The lunar south pole is the hardest driving environment humans have ever faced: powdered, abrasive regolith; slopes strewn with ejecta blocks; temperature swings from +120 °C in sunlight to −170 °C in shadow; a 14-day night; and no atmosphere to cool a motor or cushion an impact. Prometheus is designed from first principles for exactly that — and to NASA's Lunar Terrain Vehicle reference requirements: two crew, 8-hour sorties, 20 km minimum range per charge, survival through the lunar night, and ten years of crewed and remote operations.
Every subsystem below was downselected through a weighted trade study — the same discipline we applied to Cosmos. Where flight-proven hardware exists, we baseline it. Where it doesn't, we develop it.
Electric is not just the best answer — on the Moon it is effectively the only answer. No atmosphere rules out combustion; consumable-based systems trade poorly against a 10-year service life. The real question is the drive topology. Six independent in-wheel motors give per-wheel torque vectoring, which is decisive on slopes and in soft regolith: slip is detected and corrected at each contact patch individually, the way Curiosity and Perseverance do it on Mars.
| Architecture | Traction control | Mass & complexity | Fault tolerance | Heritage | Score |
|---|---|---|---|---|---|
| Battery-electric, 6 × in-wheel PM motorsSelected | Per-wheel torque vectoring; regen braking; zero-radius turns via differential steer | No driveline, no gearbox tunnel; motor mass at wheels | Fail-operational: drives on 4 of 6 wheels | Mars rovers (6-wheel independent drive), Apollo LRV (4 hub motors) | 9.1 |
| Battery-electric, 2 central motors + driveshafts | Axle-level control only; mechanical differentials clog with dust | Sealed driveline in vacuum is a major dust/lubrication liability | Single motor loss strands an axle set | Terrestrial only | 5.8 |
| H₂/O₂ fuel cell + electric drive | Same drive benefits | Cryogenic storage, plumbing, consumable logistics per sortie | Good, but refueling infrastructure doesn't exist yet | Shuttle fuel cells (power, not traction) | 6.2 |
| Solar-direct drive | Power-limited on slopes; dies in shadow — the south pole is shadow | Lightest | None in darkness | Small robotic rovers | 3.4 |
Battery-electric with six independent in-wheel permanent-magnet motors, one per wheel, each with its own inverter channel. A deployable solar array recharges between sorties and trickle-heats the vehicle through the night. Regenerative braking recovers energy on descents — and doubles as the primary service brake, backed by a passive parking brake in each hub.
The honest answer to "who sells the best lunar in-wheel motor" is: nobody sells one. Every terrestrial drive unit — however brilliant — assumes air or oil cooling, elastomer seals and wet lubricants, all of which fail in vacuum at −170 °C. The best electromagnetics in the world live in EV drive units; the only vacuum-proven drivetrains live in space actuation houses. Prometheus takes one from each.
Supplier ground rule: US manufacturers only. There is no absolute NASA ban on foreign hardware, but ITAR export controls on a crewed lunar vehicle, Buy American provisions, and plain proposal optics all cut the same way. That rules out Mercedes (whose quad-motor electric G 580 is arguably today's best electric off-roader), Porsche, Land Rover, Slovenia's Elaphe and Chinese-owned Protean — noted, admired, excluded.
| US supplier | What they bring | Off-road / vectoring pedigree | Fit for Prometheus | Score |
|---|---|---|---|---|
| Rivian — quad-motor drive unitsSelected partner | One motor per wheel in production today; in-house inverters and drive software; US-built (Normal, IL) | Best US torque-vectoring stack in dirt — independent per-wheel slip control is their core product | License electromagnetics + vectoring software; we repackage for vacuum | 8.6 |
| Lucid — Sapphire-class drive units | Highest power density in the industry (~500 kW from a ~74 kg unit); already licenses tech to other OEMs | Track-focused, not off-road | Strong second — ideal electromagnetics, vectoring software would be ours to write | 7.9 |
| Tesla — Model S Plaid / Cybertruck units | Unmatched manufacturing scale, carbon-wrapped rotors | Tri-motor Cybertruck vectoring is capable | Vertically integrated; does not supply or license drive units to outside programs | 5.7 |
| GM — Ultium-derived lunar drive | The one US OEM already re-engineering EV drive for the Moon | Good (Hummer EV) | Exclusive to Lunar Outpost's competing LTV team | 5.5 |
| Moog / CAES-class space actuation (US) | Flight-proven BLDC actuators, dry-film (MoS₂) gearing, rad-tolerant drive electronics — decades of rover and spacecraft heritage | n/a — they qualify, not vector | The qualification backbone for whichever electromagnetics we choose | 8.2 |
Marry Rivian's electromagnetics and torque-vectoring software to Moog-class space qualification, inside our own sealed hub. Rotor/stator designs and per-wheel slip-control software licensed from the quad-motor program; bearings, dry-film-lubricated gearing, seals and conduction cooling re-engineered by a US space actuation house; packaged into six identical Void Orbital hub units with labyrinth dust seals. All-American bill of materials, one spare qualifies all six positions.
