To build a motorcycle that runs in space, you have to replace the three things a bike takes from Earth for free: air to burn, air to carry heat away, and a road to push against. That means carrying liquid oxygen (about 3.4 kg for every kg of petrol) or going electric, fitting a radiator that can only glow its heat away (about 23 m² for a petrol engine's waste heat), and swapping tyres for gas thrusters and a reaction wheel. Do all that and you have rebuilt NASA's 1984 Manned Maneuvering Unit with a seat on it.
Calculated in this post from the 14.7 : 1 air-fuel ratio (Wikipedia: Air–fuel ratio) and oxygen's 23.14% share of air by mass (Wikipedia: Atmosphere of Earth) and the radiator working below, with these example inputs (15 kW engine, 30% efficient, 77 °C coolant, white-painted panel).
- 3.4 kg of oxygen must ride along for every kg of petrol, because space has no air to burn.
- About 23 m² of radiator is needed to dump a 15 kW petrol engine's 35 kW of waste heat at 77 °C.
Calculated in this post from the Stefan–Boltzmann constant (NIST CODATA) and the example inputs below (15 kW engine, 30% efficient, 77 °C coolant).
- 80° per second: how fast the bike would flip backwards if you spun the rear wheel to 1,000 rpm in orbit.
- 33 m/s (about 120 km/h) of total speed change from the MMU's 11.8 kg of nitrogen, applied to a 250 kg bike.
Calculated in this post from the MMU's two 5.9 kg nitrogen tanks (Wikipedia: Manned Maneuvering Unit) and an assumed 70 s specific impulse and 250 kg bike, with these example inputs.
- Verdict: possible with changes. Electric power, cold-gas thrusters and a reaction wheel work. A wheel-driven petrol bike does not.
This is a tier-2 what-if, so it gets solved like an engineering problem: assumptions, the governing laws, the numbers carried through, a sensitivity check and a verdict. The laws are old and well tested. The only fictional part is the motorcycle.
On 7 February 1984, four days after Challenger launched on mission STS-41-B, Bruce McCandless II unclipped from the shuttle and flew away from it with nothing holding him. He sat in the Manned Maneuvering Unit, a backpack with 24 small nitrogen thrusters, and steered with hand controllers at the ends of its two arms. He travelled about 90 metres from the orbiter and came back.
The MMU flew on three missions that year, logging 6 hours and 29 minutes in total, before NASA judged it too risky and retired it after the 1986 Challenger disaster. It is the closest anyone has come to riding a motorcycle in orbit.
Physics Today: The first untethered spacewalkWhat are the assumptions?
Every number later depends on these ground rules. Where a value is my estimate rather than a measured figure, the table says so.
| Symbol | Meaning | Value used |
|---|---|---|
| m | Bike plus suited rider | 250 kg (estimate) |
| Pshaft | Engine output, a mid-size single | 15 kW (estimate) |
| η | Petrol engine efficiency | 30% (estimate) |
| AFR | Stoichiometric air-fuel ratio for petrol | 14.7 : 1 |
| YO2 | Oxygen share of air by mass | 0.231 |
| LHV | Energy in petrol (E10) | 31.2–32.4 MJ/L, about 43 MJ/kg |
| σ | Stefan–Boltzmann constant | 5.67 × 10-8 W/m²K⁴ |
| ε | Radiator emissivity | 0.9 (estimate, white paint) |
The bike floats freely in orbit, far from any surface. "Runs" means three things: the engine keeps turning, it does not overheat, and the rider can actually go somewhere.
Why won't a normal engine start in space?
A petrol engine is an air engine that happens to burn a little fuel. By mass, it breathes 14.7 parts of air for each part of petrol, and only the oxygen in that air does any work. In a vacuum the intake pulls in nothing, so the spark has nothing to light.
