In 1981, launching a kilogram to low Earth orbit cost roughly $85,000 in today’s dollars. The Space Shuttle, for all its technical achievement, was not cheap. In 2006, SpaceX was a startup that had not yet reached orbit. In 2024, Falcon 9 routinely launches payloads to LEO at approximately $2,700 per kilogram.
That change is larger than it first appears. A 97 percent reduction in launch cost over 40 years, concentrated mostly in the last decade, is not a gradual improvement. It is a discontinuity.
How it happened
The conventional answer is SpaceX and reusability. That is correct but incomplete.
SpaceX changed the economics of launch primarily through vertical integration and manufacturing efficiency before reusability was even demonstrated. By building its own rockets and engines, by treating rocket manufacturing as a production engineering problem rather than a custom fabrication problem, the company drove down costs on expendable vehicles faster than the industry had managed in decades.
Reusability then applied a multiplier. A Falcon 9 first stage can fly 10 to 20 times before retirement. The booster is the most expensive part of the rocket. Spreading that cost over many flights, rather than throwing it away after each launch, changes the per-launch economics dramatically.
Falcon Heavy, which can lift 64,000 kilograms to LEO with full reuse of side boosters, currently charges approximately $97 million per launch. That is more than Falcon 9 in absolute terms but substantially cheaper per kilogram for heavy payloads than any competitor.
Starship, when it reaches full operational capacity, is targeting costs of roughly $100 per kilogram to LEO at high launch rates. If that is achieved, it represents another order-of-magnitude reduction.
What the cost reduction actually enables
The economic effects of cheaper launch are not simply linear, where existing customers buy more launches at lower prices. The effects are nonlinear: entirely new categories of customers become viable.
Below roughly $10,000 per kilogram, smallsat constellations become commercially interesting. Below $5,000 per kilogram, commercial Earth observation businesses become possible. Below $3,000 per kilogram, the economics of satellite internet fundamentally change. Below $1,000 per kilogram, missions that would previously require dedicated national budget commitments become accessible to universities and smaller nations.
Each threshold creates new industries, not just new customers for existing industries.
Who is competing
The framing of the “space race” as US versus China captures something real about national competition but misses the more consequential story. China has made remarkable progress in launch capability. Long March 5B is a competitive heavy-lift vehicle. China plans a Starship-class super heavy launch vehicle. Chinese smallsat and megaconstellation ambitions are significant.
But the companies pushing the cost curve fastest are American, primarily, though not exclusively. Rocket Lab’s Electron is the dominant small launch vehicle globally. ULA’s Vulcan is entering service. Blue Origin’s New Glenn is operational. Relativity Space and ABL Space are pursuing smaller vehicles.
Europe has struggled. Ariane 6, delayed and over budget, launched for the first time in 2024. Vega-C has had failures. European access to orbit is no longer guaranteed at competitive prices.
Japan’s H3 rocket is in development. India’s LVM3 has successfully carried commercial payloads. New Zealand, the UK, Brazil, and South Korea all have active launch programs or commercial launch sites in development.
The global expansion of launch capability is the story, not bilateral competition.
What this means for science
Science missions have historically been priced at the expense of launch. A $3 billion planetary mission includes a substantial fraction devoted to the rocket. When launch costs fall, the science-to-total-cost ratio improves. Missions that would previously require major national commitment become achievable at smaller scale.
This has already changed cubesat science. Dozens of scientific missions that would have been impossible to fund as standalone programs now fly as rideshare passengers on commercial rockets. The data quality is not always equivalent to dedicated missions, but the quantity of experiments in space is increasing dramatically.
For deep-space science, the effect is slower but building. Lunar missions that would have required Apollo-scale commitments can now be flown commercially at costs accessible to smaller space agencies and even well-funded research universities.
The ceiling
The launch cost curve has a floor. Chemical rockets carrying payloads to low Earth orbit will not approach zero cost regardless of efficiency improvements, because propellant has a minimum cost and the physics of the rocket equation sets minimum mass ratios for useful payloads.
The floor for chemical launch to LEO is probably somewhere in the range of $50 to $200 per kilogram at industrial scale. Starship’s stated target of $100 per kilogram assumes very high launch rates, which in turn require demand well beyond current levels.
Whether that demand materializes depends on a self-reinforcing cycle: lower costs enable new applications, new applications create demand, demand justifies investment in higher launch rates, higher launch rates drive costs lower. The cycle is running, but it is not guaranteed to continue at the pace of the last decade.
The remarkable thing about the last fifteen years is that it happened at all. The entrenched incumbents, the cost structures, the institutional inertia of government-funded spaceflight, all predicted a future of incremental improvement. What arrived instead was a discontinuity.
The next discontinuity, if it comes, will look like Starship reaching its cost targets. Or it will look like something we have not yet built.