One battery pack feeds the whole vehicle — drive, avionics, life support and heaters — exactly as specified. The sizing driver is not range; it's surviving the 14-day lunar night, when the pack must keep itself and the avionics above minimum temperature for 336 hours with no sun.
| Chemistry | Pack-level energy | Cold behavior | Cycle life / 10 yr | Flight heritage | Score |
|---|---|---|---|---|---|
| Li-ion NCM/NCMA, aerospace cellsSelected | ~150–170 Wh/kg installed | Heated enclosure required; hibernation strategies proven in NASA studies | 2,000+ cycles at 80% DoD | ISS, Orion, every modern spacecraft; GM lunar NCMA variant in work | 8.9 |
| LiFePO₄ (LFP) | ~110 Wh/kg — heavy for the same night reserve | Poor charge acceptance below 0 °C | Excellent | Limited space use | 6.4 |
| Lithium-sulfur | 300+ Wh/kg on paper | Unproven | Hundreds of cycles, not thousands | None crewed | 4.7 |
| Regenerative fuel cell | Best for multi-week night power | Complex water/thermal management | Good | Demonstrators only | 5.9 |
A single 100 kWh Li-ion NCM pack, internally split into two isolated strings — one physical pack, as designed, but no single cell fault can take down the bus. The pack lives in an insulated, MLI-wrapped enclosure with thermal switches, survival heaters, and provision for radioisotope heater units on night-critical missions. Recharge comes from the deployable array or a base-camp charge port. Night strategy: hibernate at minimum state-of-charge ~30%, wake on sunrise.
The best wheel NASA has ever developed is the superelastic tire from NASA Glenn and Goodyear: a non-pneumatic mesh woven from nickel-titanium shape-memory alloy. Where the Apollo wire-mesh wheel deformed permanently past ~5% strain, nitinol springs recover from up to 10% strain — the tire can be crushed flat to the axle and snap back to shape. That is the difference between a wheel that survives one rock strike and one that survives ten years of them.
| Wheel | Compliance / traction | Damage tolerance | Mass | Heritage | Score |
|---|---|---|---|---|---|
| NiTi superelastic spring tireSelected | Conforms to rocks and soft regolith; large contact patch | Recovers from full deflection to the axle; no permanent set | Moderate | NASA Glenn + Goodyear; licensed terrestrially; baselined by LTV teams | 9.3 |
| Apollo zinc-coated piano-wire mesh | Good in soft soil | Permanent deformation past ~5% strain | Lightest | Apollo LRV — 3 missions, ~90 km total | 6.6 |
| Rigid aluminum + grousers (Lunokhod / MER style) | Small contact patch; digs in on slopes | Skin punctures (Curiosity's wheels) | Moderate | Robotic rovers | 5.2 |
| Compliant composite (fiberglass hoop) | Good | Fatigue and UV/radiation embrittlement over 10 yr | Light | Prototypes | 5.8 |
Six NiTi superelastic spring tires on machined aluminum hubs, sized ~750 mm diameter. Tread grousers in the mesh handle slope climbing; the open weave sheds regolith instead of caking. Licensing route: NASA Glenn's superelastic tire patents are available through NASA's Technology Transfer program — we license rather than reinvent.
