Lunar Energy Infrastructure

Lunar energy infrastructure is the generation, storage, distribution and thermal management of electrical power on the Moon's surface: photovoltaic arrays, fission reactors, radioisotope units, batteries, regenerative fuel cells, cables and beamed power, and the heaters, radiators and insulation that keep all of it alive between sunrises. It is a distinct subject from spacecraft power because the environment is distinct, and every number in this brief is meaningless without the environment stated alongside it.

The environment, stated once and repeated with every figure below. Away from the poles the Moon receives roughly fifteen Earth days of sunlight followed by roughly fifteen of darkness — a night of about 354 hours, worked by NASA Glenn as 350 h for sizing purposes. At the south-polar ridges the Artemis programme is actually targeting, the same NASA primer gives over six months of continuous sunlight at 10 m elevation with three to four days of maximum continuous darkness, and NASA Glenn independently works polar eclipse durations “nearer to 100 h.” Surface temperatures swing from about +22 °C to −220 °C, with peer-reviewed modelling putting the night surface at about 95 K. There is no atmosphere, so waste heat leaves only by radiation. And the regolith is an electrostatically adhesive, abrasive, sub-100-micrometre dust that degrades optical surfaces on contact. Any figure quoted without saying whether it is equatorial or polar, and without saying what the thermal boundary condition is, is not an engineering figure.

This brief develops a bottleneck another brief names. Lunar industry lists “surviving the two-week lunar night (power and thermal)” as both a frontier question and a bottleneck — identified and not opened. This is that question opened. Mining, in-situ resource use and manufacturing economics stay there and are not re-argued here; what stays here is everything measured in watts delivered, kilowatt-hours stored, kilograms landed and kelvin maintained.

The scope boundary with small modular reactors is worth stating precisely, because the two topics share a physics and share nothing else. That brief owns fission economics against a terrestrial grid, where the comparison is a levelised cost against alternatives that already exist. This brief owns fission against batteries, solar and isotopes at a site where the only currency is landed mass and there is no grid to be cheaper than. A reactor that would be uncompetitive on Earth can be the obvious answer on the Moon purely because a kilogram costs between $0.4 million and $1.2 million to put on the surface.

One preview, because it inverts the usual framing and the rest of the brief depends on it. The interesting question is not whether fission beats batteries — at an equatorial site it does, by an order of magnitude, and the argument is over. The interesting question is where the crossover falls at the polar sites anyone is actually going to, at what load and at what duty cycle; and the finding of this brief is that a programme now budgeted at $350 million in FY2026 rising to $500 million a year has not published the trade study that would answer it.

Established The framing everyone uses is wrong about the number that matters, and the correction is the most useful single thing in this brief. Lunar night at equatorial and mid latitudes is about 354 hours, and that figure is correct as a worst case. It is the wrong design point for Artemis. NASA's own lunar power-needs primer states that at 10 m elevation near the poles you get over six months of continuous sunlight and three to four days of maximum continuous darkness; NASA Glenn's regenerative-fuel-cell sizing paper independently puts polar shaded durations “nearer to 100 h” against 350 h at the equator. The design point at the sites that matter is therefore roughly 72 to 100 hours, not 354, and that is a factor of nearly four in stored energy before any technology has been chosen.

Established Work the arithmetic both ways, because it is the whole argument. The inputs are sourced and the multiplication is this brief's. Carrying a 10 kW load needs 3,540 kWh through an equatorial night and about 960 kWh through a 96-hour polar one. ESA states the current specific energy of space-capable lithium-ion at 200 Wh/kg; at 80% usable depth of discharge that is 22,100 kg against 6,000 kg of cells. Apply a conventional cell-to-pack fraction of 0.7 for structure, harness, battery management and survival heaters and the equatorial figure passes 31,600 kg. Against that, NASA's Kilopower charts give 10 kWe for 1,500 kg and 1 kWe for 400 kg.

Established At the equator, fission wins by an order of magnitude and the argument is over. At the poles the margin narrows to about fourfold — and it narrows further the moment you apply the survival-power lever. NASA's own power budgets give habitats and laboratories 4 to 10 kW nominal against 0.5 to 2 kW survival; large in-situ resource plants 20 to 100 kW nominal against 2 to 5 kW survival; large pressurised rovers 7 to 20 kW against 1 to 3 kW. That survival-to-nominal ratio of roughly one to five or one to ten is the single most effective mass lever available and it is routinely omitted from “power through the night” framings. A 2 kW survival load through 96 hours is 192 kWh, about 1,200 kg of cells — less than the Kilopower unit, and it needs no nuclear launch authorisation, no highly enriched uranium supply chain and no fifteen-tonne lander. Again, the arithmetic is this brief's and the inputs are sourced.

Frontier So the real frontier question is not how to power through the night. It is: at what site, at what duty cycle and at what load does the mass-and-cost crossover between vertical solar plus storage and fission actually fall — and does any mission architecture sit on the fission side of it before industrial resource extraction begins? This brief cannot answer that, and neither, on the published record, can anyone else.

Established That is the finding that cuts against the field's own framing, and the evidence for it is a pattern of omission across four government documents. NASA Glenn's flagship regenerative-fuel-cell sizing paper explicitly declines to compare against fission. NASA's own fission surface power requirements paper contains no comparison against solar plus storage. NASA's energy-storage-for-lunar-surface-exploration paper is explicitly methodological, offering a sizing framework from 100 W to 50 kW with no baseline cases. And the 2026 peer-reviewed vertical-solar-array farm study punts on storage entirely, saying survival power “will need to be provided either in the form of a nuclear-based power source, or stored power” with no specifications either way. Each document is competent within its own scope. Together they mean that a programme spending $350 million in FY2026, rising to $500 million a year from FY2027, has not published the trade study that justifies it.

Established The second finding that cuts against the framing: the binding constraint is thermal, not electrical, and thermal has a cheaper fix. The lander survival record, tabulated in the experiments section below, points one way. Every vehicle that has survived a lunar night carried radioisotope heater units or survived by accident; every vehicle that relied on batteries and insulation died. Firefly's Blue Ghost Mission 1 completed 346 hours of daylight operations and then lasted just over five hours into the night — it did not run out of energy, it lost the thermal fight first. Maintaining roughly −40 °C or better inside an equipment box for hundreds of hours against a 95 K sky with no sun is a heat problem, and the cheapest known solution to it is a plutonium pellet rather than a battery. Base rate as of August 2026: of Western commercial landers that reached the surface intact, none has survived a lunar night.

Established Fission surface power is real, is directed, is funded, and is unbuilt. KRUSTY ran four test phases over five months at the Nevada National Security Site from November 2017 to March 2018, culminating in a 28-hour full-power test on 20–21 March 2018 that covered startup, ramp, steady operation and shutdown including simulated engine and heat-pipe failures. Three 12-month design contracts at approximately $5 million each were awarded on 21 June 2022 to Lockheed Martin, Westinghouse and IX, with Phase 1A extensions in January 2025. A directive signed by Acting Administrator Sean Duffy on 4 August 2025 raised the requirement to a minimum of 100 kWe on a mandatory closed Brayton cycle, with launch readiness in Q1 FY2030. No flight hardware, no test article and no reactor exists. What exists is three paper studies and their extensions.

