Low Earth orbit demands roughly 9.4 km/s once gravity and drag take their cut. With the best chemical propellants, the equation forces a brutal answer: the vehicle must be about 85–94% propellant. The ship itself — tanks, engines, payload, people — gets the leftovers. That is why rockets are giant thin-walled fuel tanks, why payloads are typically only 2–5% of liftoff mass, and why every kilogram sent to orbit is argued over. The equation cannot be repealed; it can only be out-engineered.

Two escape routes exist. Staging — the first out-engineering trick — throws away empty tankage mid-flight so the equation restarts with a smaller, lighter ship; that's why nearly every orbital rocket in history is a stack of rockets riding rockets. Higher exhaust velocity is the other: hydrogen engines beat kerosene, and in space, ion drives beat everything — pushing gently for years with exhaust ten times faster than any chemical flame. Dawn and Hayabusa cruised the asteroid belt that way.

The benchmark for half a century was the Saturn V: 2,970 tonnes at liftoff, 111 meters tall, about 140 tonnes to low orbit — and its five F-1 engines drank propellant by the swimming pool. It flew thirteen times, never lost a crew, and put every set of human footprints on the Moon →. The Space Shuttle that followed traded raw lift for reuse-in-theory: a spaceplane that came home, strapped to throwaway tanks and boosters, that taught the industry — at terrible cost, twice — how hard "routine" spaceflight really is.

Today's giants split the philosophy. NASA's SLS is a Saturn-class expendable built for Artemis lunar missions; it flew the uncrewed Artemis I loop around the Moon in 2022. SpaceX's Starship is the opposite bet: the largest rocket ever flown, designed so that both stages come back — its booster was first caught out of the air by the launch tower's arms in October 2024, a maneuver that looked like science fiction on the way to becoming procedure.

Saturn V — still the yardstick. NASA

SLS on pad 39B — the same pad Saturn V used. NASA/KSC

Reuse attacks the only part of the price that engineering can reach: the hardware. Propellant is a rounding error — the metal was the money. Falcon 9 boosters now routinely fly, land on ships or pads, and fly again — individual boosters have logged twenty-plus missions — and launch cadence rose from a national event to a weekly errand. The consequences ripple through this whole site: cheap lift is what makes mega-constellations, bigger space telescopes →, and serious Mars architectures → arguable at all. Starship's wager is that full reuse of everything drops the cost another order of magnitude. If it pays out, the bottleneck of the entire space age quietly disappears.

Falcon 9 — the booster on this flight was already a veteran. NASA/KSC

Discovery — reuse, version one. NASA/HQ