If you built a road straight up into space, to the 100 km Kármán line, at the 7% gradient Indian hill roads allow at most, it would be 1,429 km long. Riding it, you give up first: without oxygen, useful consciousness lasts about 3 to 5 minutes at 7.6 km and 9 to 12 seconds at 15 km, and above about 19 km the water in your mouth and lungs boils at body temperature.
- 1,429 km of road at a 7% gradient to reach the 100 km Kármán line.
- 3 to 5 minutes of useful consciousness at 7.6 km; 9 to 12 seconds at 15 km (FAA).
- ~13 km: where a 30 kW petrol engine can no longer climb 7% at 30 km/h, by this post's model.
- 19.2 km: the Armstrong limit, where body-temperature water boils.
- 15.4 km for steel against 107.5 km for graphite-epoxy: the characteristic length that decides whether the tower can stand.
A petrol bike gives up next, near 13 km, where the air is a fifth as dense. The road itself would need a tapered composite tower, because a straight steel column fails under its own weight at about 15 km.
The plan for this route is a 100 km climb, so this tier-3 what-if solves it like an engineering brief:
- the shape of the road;
- the air along it;
- when the rider, the engine and the tarmac each give up;
- the energy it takes;
- what you'd have to change to actually reach the top.
The atmosphere numbers come from the US Standard Atmosphere 1976 model; the human limits from the US Federal Aviation Administration (FAA).
How long would a road into space be?
A road can't go literally straight up, because tyres need a gradient to climb, so the real question is how long a ramp to 100 km must be. The Kármán line at 100 km is the edge of space used by the Fédération Aéronautique Internationale (FAI), the body that certifies aviation and space records.
Indian hill roads are designed to IRC:52-2019, which sets a ruling gradient of 5%, a limiting 6% and an exceptional 7%. For a road of constant gradient \( G \) climbing a height \( H \), the horizontal length is
So the road to space is roughly Bengaluru to Mumbai and back, spiralled into a tower, or a ramp leaning across half of India.
My Bengaluru to Kolkata ride in November 2025 covered about 2,000 km in 26 hours 42 minutes of riding, an average of 74.9 km/h. At that pace the road to space is a 19-hour ride, though nobody would sustain that speed uphill.
What happens to the air as you climb?
It thins exponentially and gets colder, then warmer, then colder again. The US Standard Atmosphere 1976 describes the average atmosphere in layers: temperature falls 6.5 °C per km to −56.5 °C at 11 km, stays flat to 20 km, then rises slowly through the stratosphere. Pressure in each layer follows the barometric formula:
with \( M = 0.0289644 \) kg/mol for air, \( R^{*} = 8.31432 \) J/(mol·K), \( g = 9.80665 \) m/s² and \( \lambda \) the layer's lapse rate. Density follows from the ideal gas law, \( \rho = p/(R_{air} T) \). I computed these from the model's layer constants (sea level 101,325 Pa and 288.15 K; 22,632 Pa at 11 km; 5,474.9 Pa at 20 km; 868.0 Pa at 32 km).
| Altitude | Pressure | Temperature | Air density vs sea level | Oxygen partial pressure |
|---|---|---|---|---|
| 0 km | 101.3 kPa | 15.0 °C | 100% | 21.2 kPa |
| 3 km | 70.1 kPa | −4.5 °C | 74% | 14.7 kPa |
| 5.913 km (Mig La) | 47.7 kPa | −23.4 °C | 54% | 10.0 kPa |
| 11 km | 22.6 kPa | −56.5 °C | 30% | 4.7 kPa |
| 15 km | 12.0 kPa | −56.5 °C | 16% | 2.5 kPa |
| 19.2 km | 6.2 kPa | −56.5 °C | 8.2% | 1.3 kPa |
| 32 km | 0.87 kPa | −44.5 °C | 1.1% | 0.18 kPa |
| 47 km | 0.11 kPa | −2.5 °C | 0.12% | 0.02 kPa |
| 86 km | 0.0003 kPa | −88.5 °C | 0.0006% | effectively zero |
Real temperatures over India differ from this mid-latitude average, especially in the monsoon, but the pressure profile is close enough for this exercise.