Handwave The directive's own schedule, and every schedule downstream of it. The directive required a Program Executive within 30 days, an RFP within 60 days and awards within six months of RFP release — awards by roughly April 2026. As of the pack's compilation on 10 August 2026, only a Request for Information (14 August 2025) and two draft Announcements for Partnership Proposals (29 August and 5 December 2025) had issued; the final announcement was promised “early 2026” and no provider award has been announced. The Program Executive appointment did happen, on 25 September 2025. This brief records the absence of an award as an absence of news rather than as confirmed fact: the pack could not read the SAM.gov contracting record, which returns an empty JavaScript shell to automated retrieval, so the acquisition status is assembled from NASA Glenn's own updated pages and trade press.

Established Two official positions on the same programme contradict each other, and this brief does not resolve them. On power level: NASA's own ESDMD fission surface power page still reads “at least 40 kilowatts,” last updated 8 September 2025, and Idaho National Laboratory's project page still describes “a 40 kilowatt-electric nuclear reactor” with news items dated 6 August 2026 — a full year after the directive mandated at least 100 kWe. On delivery date: the directive says launch-ready Q1 FY2030, while NASA's own three-phase Moon Base plan, as reported in March 2026, places fission surface power delivery in Phase Three (2033–2036), with Phase Two (2029–2032) getting only “hundreds of watts” from radioisotope generators. These two official positions are three to six years apart. That the programme's own public pages have not been reconciled with its governing directive a year on is itself a status signal.

Established The regenerative fuel cell does not rescue the storage case as cleanly as it is advertised to. ESA states that regenerative fuel cell specific energy is “more than double” the 200 Wh/kg of lithium-ion, implying above 400 Wh/kg and roughly 8,850 kg for the equatorial 3,540 kWh case. But NASA Glenn's measured and modelled round-trip efficiency for such a system is 28 to 37% — 31.7% at 1 kW, about 36.6% at 50 kW — not the roughly 80% theoretical maximum. At that efficiency the daytime array must be about three times the night load on top of the day load. The halved storage mass is bought with a tripled generation plant, and that trade is understated in almost all promotional material on the technology. The same paper masses a 1 kW regenerative fuel cell system at about 260 kg, with reactant storage roughly 30% of that.

The genuinely open questions in lunar surface power are almost all questions about a specific site, a specific duty cycle or a specific unpublished number. Very little of what is open is open because the physics is unknown. Separating that from the topics that merely sound open is the main work of this section.

Frontier The single most load-bearing missing number in the subject is the worst-case continuous darkness at a specific candidate polar site. The generic figures exist: three to four days at 10 m elevation from NASA, and about 100 h from NASA Glenn. No site-specific worst case for the Shackleton–de Gerlache Connecting Ridge — the ridge Blue Origin's Blue Moon MK1 is targeting and the one most polar architectures assume — could be verified from any source consulted for this brief. The 2026 peer-reviewed vertical-array farm study identifies unavoidable blackout periods when “the entire hill is in darkness from the shadow of Mons Mouton” during the first third of the year, and does not quantify the longest one. That unquantified duration is the number that sizes the storage, and therefore the number that decides the solar-versus-fission crossover at the only site anyone is currently building toward. This brief states it as a gap rather than estimating it.

Frontier Regolith as thermal mass and thermal store is genuinely underexplored, and the numbers are surprising. Peer-reviewed modelling published in March 2026 finds compacted-regolith thermal storage reaching 0.25 kWh/kg over one lunar day, that roughly 1 m of external regolith substantially reduces heat flux, and that 2 to 3 m of low-density regolith smooths the diurnal variation. The same work models a solar-thermal integrated system delivering 10.8 kW average electrical during the lunar day and a stable 7.0 kW at night at 48.0% overall cycle efficiency. Note that 0.25 kWh/kg for thermal storage in regolith exceeds 0.2 kWh/kg for electrical storage in space-qualified lithium-ion — and regolith is free on site, whereas every cell is landed at up to $1.2 million a kilogram. Nobody has flown buried thermal mass. Nobody has flown a survival-mode load-shedding architecture designed around the one-to-five nominal-to-survival ratio. Almost the entire public discussion is about batteries and reactors.

Frontier The electrodynamic dust shield worked in flight and nobody has published how well. NASA's Electrodynamic Dust Shield flew on Blue Ghost Mission 1 in March 2025 and “successfully lifted and removed lunar regolith using electrodynamic forces on the glass and thermal radiator surfaces” — the first in-situ dust mitigation result anywhere. No quantitative removal percentage, and no before-and-after transmissivity, has been published for the flight unit. Ground and laboratory work reports 80 to 90% removal in air and near-100% in vacuum, but a laboratory percentage on simulant is not a flight percentage on real regolith after a real landing plume. The flight number is genuinely unpublished as far as this brief could verify, and it is a number that changes the maintenance model for every surface power system.

Frontier A de facto lunar power standard is emerging through procurement rather than through any standards body, and this brief has not seen the point stated anywhere. NASA's 2022 fission surface power requirements paper fixed 40 kWe at 120 Vdc. The December 2025 draft Announcement for Partnership Proposals specifies ten user-interface connector ports at 120 VDC, each rated at 6 kWe or more, with autonomous fault detection, isolation and recovery, and restart without external power. That is the closest thing to a lunar power interface specification in existence, and it exists as a purchase requirement rather than as a standard. Whether it becomes the bus voltage by default, and whether anyone other than NASA's eventual provider designs to it, is open.

Frontier Power beaming to the surface is at the stage where demonstrations are announced without numbers. Star Catcher and Intuitive Machines beamed laser power to a lunar terrain vehicle's standard solar panels at Kennedy Space Center on 12 November 2025, claiming to deliver “meaningful power levels” without custom receivers; no power level, distance or efficiency was disclosed. Volta Space's LightPort optical receiver was added to Blue Ghost Mission 2 on the far side in December 2025, again with no specifications, with the company's orbital transmitters notionally 2028 — a receiver manifested roughly four years before any transmitter exists. That the ground demonstration occurred is established; the performance claims are unevaluable and are treated as handwave until numbers appear.

Frontier Helium-3 extraction has one genuinely interesting recent result and it is about energy, not about fusion. Thermal release of implanted helium from regolith runs 50% at 500 °C and 95% at 800 °C, and the heating is the dominant energy line item in every mining concept. Recent agitation-based testbed work removed 70% of implanted helium at an agitation rate of 1.5 g/s and 96% at 9.0 g/s at room temperature — equivalent to roughly 500 °C thermal extraction without heating. If that holds at scale it removes the single largest energy cost in the concept. It does not make the concept an energy source, for reasons in the misconceptions section.