The fix is to carry the oxygen. The working is short:
- Oxygen per kg of fuel. \( m_{O_2} = \text{AFR} \times Y_{O_2} = 14.7 \times 0.231 = 3.40 \) kg of oxygen per kg of petrol.
- A full tank. Take 13 L of petrol at about 0.74 kg/L (both my estimates): 9.6 kg of fuel.
- Oxygen to burn it. \( 9.6 \times 3.40 = 32.7 \) kg of oxygen, more than three times the fuel's mass.
- Storage. As liquid oxygen at −183 °C and roughly 1.14 kg/L (Wikipedia: Liquid oxygen), that is about 29 L of cryogenic tank.
Calculated in this post from the 32.7 kg of oxygen above and liquid oxygen's density of 1.141 kg/L (Wikipedia: Liquid oxygen).
At that rate the 9.6 kg tank lasts about 2.3 hours, and the bike needs 3.9 g of oxygen every second to go with it. You would also need to feed the engine at a controlled pressure, since no atmosphere pushes the mixture in. Without air you have, in effect, built a rocket engine with a crankshaft.
Calculated in this post from the 1.16 g/s fuel flow above, which uses petrol's energy content from US DOE data, and the 13 L tank assumed above, with these example inputs.
How do you cool an engine where there is no air?
In space, the only way to dump heat is to glow it away as infrared light. No air means no convection, so the fins on an air-cooled single do almost nothing. The law that sets how much a hot surface radiates is the Stefan–Boltzmann law: power rises with the fourth power of absolute temperature.
Deep space sits at about 3 K, so \(T_{sink}^4\) is negligible. A flat panel radiates from both faces, so the panel area is \(A = Q / (2\varepsilon\sigma T^4)\).
Worked solution: radiator for the petrol bike
- Waste heat. At 30% efficiency, 15 kW of shaft power means 50 kW of fuel power, so \( Q = 50 - 15 = 35 \) kW to dump.
- Radiator temperature. Coolant at 77 °C, so T = 350 K. \( \sigma T^4 = 5.67\times10^{-8} \times 1.50\times10^{10} = 851\ \text{W/m}^2 \).
Calculated in this post from the Stefan–Boltzmann constant (NIST CODATA) and the 77 °C coolant assumed in this post, with these example inputs.
- Per face. \( 0.9 \times 851 = 766\ \text{W/m}^2 \); both faces give 1,532 W per m² of panel.
Calculated in this post from the Stefan–Boltzmann constant (NIST CODATA) and the 0.9 emissivity assumed above, with these example inputs.
- Area. \( A = 35{,}000 / 1{,}532 = 22.9\ \text{m}^2 \). That is a panel about 5 m by 4.6 m, bolted to a 150 kg motorcycle.
Calculated in this post from the 35 kW of waste heat above and the Stefan–Boltzmann constant (NIST CODATA), with these example inputs.
For scale, the International Space Station's main cooling system is rated at 70 kW. NASA's overview says it uses six radiator units, each 23.3 m by 3.4 m and weighing 1,122.6 kg. The station runs its ammonia far colder than an engine, so it needs much more area per kilowatt. A petrol space motorcycle would have to reject half the station's design heat load, on a vehicle that weighs less than one of its radiator units.
How sensitive is the radiator to temperature?
Very. Because of the fourth power, running the radiator hotter shrinks it dramatically. Double the absolute temperature and the area falls 16 times. This is the main design lever, and it is why some space power concepts use radiators that run at hundreds of degrees.
Calculated in this post from the fourth-power term in the Stefan–Boltzmann law shown above (constant from NIST CODATA): 2⁴ = 16.
See the data as a table
| Two-sided panel area (m²) | |
|---|---|
| 300 K (27 °C) | 42.3 |
| 350 K (77 °C) | 22.9 |
| 400 K (127 °C) | 13.4 |
| 500 K (227 °C) | 5.5 |
| 600 K (327 °C) | 2.6 |
Source: Calculated with the Stefan–Boltzmann law, σ from NIST CODATA, ε = 0.9, sink at 3 K, no sunlight · A petrol engine's coolant cannot run much above about 120 °C without boiling or damaging seals, which keeps the area large.