On 16 August 1960, US Air Force Captain Joseph Kittinger rode an open gondola under a helium balloon, Excelsior III, to 102,800 ft (31,300 m): on our 7% road, about 447 km along. During the climb the pressurisation in his right glove failed. He kept going, and his right hand swelled to about twice its normal size in the near-vacuum.
At the top he stepped off. He fell for about 4 minutes 36 seconds, reached a top speed of 614 mph (988 km/h) in the thin air, and opened his main parachute at about 18,000 ft. Nine months earlier, on Excelsior I from about 76,400 ft, an equipment fault had sent him into a flat spin of about 120 rpm and knocked him unconscious; an automatic parachute opener saved his life. His swollen hand is the Armstrong limit in miniature: one failed seal, and the body starts to give way to the vacuum.
Wikipedia: Joseph Kittinger (Excelsior I and III, with USAF and citation sources)See the data as a table
| Density (% of sea level) | |
|---|---|
| 0 km | 100 |
| 3 km | 74.2 |
| 5.9 km | 54.4 |
| 8 km | 42.9 |
| 11 km | 29.7 |
| 15 km | 15.8 |
| 19.2 km | 8.2 |
| 25 km | 3.2 |
| 32 km | 1.1 |
| 47 km | 0.12 |
Source: Computed from US Standard Atmosphere 1976 layer constants · Half the air is below about 5.5 km; 90% is below about 16 km.
The standard defines an idealised, steady-state atmosphere from the surface to 1,000 km, built from rocket, satellite and balloon data for mid-latitudes. Below 86 km it is a set of layers with fixed lapse rates: −6.5 K/km to the tropopause at 11 km (216.65 K, 22,632 Pa), isothermal to 20 km, then +1 and +2.8 K/km through the stratosphere to 47 km. It is the reference behind altimeters, aircraft performance charts and engine-rating corrections. The limit: it is an annual mid-latitude average; the tropical tropopause over India sits higher and colder, so real densities at a given height differ by several percent.
Read the standard (free)When does the rider give up?
Long before the bike. Breathing depends on the oxygen partial pressure, which is 20.95% of total pressure: 21.2 kPa at sea level, 10.0 kPa at Mig La, 4.7 kPa at 11 km.
Calculated in this post from the oxygen share of air linked above and the pressures in the US Standard Atmosphere 1976 table below.
The FAA publishes how long a healthy person at rest keeps enough judgement to act, called the time of useful consciousness (TUC).
- 5.9 km, the real ceiling. Mig La, at 5,913 m, is the highest motorable road in the world. Acclimatised riders cross it; nobody lives up there.
- 5.5 km (18,000 ft). TUC 20 to 30 minutes (FAA, Figure 2-3).
- 7.6 km (25,000 ft). TUC 3 to 5 minutes, or 1.5 to 2.5 minutes after a sudden pressure loss (FAA, Figure 2-3). On the 7% road, 7.6 km is 109 km of riding, about 3.6 hours at the 30 km/h climbing speed assumed below.
Calculated in this post from the 7.6 km height, the 7% gradient (IRC:52-2019) and the 30 km/h climbing speed assumed below, with these example inputs.
- 12.2 km (40,000 ft). 15 to 20 seconds (FAA, Figure 2-3).
- 15.2 km (50,000 ft). 9 to 12 seconds.
- 19.2 km, the Armstrong limit. Total pressure falls to about 6.3 kPa, the vapour pressure of water at 37 °C, so saliva, tears and the moisture in your lungs start to boil. Beyond this only a full pressure suit keeps you alive.