Handwave What merely sounds open: whether solar or nuclear “wins.” The site settles it more than the technology does. At an equatorial or mid-latitude site with a multi-kilowatt continuous load, fission wins on mass by an order of magnitude and there is no serious counter-argument. At a polar ridge with a survival-power duty cycle, batteries plus a vertical array can beat it on landed mass and avoid the entire nuclear launch-approval chain. Framing the subject as a technology contest rather than a siting-and-duty-cycle question is the characteristic error, and it is the reason the crossover has never been published.

Handwave What merely sounds open: radioisotope power at base scale. This is arithmetically closed, not open, and the section below on bottlenecks gives the numbers. Radioisotope power is structurally a watts-to-hundreds-of-watts technology for the foreseeable future. It solves keep-alive and instrument power. It does not solve base power, and no plausible production ramp changes that within a generation.

Established The first bottleneck is landed mass multiplied by landed cost, and every other consideration reduces to it. At $0.4 million to $1.2 million a kilogram, a power system choice on the Moon is an argument about kilograms with a very large multiplier attached, and the multiplier dominates all operational considerations including efficiency, lifetime and maintainability. The delivery market is also not improving as fast as the aspiration: trade analysis of the Commercial Lunar Payload Services programme at thirty task orders documents average cost growth of 26% and average schedule slip of 14 months per task order, and notes that every currently active lander operates below 500 kg of surface capacity against NASA's four-tonne Phase 1 aspiration. The fission directive assumes a fifteen-tonne heavy-class lander. No such lander has flown.

Established The second is thermal survival, which is a separate problem from electrical supply and is routinely folded into it. Keeping electronics and cells above their minimum temperature for hundreds of hours against a 95 K sky costs energy continuously whether or not any payload runs, and the empirical record says it is the killing constraint. Blue Ghost Mission 1 had a 346-hour charge behind it and lasted five hours past sunset. Every vehicle that has survived a night carried radioisotope heater units, or, in SLIM's case, survived unintentionally. Radioisotope heat supplies warmth without electrical conversion, which is why it wins a competition the electrical accounting does not even score.

Established The third is dust, and the correct framing of it is not the usual one. Natural deposition is negligible on mission timescales: rediscovered Apollo dust-detector data from 1969–1977 gives about 100 µg/cm²/yr, roughly one millimetre per thousand years. The dust threat is episodic and anthropogenic — landing plumes, ascent plumes, rover traffic and extravehicular activity. Apollo 11's own dust-detector cells measured a single lunar module ascent costing 7% of output on a cell with 0.15 mm cover glass and about 17% on one with 0.5 mm cover. A base with regular landings therefore faces a repeating step-function degradation, not a decay curve, and a permanent site must be designed around plume ejecta from its own traffic.

Established And within the dust problem, radiators are worse off than arrays. Ground testing reported by NASA found that only 11% areal coverage doubles the solar absorptance of a thermal control surface. Measured optical degradation is severe: AZ-93 white paint's solar absorptance rises by a factor of 2.2 to 3.0 with emissivity falling about 10%; silvered fluorinated ethylene propylene second-surface mirrors rise 1.7 to 4.2× in absorptance; the absorptance-to-emissivity ratio degrades by 1.5 to 3.4× at full monolayer coverage. The Apollo record confirms the consequence: Apollo 12's magnetometer ran about 68 °F above expectation at five internal locations because of dust on thermal control surfaces, and Apollo 15's rover batteries ran 68 to 78 °F high from dust on their radiators, with Apollo 16 repeating it. And the dust does not blow off: fine dust under 34 µm at 4.5% coverage had an absorptance of 0.245, and fifteen seconds of 10 psi nitrogen reduced it by two percent, to 0.239. The fission case's own Achilles heel is the same dust usually cited against solar.

Established The fourth is plutonium-238 production, which caps the one technology that demonstrably works. An MMRTG contains 4.8 kg of plutonium oxide at 71% plutonium-238 and produces about 110 We. Production restarted at Oak Ridge with 50 g in December 2015 after a roughly thirty-year gap; the largest shipment in over a decade was about 550 g in mid-2023; the goal is an average of 1.5 kg per year by 2026, which is one unit every 3.2 years. The misconceptions section works the consequence for a base-scale load. Whether the goal rate has been achieved is unverified.

Frontier The fifth is institutional rather than physical, and on current evidence it binds harder than any of the above. The programme of record has no provider award as of August 2026 against a directed date of roughly April 2026; is staffed at a cap of 15 full-time-equivalent engineers; was moved from the Space Technology Mission Directorate to the Exploration Systems Development Mission Directorate with the former directed to cease competing efforts; and publishes two mutually inconsistent power levels and two delivery dates three to six years apart on its own websites. The directive itself acknowledges “over $200 million invested since 2000 with no flight-ready system achieved” — a programme indicting its own predecessor in its founding document.

Established The sixth is distribution, where the most advanced hardware in existence moves one kilowatt. Astrobotic's LunaGrid-Lite, funded by a $34.6 million NASA Tipping Point award in 2023, is a 4U CubeRover that pays out 500 m of ultra-light cable to deliver 1 kW. It passed Critical Design Review in August 2025 with flight component fabrication under way. It is the most advanced lunar power-distribution hardware anywhere and its rating is one kilowatt over half a kilometre. There is no lunar equivalent of a plug, a bus voltage set by a standards body, or a grid code.

Established What is not a bottleneck, stated plainly. Reactor physics is not: KRUSTY demonstrated startup, ramp, steady operation, failure response and shutdown in 2018 at better than the required conversion efficiency and better than eight times the required turn-down. Photovoltaic efficiency is not: 30% conversion is assumed in the current farm literature and the polar illumination fraction, not the cell, is what limits yield. Battery chemistry is not the binding term either — a doubling of specific energy would not close a fifteenfold mass gap at the equator, though it would matter at a polar site where the gap is fourfold. The bottleneck is that nobody has decided what to build, at which site, for which duty cycle, and then landed it.

Established This brief carries a typed dependency edge and it is a genuine one. Lunar energy infrastructure depends on lunar industry: that brief establishes the surface presence this one powers, and it explicitly names night survival as its own bottleneck. The relation is stated in the other brief's own text, which is the standard this corpus sets for typing an edge rather than asserting adjacency.

Established Beyond that the dependencies are industrial and institutional rather than scientific, and there are five. Landed cost per kilogram, currently $0.4 million to $1.2 million, which sets the entire trade. A heavy-class lander of about fifteen tonnes, which the fission directive assumes and which does not exist — the largest cargo capacity in near-term prospect is Blue Moon MK1's 3,000 kg, and it has not flown. Plutonium-238 production, a national programme running at a rate this brief could not verify against a goal of 1.5 kg a year. Low-enriched uranium at around 19.75% assay for the reactor itself, on a supply chain that analysts identify as a critical-path risk for the adjacent Space Reactor-1 programme. And Presidential launch authorisation, which is a political dependency with no technical fix and no published schedule.