Use this to check the numbers yourself. Change the heat load, the radiator temperature or the emissivity and see the panel grow or shrink.
Space radiator sizing
Is an electric motorcycle better in space?
Yes, by a wide margin. An electric drive needs no oxygen and wastes far less heat. Assume a motor and controller about 95% efficient (my estimate for a modern permanent-magnet drive).
- Waste heat: 15 kW out needs 15.8 kW in, so only about 0.8 kW to reject.
Calculated in this post from my 15 kW example output and my 95% efficiency estimate above, with these example inputs.
- Radiator: at 350 K that is about 0.5 m², a panel the size of a large tank cover.
- Energy: batteries charged from solar arrays, the way the ISS and every satellite do it.
- Trade-off: batteries store far less energy per kg than petrol, but petrol needs 3.4 kg of oxygen per kg anyway, which closes much of the gap.

What happens when you twist the throttle in orbit?
The bike flips over backwards. On Earth, the rear tyre pushes on the road and the road pushes back. In orbit there is no road, so spinning the wheel only moves angular momentum around. The law is conservation of angular momentum: with no outside torque, the total spin of the system stays at zero.
- Wheel inertia. A rear wheel with tyre, about 0.8 kg·m² (my estimate).
- Bike and rider in pitch. About 60 kg·m² about the centre of mass (my estimate).
- Throttle. Spin the wheel to 1,000 rpm: \( \omega = 1000 \times 2\pi/60 = 104.7 \) rad/s.
Calculated in this post from my example of 1,000 rpm, with these example inputs.
- Result. \( \omega_{bike} = (0.8/60) \times 104.7 = 1.40 \) rad/s, about 80° per second. The bike does a full backflip in 4.5 seconds and keeps going until you close the throttle.
That is not a bug. It is exactly how satellites point themselves. A reaction wheel is a flywheel driven by a motor, and spinning it turns the spacecraft the other way. Your rear wheel becomes the attitude control system, but only for one axis. Three wheels at right angles give full control.

So how do you actually go somewhere?
You throw mass out of the back. With nothing to push against, the only way to change speed is to push on your own propellant. That is the Tsiolkovsky rocket equation, which links speed change to exhaust speed and how much of your mass you throw away.
Here \(I_{sp}\) is specific impulse in seconds, a measure of exhaust speed, and \(g_0 = 9.81\) m/s². Two builds show the range.
Build A: the MMU's nitrogen
- Propellant. The MMU carried two tanks of 5.9 kg of nitrogen (Wikipedia: Manned Maneuvering Unit), 11.8 kg in all.
- Exhaust. Cold nitrogen gas gives an \(I_{sp}\) of roughly 70 s (my estimate, a typical textbook value for cold gas).
- Mass ratio. 250 kg full, 238.2 kg empty: \( \ln(250/238.2) = 0.0484 \).
Calculated in this post from my 250 kg example mass and the 11.8 kg of nitrogen above (Wikipedia: Manned Maneuvering Unit), with these example inputs.
- Result. \( \Delta v = 70 \times 9.81 \times 0.0484 = 33 \) m/s, about 120 km/h of total speed change, starting and stopping included.
Calculated in this post from the 0.0484 above, which uses the MMU's 5.9 kg tanks (Wikipedia: Manned Maneuvering Unit), and my assumed 70 s specific impulse, with these example inputs.
That lands close to the 34–40 m/s (110–130 ft/s, as given by Wikipedia: Manned Maneuvering Unit, citing NASA) for a ground-charged MMU, which is a useful sanity check on the 70 s assumption.
Metric figures converted in this post from the 110–130 ft/s given by Wikipedia: Manned Maneuvering Unit (1 ft = 0.3048 m).
