water.

PWR on the Moon?
Theoretically possible but impractical. Water would need to be imported or extracted from lunar ice. No cooling towers possible in vacuum — all heat rejection must be via radiation. Mass and complexity become prohibitive vs. purpose-built designs.
♨️

Boiling Water Reactor (BWR)

Water boils directly in reactor vessel, drives steam turbine. Simpler than PWR but same water dependency. Output: 600–1,400 MWe.

🧱

TRISO / Pebble Bed (Advanced)

Uranium coated in ceramic — physically meltdown-proof. High-temperature gas cooled. X-energy proposes this for lunar use. Strongest safety profile of any reactor design.

🔬

SMR / Microreactor (DoD Project Pele)

1.5 MWe, fits in shipping containers. TRISO fuel, helium Brayton cycle, forward military bases. BWXT. Full core delivered to INL Nov 2025. Shares Brayton tech with FSP.

🌕

Moon-Optimized Reactors

No water · Vacuum · Radiator-cooled · Autonomous

🔧

Kilopower Heat Pipe Reactor (1–10 kWe)

U-Mo core + sodium heat pipes + Stirling engines. Fully passive — no pumps. Self-regulating. Only U.S. space fission system tested in 50 years. TRL-5, tested 2018.

💡

eVinci / AstroVinci (Westinghouse)

Sodium heat pipe cooled microreactor, 10–100 kWe range. Adapted from commercial eVinci design. Brayton or Stirling. Competing for FSP 100 kWe contract.

🏗️

FSP 100 kWe Brayton System

HALEU fuel, closed Brayton cycle, 15-tonne mass, 128 m² radiator panels. 1 km power cable to habitat. 10-year unattended operation. Launch target: Q1 FY2030.

❄️

The Lunar Cold Advantage

Permanently shadowed polar craters: −250°C. Carnot limit: 97.8%. A Stirling unit rated 10 kWe in daytime delivers ~16 kWe at lunar night — a 60% power bonus exactly when solar produces zero.

6

Comparative Analysis: Cost · Mass · Power

Power Output by System (We)

Efficiency Comparison (%)

MMRTG used on Curiosity and Perseverance. $8–10M/kg of Pu-238 fuel. ~900 units needed for 100 kWe. Not viable for base power.
Breakthrough program: $18M for a fully tested working reactor vs. $1B+ for SP-100 which produced nothing. Changed how NASA thinks about space nuclear development.
August 2025 directive: min 100 kWe, ≤15 tonne, closed Brayton cycle, launch Q1 FY2030. Awards to 2 companies expected March 2026, down-select to 1 at PDR.
System Power Mass W/kg Program Cost Status
MMRTG (RTG)
Thermoelectric · Pu-238
 110 We 45 kg 2.4 W/kg ~$110M/unit FLIGHT PROVEN
GPHS-RTG (Cassini)
Thermoelectric · Pu-238
 300 We 57 kg 5.3 W/kg ~$60M/unit FLIGHT PROVEN
Kilopower 1 kWe (KRUSTY)
Heat Pipe · Stirling · HEU
 1,000 We ~400 kg 2.5 W/kg $18M total GROUND TESTED
Kilopower 10 kWe (Flight)
Heat Pipe · Stirling · HEU
 10,000 We ~2,000 kg 5.0 W/kg $50–100M est. DESIGN READY
FSP 100 kWe (Target)
Fission · Brayton · HALEU
 100,000 We ≤15,000 kg 6.7 W/kg ~$2–3B est. IN DEVELOPMENT
SNAP-10A (1965)
U-ZrH · NaK pump · Thermoelectric
 590 We 435 kg 1.4 W/kg Classified (1960s) FAILED 43 DAYS
SP-100 (Canceled 1994)
UN fuel · Li coolant · Thermoelectric
 100,000 We ~5,400 kg 18.5 W/kg ~$1B, no hardware NEVER BUILT
The Kilopower Lesson: The $18M KRUSTY program produced a working, tested reactor for less than 1/50th of the SP-100 program cost. The key was designing for testability and simplicity rather than maximum performance from day one.
7

Key Findings for Action

KRUSTY passed a 28-hour full-power nuclear test in March 2018 — the first successful U.S. space fission test in 50 years. Every performance metric was met or exceeded. The physics are proven. Kilopower cost $18M total; prior programs spent over $1 billion without producing flight hardware. The question now is political will and funding continuity.
Mars dust storms drop solar irradiance by up to 99%. The 14-day lunar night produces zero solar power. ISRU operations require 31–38 kWe continuous for 12–15 months uninterrupted. Robots can hibernate; humans cannot. Nuclear is the only viable power source for any permanent human presence on the Moon or Mars.
China and Russia announced a joint lunar reactor program targeting 2035. The first nation to operate a reactor on the Moon may claim safety exclusion zones. The NASA August 2025 directive explicitly frames FSP as a strategic competition: "The first country to get a reactor on the moon could potentially declare a keep-out zone which would significantly inhibit the United States from establishing a planned Artemis presence."
The FSP program requires HALEU (19.75% enriched uranium). Centrus Energy is essentially the only U.S. enrichment source. Production capacity is severely constrained. Congress should mandate a parallel HALEU fuel production ramp-up through DOE — or the reactor will be designed and contracted before its fuel supply exists, creating a single-point-of-failure outside NASA's control.
Former NASA Associate Administrator Bhavya Lal estimates $3B over 5 years — alongside substantial DOE National Lab technical support — is required to credibly achieve a 2030 launch. The FY2026 budget requests $350M rising to $500M annually. Under-resourcing this program risks repeating the SP-100 pattern: years of effort and hundreds of millions spent, no hardware to show for it.
Arrays of 10 kWe Kilopower units offer maximum redundancy — lose one, others continue operating. A single 100 kWe Brayton system is more mass-efficient but is a single point of failure with no repair option. For a crewed Mars mission with no rescue capability, redundancy may outweigh mass efficiency. Congress should require NASA to address this tradeoff explicitly in contractor proposals and in its own mission risk assessment.