Carbon fiber alone is the wrong answer, and it's worth being precise about why. For micrometeoroid protection, what matters is shocking the impactor into vapor at the first surface — aluminum bumpers do this better than CFRP, which fails in brittle, delaminating modes under hypervelocity impact. The proven architecture is a stuffed Whipple shield: a sacrificial outer bumper, a standoff gap, ceramic/aramid stuffing (Nextel + Kevlar), and a structural rear wall. Full protection from large impactors isn't physically achievable for any vehicle — the design standard is probabilistic: no penetration of crew-critical zones over vehicle life at >99% confidence, which is how ISS modules are specified.
| Shell approach | MMOD behavior | Thermal (−170 to +120 °C) | Mass / stiffness | Score |
|---|---|---|---|---|
| Hybrid: Al-Li frame + CFRP panels + Whipple zonesSelected | Al bumper + Nextel/Kevlar stuffing over crew-critical zones; distributed spall protection elsewhere | CTE-matched joints designed in; white ceramic thermal coating sheds heat and dust | Best stiffness-to-mass; CFRP where loads are distributed, metal where they concentrate | 8.7 |
| All CFRP monocoque | Brittle hypervelocity response, delamination, hidden damage | Matrix microcracking under 300 °C swings; charging/ESD issues | Lightest, stiffest | 5.9 |
| All aluminum-lithium (2195/2050) | Excellent bumper material, well-characterized | Predictable, weldable, repairable | ~15–20% heavier than hybrid | 7.8 |
| Titanium primary structure | Good | Excellent | Heavy and expensive to machine | 5.4 |
Aluminum-lithium primary chassis and roll structure, CFRP sandwich body panels, and stuffed-Whipple armor over the cockpit and battery. The white outer skin is a ceramic thermal-control coating chosen for low solar absorptance and dust release. Secondary rock-fall and rollover protection comes from the roll hoop behind the crew — the same philosophy as a rally car, sized for lunar gravity but Earth-mass inertia.
Your instinct about regolith is exactly the failure mode NASA worries about. Every pressurized hatch that opens on the surface becomes a dust airlock problem: charged, barbed regolith particles embed in seals and abrade them with each cycle. There is a clean way around it — the suitport, where suits dock backward through the rear bulkhead and never enter the cabin — but it belongs on a larger vehicle. The trade below is why Prometheus Block 1 flies unpressurized.
| Architecture | Crew benefit | Mass & power | Dust risk | Fit to NASA LTV req. | Score |
|---|---|---|---|---|---|
| Unpressurized + seat-side umbilicalsSelected — Block 1 | Suits stay on; vehicle O₂, power, cooling and comms via umbilical extend EVA well past 8 h | ~1,500–2,500 kg class; 2–4 kW | No pressure seals to contaminate | Matches LTV requirement exactly | 8.9 |
| Fully pressurized cabin (helmet-off) | Shirtsleeve driving, meals, rest — true mobile habitat | 5,000–10,000 kg; 5–10 kW continuous; needs a much larger lander | Hatch seal degradation with every ingress | Exceeds LTV scope (this is JAXA/Toyota Lunar Cruiser territory) | 5.6 |
| Pressurized + rear suitports | Shirtsleeve cabin and dust stays outside — suits never come in | Same mass class as pressurized | Best-in-class dust control | Future pressurized-rover program | 7.2 (Block 2) |
Block 1 is unpressurized: the crew rides suited, plugged into vehicle life support through fighter-jet-style umbilical panels beside each seat — oxygen, power, cooling water and data, with PLSS recharge while driving. That turns drive time into consumables recovery time instead of consumables burn. Block 2 grows into a pressurized cabin with rear suitports, which solves the regolith-on-the-seals problem by keeping suits permanently outside the pressure vessel.