Established What depends on this is the more consequential direction. Continuous surface power is the precondition for resource extraction, propellant production and manufacturing at anything beyond demonstration scale — moon-based manufacturing and space resource economies both assume it generically. No typed enabling edge is claimed to either, because they assume power in the abstract rather than depending on this particular route to it. Helium-3 extraction is the extreme case of the same relation: the reference mining concept needs 12.3 MW of thermal input plus 350 kW electrical, which is two to three orders of magnitude beyond any lunar power system that has been designed, let alone built.

Frontier One dependency runs the wrong way round and is worth naming. The strongest candidate for a lunar power interface specification is a NASA procurement requirement — 120 VDC, ten ports at 6 kWe or more — which means every future surface system's interoperability depends on the outcome of a single unresolved acquisition. If the award slips again, the de facto standard slips with it.

Established The most informative experiment in this subject is not an experiment anyone designed. It is the lander survival record, and it is unusually clean because the outcomes are binary, public, and now numerous enough to have a base rate.

Established The base rate that falls out of it: of Western commercial landers that reached the surface intact, none has survived a lunar night, and every vehicle that has survived one carried isotopes or did it by accident. Two entries in the record could not be confirmed from a primary source for this brief and are named rather than tabulated: Intuitive Machines' IM-2 Athena, reported to have landed sideways in a south-polar crater with the mission declared over on 7 March 2025, and ispace's HAKUTO-R M1 and M2, both lost during landing. Trade coverage exists for both; no company or agency statement was retrieved.

Established The decisive ground experiment already ran, in 2018, and nothing has superseded it. KRUSTY's four test phases over five months, culminating in the 28-hour full-power test of 20–21 March 2018 with simulated engine and heat-pipe failures, established that the physics and the control behaviour work at kilowatt scale with power conversion attached. It remains the only space fission reactor the United States has taken critical since SNAP-10A in 1965, against a cumulative national space-nuclear spend that the NASA Administrator has put at over $20 billion since the 1960s for one flight.

Established The one in-situ power-technology result of the current era is qualitative. Blue Ghost Mission 1 carried a 400 W solar array, ran 346 h of daylight operations, and flew NASA's Electrodynamic Dust Shield, which removed regolith from glass and thermal radiator surfaces. That is the entire in-situ 2025–26 lunar power-technology dataset, and its headline result has no number attached to it.

Frontier The experiment that would settle the central question is a published trade study, and it is cheap. Nothing needs to be launched. What is needed is a site-specific worst-case darkness duration for a named candidate ridge, a stated survival-power duty cycle, and a landed-mass comparison between vertical solar plus storage and a 40-to-100 kWe reactor, at that site, over ten years, with the launch-approval and thermal-survival costs on both sides. Four separate government and peer-reviewed documents have declined to perform exactly this comparison. It would cost a fraction of one month of the programme's FY2026 budget.

Frontier The experiments nobody has run in flight, in order of what they would settle. A regenerative fuel cell through a real lunar night at either site type — the sizing work exists and the hardware does not, so the most promising non-nuclear route for long nights is entirely unvalidated in the environment that matters. Buried regolith thermal mass, which the modelling says beats lithium-ion on stored energy per kilogram and costs nothing to land. A survival-mode architecture designed around the one-to-five nominal-to-survival ratio rather than around full-power operation. And a quantified dust-mitigation result on a real radiator after a real plume event.

Established One national programme is now running the thermal experiment deliberately. The Indian Space Research Organisation and the Department of Atomic Energy are jointly developing artificial heating intended to extend lander life from about fourteen days to 100 to 200 days, targeting Chandrayaan-4 at around 2028 on a budget above ₹2,100 crore. Whether the heating will be isotopic is not stated in the reporting consulted. The date is a programme intention and is treated as such.

Established Negative results worth recording, because they are the field's best data. Chandrayaan-3 died for want of heat, not power. IM-1 died with its mission complete and said so honestly. Blue Ghost died five hours into a night it had 346 hours of sunlight to prepare for. Each is a cleaner result than any surviving mission, because each isolates thermal survival from every other variable, and together they are the reason this brief treats thermal as the binding constraint.

Established Start with the illumination, because the polar solar case is an argument about height above the ground and nothing else. Illuminated area at the south pole scales steeply with elevation, and the scaling is the reason vertical arrays exist as a technology programme. In a 50 × 50 km box at the south pole, the area receiving more than 70% illumination is a fraction of a square kilometre at ground level and tens of square kilometres at a hundred metres up.

Established Only above roughly 100 m of elevation is more than two square kilometres illuminated for over 90% of the time. The Shackleton–de Gerlache connecting ridge reaches a maximum average illumination of 88% at two metres above the ground, averaged over twenty years, with slopes mostly under 10° along the ridge and under 20° down it, and within about 10 km of several small permanently shadowed regions. An independent peer-reviewed treatment gives Shackleton crater an illumination fraction near 80%, about 95% at 89° latitude, and an annual yield of 10,988 kWh/m²/yr for a vertical tracker at 89° against 3,822 kWh/m²/yr for a fixed system. The same work notes that the Sun sits so near the horizon that it is partially occluded 4.3 to 7.8% of the time, which erodes real capture below the headline illumination fraction. None of these figures transfers to an equatorial site.

Established Vertical solar array technology is the engineering response, and it is a structures problem as much as a photovoltaic one. NASA's programme specifies deployment heights up to 32 feet, about 10 m. The 2026 peer-reviewed farm study models 5 × 10 m panels with the bottom edge 5 m up and the top at 15 m, at 30% conversion efficiency, giving 15 kW per panel at full sun, arranged as a hundred-panel farm on 20 m grid spacing for a notional 2030 site. Such a mast must deploy autonomously, stand in one-sixth gravity on slopes under 20°, track a sun that circles the horizon rather than crossing the sky, survive thermal cycling between roughly +22 °C and −220 °C, and shed or tolerate dust with no maintenance crew. Two of the authors of that farm study are employed by Shell Global Solutions, which is disclosed here because the paper is peer-reviewed but written by an interested party in surface energy systems.

Established Storage, worked with its assumptions exposed. ESA gives 200 Wh/kg as the current specific energy of space-capable lithium-ion. At cell level and 100% depth of discharge, 3,540 kWh needs 17,700 kg; at 80% usable depth of discharge, 22,100 kg; with a 0.7 cell-to-pack fraction covering structure, harness, battery management and the survival heaters the cells themselves need, about 31,600 kg. The defensible band is 18 to 32 tonnes of battery to carry 10 kW through one equatorial night. The specific-energy input is established; the pack-level derating is engineering judgement and is flagged frontier. Note two framings that follow. 17.7 tonnes exceeds the fifteen-tonne heavy-class lander the fission directive itself assumes as the upper bound of available landed mass, and it is 5.9 times the entire 3,000 kg cargo capacity of Blue Moon MK1, and 74 to 177 times a single Commercial Lunar Payload Services lander's payload.