Build B: burn the petrol as a rocket
- Propellant. 9.6 kg petrol plus 32.7 kg oxygen, 42.3 kg in all.
Calculated in this post from my 9.6 kg of petrol and the 32.7 kg of oxygen worked out above (14.7 : 1 air-fuel ratio, Wikipedia: Air–fuel ratio), with these example inputs.
- Exhaust. About 300 s, my rough figure for a small kerosene-type and oxygen thruster.
- Result. \( \Delta v = 300 \times 9.81 \times \ln(292.3/250) = 460 \) m/s.
Calculated in this post from the 42.3 kg of propellant above and my rough 300 s specific impulse, with these example inputs.
In space, petrol only moves you if you burn it as rocket propellant, and then it gives about 14 times the speed change of the MMU's nitrogen. Driving a wheel with it gets you nowhere, because the wheel has nothing to grip.
Which build wins?
Lay the options side by side and the shape of the real machine appears.
| Option | Needs carried oxygen? | Waste heat (15 kW out) | Can it move you? |
|---|---|---|---|
| Petrol engine driving a wheel | Yes, 3.4 kg per kg fuel | 35 kW, 22.9 m² radiator | No, only flips the bike |
| Electric motor driving a wheel | No | 0.8 kW, 0.5 m² | No, but steers attitude |
| Cold-gas thrusters (MMU) | No | Almost none | Yes, about 33 m/s total |
| Petrol and oxygen rocket | Yes | Carried off by exhaust | Yes, about 460 m/s total |
Written before the first flight, this paper describes how the MMU was built for redundancy: two identical propulsion systems, each with a high-pressure nitrogen tank, an isolation valve, a regulator and twelve 7.5 N (1.7 lbf) thrusters, giving six-degree-of-freedom control from hand controllers. It also describes an automatic attitude-hold mode on all three rotation axes. The limit to know: it covers the controller arm and propulsion layout, not measured flight performance.
Read the paper (free)NASA's description of how the station rejects heat. The external system moves heat in liquid ammonia to two rotating radiator wings of three units each; every unit is 23.3 m by 3.4 m and weighs 1,122.6 kg, and the system is designed for 35 kW per loop, 70 kW in total. An earlier temporary system managed 14 kW. The limit: it is a programme overview with design ratings, not a record of on-orbit performance.
Read the report (free)In September 2008, imagery showed a face sheet on one starboard radiator panel lifting off its honeycomb core. NASA isolated that radiator and vented its ammonia, then tracked every panel by photograph for years. In January 2025 the radiator was brought back into the loop to add heat-rejection capacity, damaged panel and all. The limit: the record summarises the investigation; detailed thermal numbers are in the full paper.
Read the paper (free)Space is cold, but it is a terrible place to cool anything. With no air, every watt has to leave as light.From the radiator working above
So, can it be done?
Pulling the numbers together:
- Engine: a petrol engine runs if you feed it 3.4 kg of oxygen per kg of fuel, but it needs a 23 m² radiator.
Calculated in this post from my example inputs and the figures worked out above, with these example inputs.
- Wheels: driving a wheel in orbit only rotates the bike, 80° per second at 1,000 rpm.
- Moving: thrusters work, 33 m/s from MMU-sized nitrogen or about 460 m/s from petrol burned as a rocket.
- Rider: needs a full pressure suit, as McCandless wore.
A motorcycle that runs in space is possible only if you stop treating it as a motorcycle. Replace the petrol engine with an electric drive (0.8 kW of heat instead of 35 kW), use the wheels as reaction wheels for attitude, and move with gas thrusters, which give about 33 m/s from 11.8 kg of nitrogen. That machine already flew: the 148 kg Manned Maneuvering Unit of 1984.
Calculated in this post from my 95% electric efficiency and 15 kW example output above, and the MMU's 11.8 kg of nitrogen (Wikipedia: Manned Maneuvering Unit), with these example inputs.