An F1 driving position is built around a soft human in a carbon tub. Our crew is inside a semi-rigid pressure suit with a hard upper torso and a life-support backpack — a reclined F1 posture is nearly impossible to enter in a suit and wrecks sightlines over the nose. The fighter-jet reference is much closer, with one lunar twist: in 1/6 g, standing for hours is easy, which is why several LTV concepts use a stand-up "chariot" station.
| Posture | Suit compatibility | Visibility | Rough-terrain restraint | 8-hour fatigue | Score |
|---|---|---|---|---|---|
| Fighter-style upright, ~15° recline, open-back seatSelected | Open seat back accepts the PLSS; umbilical panel at hip height like an ejection-seat connector | High eye point over the sloped nose | 5-point harness works with suit hard points | Good — suit carries some load in 1/6 g | 9.0 |
| F1 reclined tub | Ingress in a pressurized suit essentially impossible | Poor over-nose vision for rock avoidance | Excellent | Good | 4.1 |
| Stand-up chariot (lean restraint) | Easiest ingress of all | Best | Marginal over rocks at 15 km/h | Surprisingly good in 1/6 g, poor past ~4 h | 6.8 |
| Kneeling / saddle (motorcycle style) | Awkward with suit bearings | Good | Moderate | Pressure points in suit | 4.9 |
Two side-by-side fighter-style stations: upright seats with open backs for the life-support pack, five-point restraints to suit hard points, and the umbilical quick-disconnect panel on the inboard console beside each hip — connect by feel, gloved, without looking, exactly like strapping into an ejection seat. Hand controllers (translation + rotation stick, no pedals) sit on armrests so drive inputs stay stable when the terrain isn't.
Cockpit displays on the Moon fight three enemies: raw sunlight with no atmosphere to soften it, gloves with no capacitive touch, and radiation flipping bits in commercial silicon. The answer is layered — nothing safety-critical lives only on a screen.
| Layer | Role | Why it wins |
|---|---|---|
| 2 × ruggedized high-brightness displays (>1,500 nit, heated)Core | Terrain map, traverse plan, energy & consumables ledger, camera feeds | Sunlight-readable panels with physical bezel keys and rotary encoders — every function reachable with a pressurized glove; touch is a convenience, never a dependency |
| Helmet HUD via suit data linkCore | Speed, heading, power reserve, RTB margin, nav cues | Eyes stay on terrain; data follows the crew off the vehicle on foot |
| Hardwired annunciator row + voice assistant | Cautions and warnings; hands-free queries and checklists | Works when everything else doesn't; voice is glove-proof by definition |
| Telemetry backhaul — LunaNet / DSN relay | Full vehicle telemetry to Earth and base camp; remote driving between crews | NASA requires remote operation between Artemis missions — telemetry is a revenue system, not overhead |
Maps come from LRO laser-altimeter terrain data fused with onboard stereo cameras and visual odometry — there is no GPS on the Moon, so Prometheus navigates like a spacecraft: star tracker + IMU + terrain-relative navigation. Radiation-tolerant flight computers run the vehicle; the pretty displays are clients, never authorities.
Baked into the flight software, exactly as you specified: Prometheus continuously solves the return problem. Every few seconds it recomputes the energy and consumables needed to reach the landing site or base camp over real terrain — slopes, temperature, motor efficiency, suit O₂ state — and compares it against what's on board.
The rule is simple: the vehicle must always be able to get home with 10% energy and consumables to spare. When the margin trends toward the floor, Homeward advises; when it reaches it, Homeward turns the vehicle around and drives — crew aboard or not. The same autonomy stack handles remote science traverses between crews, which is how the vehicle earns its keep for ten years, not ten days.
Mass stays; weight divides by six. A 2,400 kg Prometheus presses on the regolith with the force of a 400 kg Earth vehicle — roughly 650 N per wheel. That's the whole game: traction is proportional to weight, but inertia is proportional to mass. The vehicle is easy to get stuck-free and hard to stop; it brakes, corners and absorbs rock strikes like a fully loaded truck while gripping like a golf cart. This is why torque vectoring and compliant tires beat raw power — the limit is almost never motor torque, it's slip at the contact patch. It's also why Apollo crews reported the LRV going airborne over bumps at 10 km/h: less weight means less suspension force is needed to launch you. Suspension travel is sized long and soft, and Homeward caps speed over rough terrain automatically.
Range is sized to the reserve rule, not the brochure. NASA's LTV floor is 20 km per charge and 8-hour sorties. Prometheus targets 50 km nominal: a 20 km science radius means 40 km round trip, plus slope and thermal penalties, plus the 10% Homeward floor — with margin left for the sortie after this one.
The subsystems above are the headline trades. A NASA-credible proposal has to close all of these too — this is the honest bill of work.