Handwave Any battery figure quoted at 500 Wh/kg. NASA's battery research briefing lists a lunar rover target of over 500 Wh/kg and a cold-rechargeable target of over 150 Wh/kg at C/5. Those are stated goals. That they are goals is established; citing either as an available lunar-qualified product is handwave, and the briefing lists no achieved 200-plus Wh/kg lunar-qualified pack at all.

Established Fission, with the demonstrated numbers rather than the design-family ones. KRUSTY produced a measured 4 kWt at 800 °C, driving Stirling convertors at better than 35% conversion efficiency against a requirement of 25%, with better than 16:1 turn-down against a requirement of 2:1, reaching 800 °C in 1.5 h against a 3 h requirement. The core was solid cast uranium-235 “about the size of a paper towel roll,” with passive sodium heat pipes and Stirling conversion. The reference masses from the same charts are 400 kg for 1 kWe and 1,500 kg for 10 kWe. NASA's higher-power concepts work puts the thermal requirement for 10 kWe at about 43 kWt, which is the number that sizes the radiators.

Established Heat rejection is the dominant mass term at high power and the least discussed part of the problem. There is no convection, so waste heat leaves only by radiation, and radiated power scales with the fourth power of surface temperature. A 100 kWe closed-Brayton system rejects roughly 400 kWt through radiators sized on a clean ratio of solar absorptance to infrared emissivity. That ratio is exactly what dust destroys, which is why the dust problem bites the fission case at least as hard as the solar one — and unlike an array, a radiator cannot be tilted vertical to shed dust, because its whole function is to face cold sky.

Established The governing requirement set, for reference. The 2021 request for proposals required 40 kWe continuous for at least ten years, under 6,000 kg, stowed within a 4 m × 6 m cylinder, low-enriched uranium, deployed by 2030. NASA's 2022 requirements paper confirms 40 kWe at 120 Vdc, the 6,000 kg limit and a dose limit of 5 rem/yr at 1 km. Both government reference designs came in around 7 tonnes, already over the cap, before the requirement was raised to 100 kWe on a 15 t lander assumption.

Established Landed cost per kilogram, derived rather than quoted, because there is no published tariff. It has to be inferred from task order value divided by lander payload capacity. Intuitive Machines' Nova-C IM-1 flew on a task order that grew from $77 million to $118 million against a surface payload capacity of 100 kg — about $1.18 million per kilogram. Firefly's Blue Ghost Mission 1 was a $93.3 million task order of February 2021 against a stated capacity of up to 240 kg — about $389,000 per kilogram. The working band used throughout this brief is therefore $0.4 million to $1.2 million per kilogram landed, 2026, and it is an optimistic one: the divisor is lander capacity, not delivered mass, because NASA's Commercial Lunar Payload Services page publishes no payload masses.

Lunar industry is the nearest neighbour and the relation is a dependency rather than an adjacency: it names the bottleneck and this brief opens it. The division of labour is that mining, in-situ resource use and manufacturing economics stay there, and everything measured in watts, kilowatt-hours, landed kilograms and kelvin stays here.

Small modular reactors share a physics with lunar fission and share no economics at all. Terrestrially the question is levelised cost against a grid that already exists; on the surface there is no grid to undercut and the only currency is landed mass, which is why a design that would be commercially hopeless on Earth can be the obvious answer at a lunar site. Reading the two together is the cleanest way to see how completely context rather than technology determines whether small fission makes sense.

Energy storage revolutions is adjacent in the strong sense that the mass penalty of storage is the entire argument here, amplified by a factor of about a million in delivery cost. A specific-energy improvement that would be a footnote terrestrially moves billions of dollars of landed mass on the Moon — and, at a polar site with a fourfold gap rather than a fifteenfold one, could plausibly change which technology wins.

Commercial fusion is adjacent only through helium-3, and the relationship is almost entirely negative: the lunar resource is a fuel for a reaction that is substantially harder than the one no reactor has yet made commercially net-positive. Moon-based manufacturing and space resource economies are the prospective loads that would justify tens or hundreds of kilowatts, and neither has a firm demand figure this brief could use.

Research integrity is adjacent in a quieter way. This is a field where vendor product pages, agency press releases and programme roadmaps say materially different things about the same hardware, and where the most-cited numbers are goals. The ethical section treats source standing as substance rather than as an aside.

Established The defining institutional fact is that there is exactly one buyer. Every figure in this brief traces to a NASA requirement, a NASA task order, a NASA technology award or a company positioning itself for one. There is no commercial demand for lunar surface power that is not derived from a government exploration programme, no second national programme buying at volume in the West, and therefore no market discipline on price, no second specification for vendors to design against, and no reason for anyone to qualify hardware NASA has not asked for. When the single buyer's requirement moves — and it has moved three times in five years — the entire supplier base moves with it or stops.

Established The pattern of that single buyer is a programme accelerating its target faster than it builds hardware. A 2018 ground demonstration at roughly 1 kWe. Design studies at 40 kWe in 2022 at about $5 million each. A directive in August 2025 raising the requirement to at least 100 kWe on a mandatory closed Brayton cycle with launch readiness in Q1 FY2030, procured under Space Act authority for “maximum flexibility to industry,” with two providers, a down-select option at preliminary design review and a minimum of 25% of payment withheld until successful checkout and delivery. The programme was moved from the Space Technology Mission Directorate to the Exploration Systems Development Mission Directorate, with the former directed to cease competing efforts, and staffed at a cap of 15 full-time-equivalent engineers. Each step is real. None of them is a reactor.

Established The institution's own public record does not agree with itself, and that is a governance finding rather than a clerical one. NASA's programme page said “at least 40 kilowatts” as of September 2025; Idaho National Laboratory's said 40 kWe with news items dated August 2026; the directive and NASA Glenn's own releases say at least 100 kWe; and NASA's Moon Base roadmap places delivery in 2033–2036 rather than FY2030. A programme whose four official public artefacts carry two power levels and two dates a year after its governing directive is a programme without an internal owner of its own external communication, and a reader outside it cannot tell which number is operative.

Established Above NASA sits a policy layer that has recently become directive. An executive order of 18 December 2025 produced a White House National Security and Technology Memorandum on 14 April 2026 requiring NASA to field a lunar fission variant ready for launch by 2030 and the Department of Defense a mid-power reactor by 2031. A NASA–Department of Energy memorandum of understanding was signed 14 January 2026. This is the first period since the 1960s in which lunar surface power has had explicit White House-level direction, and the cumulative record it inherits is over $20 billion of space-nuclear spending for one flight.

Established Commercial lunar delivery has changed the cost basis and not the physics. Landed mass remains the currency. The landers flying are small — every currently active vehicle below 500 kg of surface capacity against a four-tonne programme aspiration — with power systems sized for a single lunar day and no night-survival capability stated for any of the four vehicles manifested for 2026. That is a rational commercial choice given who pays and what a task order buys. Its consequence is that the surviving-the-night problem is being deferred by the current flight rate rather than solved by it, and that the operational dataset the whole field needs is not being generated.