A petrol engine driving a rear wheel is ruled out by conservation of angular momentum, since with no road the throttle can only spin the bike. For a rough sense of cost, NASA's May 2026 awards for crewed lunar rovers were about $219–220 million each, and an orbital bike would not be cheaper.
Gravity changes everything again once there is ground under the wheels, which is where riding on the Moon picks up. For how power and torque behave when the road does push back, read how engine power and torque work. Speed taken to its limit is in riding at nearly the speed of light, and how a bike really turns on Earth is in the physics of countersteering. More machines live on Yantra.
What's new
- NASA awards Astrolab and Lunar Outpost about $220 million each for crewed lunar rovers NASA, May 2026
- NASA paper explains how a damaged ISS radiator was returned to service in 2025 NASA NTRS, Apr 2026
- Moon-base rover deals detailed: 200 km range, 10 km/h, spacesuited crews SpacePolicyOnline, May 2026
What if next?
- What would riding on the Moon feel like? There is ground again, but your brakes lose five-sixths of their grip.
- Could a petrol engine run on Mars? Mars has air. It just has almost no oxygen in it.
- What does an AI-native motorcycle look like in 2045? Back on Earth, with a bike that sees the truck three cars ahead.
Questions people ask
Can a petrol engine work in space?
Only if you bring the oxygen. A petrol engine needs about 14.7 kg of air per kg of fuel, which works out to 3.4 kg of oxygen. Supply that from a tank and the engine runs, but it must still shed its waste heat through a large radiator.
Why do spacecraft need radiators if space is cold?
Because cold only helps if heat can flow into it. In a vacuum there is no air to carry heat, so the only route is infrared radiation. That is why the ISS needs six radiator units of 23.3 m by 3.4 m to reject 70 kW.
How did astronauts move around without a tether?
With the Manned Maneuvering Unit, a backpack with 24 nitrogen thrusters of 7.5 N each, flown on three Shuttle missions in 1984. Today's spacewalkers stay tethered and carry a smaller emergency jetpack instead.
What would happen if you spun a wheel in space?
Your body and bike would rotate the opposite way. Angular momentum is conserved, so a 0.8 kg·m² wheel at 1,000 rpm would turn a 60 kg·m² bike and rider at about 80° per second.
Calculated in this post from my example wheel and bike inertias and 1,000 rpm worked out above, with these example inputs.
Sources
- Physics Today: The first untethered spacewalk (source for 7 February 1984, 24 thrusters, about 90 m, 6 h 29 min total, retirement after 1986)
- Wikipedia: Manned Maneuvering Unit, citing the NASA MMU Users Guide (1978) (source for 148 kg, 2 × 5.9 kg nitrogen, 110–130 ft/s delta-v)
- Falkner, K. E.: The Manned Maneuvering Unit Flight Controller Arm, NASA NTRS (1983) (source for 7.5 N thrusters, two redundant systems of twelve)
- NASA: ISS Active Thermal Control System overview (source for 70 kW, 35 kW per loop, radiator unit size and mass)
- Rodriguez, A. and Morton, R.: ISS Radiator Face Sheet Anomaly Investigation and Return to Function, ICES-2026-281 (2026) (source for the 2008 damage and 2025 return to service)
- NIST CODATA: Stefan–Boltzmann constant (source for σ = 5.67 × 10-8 W/m²K⁴)
- NASA NSSDC: Earth fact sheet (source for solar irradiance of 1,361 W/m²)
- US DOE Alternative Fuels Data Center: Fuel properties comparison (source for petrol at 112,114–116,090 Btu/gal, converted to 31.2–32.4 MJ/L)
- Wikipedia: Air–fuel ratio (source for the stoichiometric ratio of about 14.7 for petrol)
- NASA: Update on Moon base rovers, landers and missions (2026) (source for the $219 million and $220 million rover awards)
- SpacePolicyOnline: NASA awards initial Moon base rover and lander contracts (2026)