The FAA's training circular for high-altitude pilots covers the physics of the atmosphere, hypoxia, decompression sickness and pressurisation. Its Figure 2-3 lists times of useful consciousness: 20 to 30 minutes at 18,000 ft, 10 minutes at 22,000 ft, 3 to 5 minutes at 25,000 ft, 2.5 to 3 minutes at 28,000 ft, 30 seconds to 1 minute at 35,000 ft, 15 to 20 seconds at 40,000 ft and 9 to 12 seconds at 50,000 ft, with much shorter times after rapid decompression. The FAA notes these are averages for healthy people at rest in altitude chambers, and that exercise, fatigue and individual differences shorten them. The limit for this post: riding a motorcycle uphill is exercise, so real times would be shorter than the table.
Read the circular (free)
When does the engine give up?
A naturally aspirated petrol engine runs out of breath at roughly 12 to 14 km on a 7% road, by my estimate. An engine burns fuel with the oxygen it breathes in; at full throttle the mass of air per stroke is proportional to air density, so maximum power falls roughly with density. Mechanical friction inside the engine does not fall, so there is a height where what is left can't push the bike uphill.
Assumptions. Bike plus rider 250 kg; speed 30 km/h (8.33 m/s); 7% gradient; rolling-resistance coefficient 0.012; air drag ignored (the air is thin). Sea-level engine power \( P_0 = 30 \) kW, with internal friction a fixed 15% of \( P_0 \) (my assumption). A fuel-injected engine that keeps the mixture right; a carburetted one does worse.
- Power needed to climb.
\[ P_{need} = m g v\,(\sin\theta + C_{rr}\cos\theta) = 250 \times 9.81 \times 8.33 \times (0.0698 + 0.012) = 1{,}671\ \text{W} \]
Calculated in this post, with these example inputs, from the scenario premises stated above (bike and rider mass, climbing speed, 7% gradient (IRC:52-2019), rolling-resistance coefficient).
- Power available at density ratio \( \sigma \). \( P_{avail} = \sigma P_0 - 0.15 P_0 \).
- Set them equal.
\[ \sigma = \frac{P_{need} + 0.15 P_0}{P_0} = \frac{1.67 + 4.5}{30} = 0.206 \]
- Find the height. From the standard atmosphere, \( \sigma = 0.206 \) at about 13.3 km. For a 20 kW engine it is 12.5 km; for 40 kW, 13.8 km.
Calculated in this post, with these example inputs, from the power balance above and the density ratios in the US Standard Atmosphere 1976 table.
That is 179 to 197 km along the road. I saw the start of this curve on the 2022 Gurudongmar ride, where my carburetted Pulsar NS200 at times could not go faster than about 40 km/h.
Calculated in this post by dividing the stall heights above by the 7% gradient (IRC:52-2019), with these example inputs.
Use this to check what you just read: change the engine, the load and the gradient.
How high can a petrol bike climb?
When does the tarmac give up?
The road surface fails from cold; the structure under it fails from its own weight. Above 11 km the air is −56.5 °C, far colder than paving bitumen is graded for in India, so the surface would turn brittle and crack under every wheel. But that is the small problem.
The big one is support. A road climbing into space has to stand on something, in effect a tower. NASA engineer Geoffrey Landis uses the characteristic length of a material, the height at which a column of constant cross-section crushes itself under its own weight:
For cast steel, strength 1.2 GPa and density 7,800 kg/m³ give \( L_c = 15.4 \) km (Landis, 2003). To go higher, a tower must taper, wider at the base, and for a tower carrying its own weight the required base-to-top area ratio grows exponentially with height:
- Cast steel, with safety factor (\( L_c \) = 12.3 km). \( e^{100/12.3} = 3{,}400 \). The base would need 3,400 times the cross-section of the top: absurd.
Calculated in this post from the 100 km height and the safety-factor characteristic length in Landis's Table 1.
- Graphite-epoxy composite, with safety factor (\( L_c \) = 53.8 km). \( e^{100/53.8} = 6.4 \). Buildable in principle.