Frontier The institution that does not exist is a standards body, and its absence is cheap to fix now and expensive later. LunaNet, the one real interoperability specification, covers communications, navigation and information services and does not touch power. The Lunar Surface Innovation Consortium runs a surface power interoperability working group with monthly meetings and publishes no quantitative targets, standards or dates. DARPA's lunar economy study named aggregated thermal generation and rejection “as a service” as an emerging concept and published no power figures, no grid architecture and no cost per kilowatt-hour. What is filling the vacuum is a procurement document: 120 VDC with ten user ports rated at 6 kWe or more, from an unawarded partnership announcement. A bus voltage established by one agency's unawarded contract is not a standard, and every future connection on the surface currently depends on how that acquisition resolves.

Established Other actors, for completeness, since the field is usually described as though only NASA existed. China and Russia plan an International Lunar Research Station reactor in 2033–2035. India's space agency and its Department of Atomic Energy are jointly developing lander heating to extend surface life from about fourteen days to 100–200 days for Chandrayaan-4 around 2028. And China is the only actor with a demonstrated, repeated capability to survive lunar nights at all, using radioisotope heater units on both landers and rovers — a capability the West's operational record does not yet contain.

Frontier Launching fission material carries a distinctive risk profile at exactly one moment: ascent. The material is comparatively benign before criticality, which is why designs start cold and are brought critical only on the surface, and that is the correct engineering answer to the public concern. It should be stated plainly rather than avoided. The requirement set includes a dose limit of 5 rem/yr at 1 km from the deployed reactor, which is a real constraint on siting near a crewed habitat and on any future traffic through the area. Presidential launch authorisation exists as a check on exactly this, and analysts treat it as a live schedule risk rather than a formality.

Established Public money is being spent at scale on a comparison that has not been published, and that is the clearest accountability failure in this subject. The programme is budgeted at $350 million in FY2026 rising to $500 million a year from FY2027. Four documents that could have carried the justifying trade study do not: NASA Glenn's regenerative fuel cell sizing paper explicitly declines to compare against fission; NASA's own fission requirements paper contains no comparison against solar plus storage; NASA's energy storage paper is methodological with no baselines; and the 2026 peer-reviewed vertical-array study explicitly places storage out of scope. Each choice is defensible in isolation. The aggregate is that half a billion dollars a year is being appropriated against a technology choice whose supporting analysis is not in the public record.

Established Evidence quality in this field is unusually uneven, and this brief marks interested parties throughout. A substantial share of what is publicly known about lunar power hardware comes from vendors describing their own products: Astrobotic on LunaGrid-Lite, Firefly on Blue Ghost, Intuitive Machines on Odysseus and on power beaming, Lockheed Martin on its own fission contract. Those sources are reliable on facts they would be embarrassed to get wrong — a mission end date, a cable length, a payload capacity — and unreliable on framing. Firefly's product page listing “Lunar day + night” against a flown outturn of five hours is the cleanest example available of why the distinction matters. Two authors of the most detailed peer-reviewed vertical-array farm study are employed by an oil major; the work appears sound and the affiliation is disclosed here rather than discovered later.

Frontier The plutonium-238 question is an allocation question that is rarely posed as one. Production is below its own goal, the current rate is unverifiable from public documents, and every gram committed to a heater on a lander is a gram unavailable to an outer-planets mission that has no alternative power source at all. Radioisotope heat is what has actually let vehicles survive lunar nights, which makes the trade real and immediate rather than hypothetical. It is being made continuously and is almost never discussed as a trade.

Established This brief's own unresolved questions, stated as an obligation rather than buried. It does not know whether any fission surface power provider award has been made, because the contracting record could not be read and only the absence of announcement is established. It does not know the current annual plutonium-238 production rate, nor the total United States inventory available to NASA; the reports that would settle the latter are from 2017 and were not consulted. It does not know the worst-case continuous darkness at the Shackleton Connecting Ridge, which is the single most load-bearing missing number in the storage-versus-fission comparison. It does not know how well the Electrodynamic Dust Shield performed in flight, because no removal percentage has been published. It has no measured array or radiator degradation from any post-Apollo lander — not from Chang'e, Chandrayaan, SLIM, Intuitive Machines or Blue Ghost — so the entire measured degradation base for this brief is Apollo-era. It does not know the thermal output in watts of the Chang'e radioisotope heater units, nor whether Chang'e-4 carries a generator as well as heaters; the paper that would settle it is paywalled. It does not know the outcome of the vertical solar array Phase 2 selection. And it does not resolve the helium-3 concentration discrepancy between roughly 30 µg/g and 11.8 ppb. Several relevant sources were unreachable rather than absent — SpaceNews reporting on the NASA–DOE collaboration and on the Astrobotic award, an Acta Astronautica paper on night-survival storage selection, an Applied Energy characterisation of the Chang'e-4 radioisotope generator, a 2024 Philosophical Transactions paper on the lunar dust environment, and a Department of Energy history of the plutonium-238 restart — and are named here so a reader can go where this brief could not.

Frontier A permanent presence anywhere off Earth is a power problem before it is anything else. Habitats, resource extraction, propellant production and manufacturing all reduce to kilowatts delivered continuously through a darkness measured in days, at a site where a kilogram of anything costs between $0.4 million and $1.2 million to deliver and there is no repair crew. The Moon is the first place where humanity has to build an energy system rather than tap one, and the discipline it imposes — state the site, state the duty cycle, state the thermal boundary condition, then count kilograms — is the discipline that any off-Earth infrastructure will need.

Established The general principle this case illustrates is that the constraint was never the physics. Fission surface power was demonstrated on the ground in 2018 with better-than-required conversion efficiency and eight times the required turn-down. Photovoltaics at 30% conversion are conventional. Lithium-ion at 200 Wh/kg is off the shelf. Every technical ingredient of a working lunar power system has existed for years. What has not existed is a decision about which site, which duty cycle and which architecture, followed by the appropriation and the lander to execute it. This is not a technology awaiting a discovery. It is one awaiting a decision, and the decision is waiting on a trade study nobody has published.

Established A sharper version of the same principle, visible in the spending record. Over $20 billion of United States space-nuclear expenditure since the 1960s has produced one flight, in 1965. Over $200 million since 2000 produced no flight-ready system, a fact conceded in the founding paragraph of the directive that superseded that effort. A programme now budgeted at half a billion dollars a year has moved its own requirement three times in five years — up in power, up in mass, out in date — while staffing fifteen engineers and publishing two contradictory power levels on its own websites. That pattern is not unique to space nuclear power and it is not a story about incompetence; it is what happens when a programme's requirement is set by strategic signalling rather than by a published mission need.

Speculative What would actually change at civilisational scale. Tens of kilowatts continuously available at a polar site is the threshold at which resource extraction stops being a demonstration and starts being an industry, and it is a threshold nothing on the current manifest reaches. Beyond it lies propellant production, which changes the cost of everything further out, and beyond that the surface manufacturing that would make lunar activity self-sustaining rather than resupplied. Each step is individually plausible. The conjunction on the schedules currently published is speculative, and this brief declines to attach a date to any of it.