Calculated in this post from the 100 km height and the safety-factor characteristic length in Landis's Table 1.
- Gravity barely helps. At 100 km, \( g \) is \( (6{,}371/6{,}471)^2 = 97\% \) of its surface value, so the tower and the bike still weigh almost as much.
Calculated in this post from Earth's mean radius and the 100 km height (NASA NSSDC Earth fact sheet).
Landis studies tall compression towers as launch platforms. He defines characteristic length as ultimate strength over density times surface gravity, tabulates it for grey iron (16.5 km), cast steel (15.4 km), graphite-epoxy (107.5 km) and boron-epoxy (122.5 km), and also gives each with a safety factor applied (cast steel 12.3 km, graphite-epoxy 53.8 km). A 15 km tower sized to carry a 2,000-tonne launch vehicle would weigh about 280 tonnes in graphite-epoxy, needing no taper, or 5,300 tonnes in cast steel with a taper ratio of 2.6. He notes 15 km puts a launch above 83% of the atmosphere. The limit: these are static compression estimates; wind loads, buckling and foundations would dominate a real design, and he stops well short of 100 km.
Read the paper (free)
How much energy would the climb take?
Less than you'd think: about 67 kWh to lift a 250 kg bike and rider 100 km, before friction. The potential energy gain is
Calculated in this post from Earth's GM and radius (NASA NSSDC Earth fact sheet), the 100 km height and an assumed bike and rider mass, with these example inputs.
which is 242 MJ, or 67 kWh, slightly less than the flat-gravity estimate \( mgH = 245 \) MJ because gravity weakens with height.
Calculated in this post from the same NASA values and example inputs, by integrating gravity from the surface to 100 km.
Add rolling resistance over 1,429 km, \( 0.012 \times 250 \times 9.81 \times 1.429\times10^{6} = 42 \) MJ, and motor losses, and a well-designed electric bike would need on the order of 100 kWh of battery: a large electric-car pack. You would come back down using regenerative braking.
Calculated in this post from an assumed rolling-resistance coefficient, the same example bike and rider and the road length worked out above, with these example inputs. The battery size is my own rough estimate.
Reaching 100 km is not reaching orbit. Stop at the top and you are standing on a very tall tower; ride off the end and you fall straight back.
Staying up needs sideways speed of about 7.8 km/s, which is why rockets, not roads, go to space.
Calculated in this post as the circular-orbit speed at the top of the road, the square root of GM divided by Earth's radius plus 100 km, using the NASA NSSDC Earth fact sheet values.
As a road for a normal rider on a petrol bike, it is impossible: the rider loses useful consciousness within minutes above 7.6 km, the engine can't climb 7% above about 13 km, and above the Armstrong limit at 19.2 km unprotected human tissue boils.
As an engineering project it is possible with three changes: a pressure suit, an electric motorcycle carrying about 100 kWh of battery, and a 1,429 km spiral ramp on a tapered graphite-epoxy tower with a base about 6.4 times the area of its top. The deciding number is the characteristic length: 15.4 km for steel, 107.5 km for graphite-epoxy. No one has priced a 100 km tower, so I won't invent a cost; Landis's own designs stop at 15 to 25 km.
How sensitive are these answers?
| If this changes | Then |
|---|---|
| Gradient 5% instead of 7% | Road 2,000 km; climbing power falls to about 1.3 kW, so the engine lasts to about 14 km |
| A turbocharged engine | Sea-level density can be held to a higher altitude, gaining several km before it also runs out |
| Carburetted engine | Mixture goes rich as air thins; power falls faster than density, so it quits lower |
| Rider uses supplemental oxygen | Useful consciousness extends to roughly the altitudes airliner masks are designed for, but not past the Armstrong limit |
| Space defined as 80 km | Road at 7% is 1,143 km; nothing else changes |
Where the model breaks: I treat engine power as proportional to density and friction as fixed, ignore cooling (thin air removes heat poorly, so an air-cooled engine would overheat sooner), and use a mid-latitude atmosphere. The engine numbers are estimates within about ±1.5 km.