Established What already happened, because the timeline in this subject is usually told as though it started in 2025. SNAP-10A flew in 1965 and remains the United States' only flown space fission reactor. KRUSTY ran November 2017 to March 2018 with a 28-hour full-power test. NASA's vertical solar array awards were made in March 2021 — five companies at up to $700,000 each for twelve-month fixed-price design contracts, with a Phase 2 of up to $7.5 million for up to two companies. The 40 kWe design contracts, three at about $5 million each, were awarded 21 June 2022, with Phase 1A extensions in January 2025. Astrobotic's $34.6 million Tipping Point award for LunaGrid-Lite came in 2023. Blue Ghost Mission 1 landed and died in March 2025.

Established Then the 2025–26 acquisition trail, which is the substance of the current programme. Directive signed 4 August 2025. Request for Information issued 14 August 2025 with responses due 21 August. First draft Announcement for Partnership Proposals 29 August 2025; industry day at Glenn 9 September; Program Executive named 25 September; second draft announcement 5 December 2025 with the final promised “early 2026.” NASA–DOE memorandum of understanding 14 January 2026. A White House National Security and Technology Memorandum on 14 April 2026, mandated by an executive order of 18 December 2025, directing NASA to field a lunar fission variant ready for launch by 2030 and the Department of Defense a mid-power reactor by 2031. A separate Lunar Enabling Infrastructure Accelerator draft solicitation open in July 2026. Every item on that list is a document. None is hardware.

Handwave Q1 FY2030 for a 100 kWe reactor ready to launch. The directive's own intermediate milestones have already slipped: awards were directed within six months of an RFP that was directed within sixty days, and no provider award has been announced roughly ten months past that date. The requirement itself has moved three times — 10-plus kW at 2 to 3.5 t by end-2026 in 2020; 40 kWe at 6 t by end of decade in November 2021; 100 kWe at 15 t by Q1 FY2030 in August 2025 — while the mass of the government's own reference designs sat at about 7 t, over the then-current cap. A schedule that has been re-baselined upward three times without hardware is an intention.

Handwave 2033–2036 for the same system, per NASA's own Moon Base plan. This brief reports both dates because both are official and neither has been withdrawn, and it does not resolve them. The same plan gives Phase Two, 2029–2032, only “hundreds of watts” of nuclear surface power from radioisotope generators — which, if correct, means the fission programme delivers nothing usable during the first Artemis surface campaign.

Handwave Near-term hardware dates, all of them intentions. LunaGrid-Lite flight readiness was targeted for Q2 2026 as of August 2025; whether it has flown could not be verified. Blue Moon MK1 was targeting launch before end-2026 and was disrupted by the New Glenn hotfire anomaly of 28 May 2026. Blue Ghost Mission 2, IM-3 and Griffin-1 were manifested for 2026, and the survey that lists them notes that no lunar-night survival capability is stated for any of the four 2026 landers. Space Reactor-1 “Freedom,” a nuclear-electric-propulsion spacecraft above 20 kWe announced 23 March 2026 for a December 2028 launch, had no publicly announced prime integrator as of 16 April 2026, and analyst assessment judges a 12-to-18-month slip plausible on integration, enriched-uranium delivery or Presidential launch authorisation.

Speculative 2033–2035 for a Chinese and Russian lunar reactor under the International Lunar Research Station. Recorded because it is the only competitor date on the board, and recorded with the observation that it is later than NASA's stated 2030 — so the “beat China” framing common in 2025 press coverage is not supported by the dates in that same coverage.

• Depends on A surface presence to power, which is lunar industry's subject. This edge is typed because that brief names night power and thermal survival as its own bottleneck in its own text, so the dependency is stated rather than inferred from adjacency. Nothing else this brief depends on is a scientific result: the reactor physics was demonstrated in 2018, the photovoltaics are conventional, and the storage chemistry is off the shelf. What is missing is landed mass, a decision, and a built article.
• Requires (not on this map) Launch and landing cost per kilogram to the surface, currently $0.4 million to $1.2 million and improving more slowly than programme aspirations assume. A heavy-class lander near fifteen tonnes, which the governing fission directive assumes and which does not exist — the largest near-term cargo capacity is 3,000 kg and has not flown. Plutonium-238 production, a national programme whose current annual rate this brief could not verify from any document. Low-enriched uranium supply at reactor assay, identified by analysts as a critical-path risk on the adjacent nuclear-propulsion programme. Presidential launch authorisation, a political dependency with no technical substitute. A lunar power interface standard, which currently exists only as an unawarded procurement requirement at 120 VDC. And a buyer other than NASA, of which there is none. Six of those seven are industrial, financial or institutional rather than technical, which is the shape of this whole subject.
• Enables Resource extraction, propellant production and surface manufacturing at beyond-demonstration scale. No typed enabling edge is claimed, because those briefs assume power generically rather than depending on this particular route to it — and because the honest position is that nothing downstream is currently waiting on this brief, since nothing downstream has been built either.
• Adjacent Small modular reactors, the same physics of small fission against a completely different economics, since on the Moon there is no grid to be cheaper than; energy storage revolutions, where the mass penalty of stored energy is the same argument in a place where a kilogram costs a thousand times more; moon-based manufacturing, the largest prospective load; space resource economies; and commercial fusion, which is where the helium-3 argument would have to be settled and is not.

“The lunar night is fourteen days, so you need two weeks of storage.” Established Correct for equatorial and mid-latitude sites, and correct as the worst case. Wrong as a design point for Artemis. NASA gives three to four days of maximum continuous darkness at 10 m elevation near the poles, and NASA Glenn works polar eclipses at about 100 h against 350 h at the equator. Sizing a polar mission for 354 hours overstates the store by nearly fourfold, and the error favours the wrong technology. Run both numbers; the polar one is where the argument actually happens.

“KRUSTY demonstrated a 10-kilowatt reactor in 2018.” Handwave KRUSTY was a 4 kWt, roughly 1 kWe-class demonstration. NASA's own press release describing the system as “capable of providing up to 10 kilowatts of electrical power” refers to the design family, not to what was tested. The distinction matters because the 10 kWe reference mass of 1,500 kg, which carries the whole fission-versus-batteries comparison including this brief's, is a design figure resting on a 1 kWe-class test.

“NASA awarded contracts in 2022 to build a lunar reactor.” Established It awarded three twelve-month design studies at approximately $5 million each — Lockheed Martin with BWXT and Creare; Westinghouse with Aerojet Rocketdyne; and IX, the Intuitive Machines and X-energy joint venture, with Maxar and Boeing. Base-scope awards were capped at $5 million by the governing request for proposals. These were paper studies, extended in January 2025 by Phase 1A awards, one of which added a ground testbed for a power conversion system. Nothing in that sequence is a reactor.