What is the real highest you can ride?
About 5.9 km. Read the story of Mig La, the world's highest motorable road, the ranking of India's highest motorable passes, and before you go, altitude sickness on a motorcycle tour and the acclimatisation plan for a Ladakh ride. For why a 7% road is the steepest you'll find, see how ghat roads and hairpins are designed. RoadHook Machines explains fuel injection vs carburettor, the reason an old carburetted bike struggles first at altitude.
What's new
- Mig La at 19,400 ft (5,913 m) becomes the world's highest motorable road, built by BRO's Project Himank Autocar India, Oct 2025
- India's national highways reach 1,46,572 km in FY 2025-26, built at nearly 34 km a day Hindusthan Samachar, Jun 2026
What if next?
- What if you rode for a day without food? A 20-hour climb burns through more than your fuel tank.
- What if you held your bladder for a 10-hour ride? Pressure suits make that question much worse.
- What if you rode 24 hours without sleep? Hypoxia and sleep loss impair the same judgement.
Questions people ask
Where does space begin?
The FAI, which certifies aviation and space records, uses the Kármán line at 100 km. Some US conventions use a lower line, about 80 km (50 miles).
At what altitude does a motorcycle engine stop working?
There is no single number. A naturally aspirated engine loses power roughly in step with air density, about 46% less at 5.9 km. By my estimate a 30 kW engine can't climb a 7% grade at 30 km/h above about 13 km. Electric motors are unaffected by air.
Calculated in this post from the 54% air density at Mig La in the US Standard Atmosphere 1976 table above.
How high can a human go without oxygen?
Acclimatised people work at around 5 to 6 km. The FAA lists 3 to 5 minutes of useful consciousness at 25,000 ft (7.6 km) and 9 to 12 seconds at 50,000 ft (15.2 km), for healthy people at rest.
What is the Armstrong limit?
The altitude, about 18 to 19 km (63,000 ft in the US Standard Atmosphere), where air pressure falls to about 6.3 kPa and water boils at body temperature. Above it, exposed fluids such as saliva boil, and survival needs a pressure suit.
Could you build a tower 100 km tall?
Not with steel or concrete. A straight steel column fails under its own weight at about 15 km. Strong composites such as graphite-epoxy could in principle reach 100 km with a tapered shape, but no one has designed one; studies stop at 15 to 25 km.
Sources
- NOAA, NASA, USAF: U.S. Standard Atmosphere, 1976 (layer constants used for pressure, temperature and density)
- Wikipedia: U.S. Standard Atmosphere (layer table: 101,325 Pa; 22,632 Pa at 11 km; 5,474.9 Pa at 20 km; 868.0 Pa at 32 km)
- FAA: Advisory Circular 61-107B Change 1 (2015; times of useful consciousness by altitude)
- Landis GA: Towers for Earth Launch, NASA Glenn (characteristic length: cast steel 15.4 km, graphite-epoxy 107.5 km; 15 km tower masses)
- Wikipedia: Armstrong limit (18 to 19 km, 0.0618 atm, 63,000 ft in the standard atmosphere)
- Wikipedia: Kármán line (FAI definition of space at 100 km)
- Autocar India: Mig La becomes the world's highest motorable road (Oct 2025; 5,913 m, previous record Umling La)
- Indian Roads Congress: IRC:52-2019 (gradients 5% ruling, 6% limiting, 7% exceptional)
- NASA NSSDC: Earth fact sheet (mean radius 6,371 km, GM)
- Wikipedia: Joseph Kittinger (story block: Excelsior III, 102,800 ft, glove failure, 4 min 36 s fall; Excelsior I flat spin)
- RoadHook Rides: Mig La, the world's highest motorable road (5,913 m)