“A 100-kilowatt reactor is being built for 2030.” Handwave A directive requiring one exists and is funded; as of August 2026 no provider award, no test article and no flight hardware exists, only a Request for Information and two draft announcements against a directed award date of roughly April 2026. Meanwhile NASA's own programme page and Idaho National Laboratory's both still say 40 kWe — the latter with news dated 6 August 2026 — and NASA's Moon Base roadmap places delivery in Phase Three, 2033–2036. This brief reports all three positions and resolves none. It also records that it could not read the SAM.gov contracting record, so the missing award is an absence of announcement rather than a confirmed negative.

“Radioisotope generators are how you power a lunar base through the night.” Handwave They are how you keep one alive, not how you power one. A 10 kW night load would need about 91 MMRTGs, roughly 437 kg of plutonium-238, or about 290 years of production at the 1.5 kg-per-year goal rate; the arithmetic is this brief's on sourced inputs. Radioisotope power is structurally a watts-to-hundreds-of-watts technology, which is exactly what NASA's own Moon Base Phase Two assumes: “hundreds of watts” from radioisotope thermal generators between 2029 and 2032. The one real lever is conversion — Stirling radioisotope generators yield roughly four times the electrical output per gram of plutonium-238 that thermoelectric conversion does, and NASA's 2026 technology solicitation names them as a target area.

“Plutonium-238 production is running at about half a kilogram a year.” Frontier Widely repeated, plausible, and this brief could not source it. What is confirmed is the goal of 1.5 kg per year by 2026, an initial capability target of 300 to 400 g a year, and a single shipment of about 550 g in mid-2023 described as the largest in over a decade. No Department of Energy, NASA or Oak Ridge document consulted states a 2025 or 2026 annual outturn, and none gives the total United States inventory available to NASA. The honest formulation is “on the order of half a kilogram a year, not confirmed for 2025–26.”

“NASA has recently funded vertical solar arrays.” Established The awards were March 2021 — five companies at up to $700,000 each, with a Phase 2 of up to $7.5 million for up to two, deployment heights up to 32 feet, target “end of this decade.” There has been no 2025–26 round, and the Phase 2 outcome could not be established at all. As of June 2026 NASA is still soliciting vertical array maturation to TRL 5–6 by ground testing, five years on. The live hardware line is Astrobotic's, funded not by that programme but by a 2023 $34.6 million Tipping Point award.

“LunaNet is the lunar interoperability standard.” Established LunaNet Version 5, published 7 February 2025 jointly with ESA and JAXA, covers communications, position-navigation-timing, and information services such as space weather. It does not address electrical power interfaces at all, and no lunar power interoperability standard exists. The institutional section sets out what is filling the vacuum instead.

“Landers routinely survive the lunar night now.” Established Of Western commercial landers that reached the surface intact, none has. Three refinements to the usual list. SLIM survived three nights, not one, and JAXA states it was not designed to. Blue Ghost Mission 1 lasted just over five hours into the night, which is a more damning number than “did not survive” because it had a complete 346-hour charging period immediately beforehand. And Blue Moon MK1 has not flown as of August 2026, so it does not belong in an outcomes list in either column.

“This lander is rated for lunar day and night operations.” Handwave Firefly's Blue Ghost product page lists “Surface Operations: Lunar day + night” as a capability. The flown outturn was five hours. Firefly does qualify elsewhere that night operations are a customisation option, so the claim is not false so much as unqualified in the place a customer reads it. The product-page claim is handwave; the five hours is established. This brief reports the disagreement between a vendor's specification sheet and its own mission page rather than choosing between them.

“Helium-3 makes the Moon an energy source.” Handwave Three separate problems, each sufficient. First, concentration: Apollo 11 regolith samples measured 9.22 to 17.9 ppb of helium-3, mean 11.8 ppb, with about 88% of the helium in sample 10084 held in particles under 100 µm. Obtaining 33 kg of helium-3 a year — one 400 MW plant's supply — requires excavating nearly 5 million tonnes of regolith at 10 ppb, about 150,000 tonnes of regolith per kilogram. Second, energy: the reference Mark III miner excavates 1,258 t/hour, covering one square kilometre to 3 m depth per year over 3,942 operating hours, and consumes about 350 kW electrical plus a thermal system needing 12.3 MW of solar input with 85% heat recovery, or 82 MW without. That is two to three orders of magnitude beyond any lunar power system that has been designed. Helium-3 is not a lunar energy source; at present it is a lunar energy sink of extraordinary size. Third, the reaction: deuterium–helium-3 needs an optimal temperature near 60 keV against about 13 keV for deuterium–tritium, and a confinement figure of merit 24 times higher, or 61 times higher for the 15% D / 85% He-3 mixture needed to suppress neutron-producing side reactions. A machine reaching Q > 10 on deuterium–tritium would “hardly achieve a gain of 1” on deuterium–helium-3, and the authors conclude the fuel “would only become a potential technology for a second or third generation of fusion reactors.” There is no working deuterium–helium-3 reactor anywhere.

“Helium-3 is present at about 30 micrograms per gram.” Frontier A peer-reviewed 2021 analysis states an abundance of about 30 µg/g, which is roughly 2,500 times the Apollo-sample figure. This brief could not reconcile the two: 30 µg/g is the usual figure for total helium in mature mare regolith, while 11.8 ppb is what the Apollo samples measured for helium-3 specifically. Use the parts-per-billion figure, and note that the discrepancy is reported here rather than resolved.

“The real helium-3 business is fusion fuel.” Established The actual customers of the one funded lunar helium-3 company are cryogenics, quantum computing, a US Air Force research contract and the Department of Energy — dilution refrigerators and neutron detectors, where terrestrial helium-3 is scarce because it comes from tritium decay in weapons stockpiles and sells accordingly. The company raised $5 million on a simple agreement in January 2026 on top of an $18 million seed, and its first lunar mission is a multispectral prospecting camera. That is a real, small, non-fusion market at kilogram scale, and even its own investors are funding cameras rather than miners.

“Dust will gradually bury the solar arrays.” Established Two errors in one sentence. Natural deposition is about 100 µg/cm²/yr, roughly a millimetre per thousand years, which is negligible on mission timescales; the threat is episodic and anthropogenic, and a single Apollo 11 ascent cost about 17% of output on a cell with 0.5 mm cover glass. And the worse victim is not the array but the radiator, which is to say the fission case: 11% areal coverage doubles solar absorptance, the absorptance-to-emissivity ratio degrades 1.5 to 3.4× at full monolayer coverage, and a 100 kWe closed-Brayton reactor rejects roughly 400 kWt through radiators sized on a clean ratio. An array can be tilted vertical to shed dust; a radiator cannot, because its function is to face cold sky. The dust argument usually deployed against solar cuts harder against fission.

“The United States is racing China to a lunar reactor.” Established China and Russia plan an International Lunar Research Station reactor between 2033 and 2035 — later than NASA's stated 2030 and roughly contemporaneous with NASA's own Moon Base Phase Three. The urgency framing common in 2025 press coverage is not supported by the dates carried in that same coverage.