I had an argument with my friend recently about the effect of technology maturation on spaceflight cost. In the process, I managed to find this infographic: Cost of Spaceflight Image source

The trend for cost reduction was clear to me, even discounting recent data, but my friend pointed out that: a) this picture seems to only have first launches, while, for example, Soyuz spacecraft modifications were in use for all of the relevant period, b) it collates very different vehicles, which affects the spread (for example, Space Shuttle and Saturn V), and c) majority of data points still fall in a relatively narrow price range that seems to be present over all of the covered timeframe.

My question is: do we observe that spaceflight costs are indeed reduced with technology improvement? I would like to find data on how cost of operation of long-running programs like Soyuz or Apollo and Space Shuttle changed over their lifetime, preferrably both with and without adjuctment for inflation.

  • $\begingroup$ Are you asking about the cost to the launch provider or the price they charge to their customers? $\endgroup$ Commented Aug 5, 2022 at 12:50
  • $\begingroup$ @DaveGremlin cost to the launch provider, I guess. $\endgroup$ Commented Aug 5, 2022 at 15:52
  • $\begingroup$ The price reductions shown in that graph are impressive. Then I realized that the Y axis is logarithmic, and it really blew me away how dramatic the cost reduction has been. $\endgroup$ Commented Aug 5, 2022 at 15:56
  • $\begingroup$ One problem is that most of the dramatic cost reduction is down to SpaceX and as a private company they don't publish costs so everything is an estimate. That Said it's probably broadly true. $\endgroup$
    – Slarty
    Commented Feb 26 at 10:15
  • $\begingroup$ Throw out Starship, which hasn't yet flown, and may never come close to its cost target, and the rest just looks like a scatter plot. $\endgroup$
    – John Doty
    Commented Feb 26 at 15:27

2 Answers 2


Focusing on your question...

My question is: do we observe that spaceflight costs are indeed reduced with technology improvement?

I studied this very topic for the Space Shuttle program and NASA's Commercial Resupply Services and included my findings in a paper published last year.

Keep in mind while reading this that what this article is discussing is the cost of sending cargo (and sometimes crew as well) to the ISS's orbit. These vehicles are designed to also safely reenter with cargo (and sometimes crew) so they can't deliver as much payload to orbit relative to a system that only launches satellites. This leads to higher apparent per-kg costs.

To arrive at a cost-versus-year trend for the Space Shuttle I assumed that all Space Shuttle missions were ISS missions, and then plotted the total amount of payload delivered versus the amount of money that was spent on the program, using 2023 inflation-adjusted dollars.

enter image description here

This approach gives us the plot shown above in Figure 3. By fitting a curve to the data in Figure 3 and plotting the inverse slope of that curve versus time, we end up with the orange curve of a USD-per-kg versus year shown in Figure 2.

enter image description here

We can see that the Space Shuttle’s cost-per-kg came down over time, but we can also see that there are a few peaks and valleys in the curve. The peaks were caused by the Challenger and Columbia disasters. Between these events, cost dipped to as low as 59,000 USD/kg.

If we apply the same methodology to Commercial ISS resupply services in general (blue curve) and the cost with the market leader, SpaceX (purple curve), using data from usaspending.gov, we observe that in practice the costs have only recently managed to achieve cost parity with the Space Shuttle Program. An Aug 2018 independent audit from the Office of Inspector General, on page 27, projected that the cost of commercial ISS resupply with SpaceX would be 71,800 USD/kg, which aligns well with the usaspending.gov data shown in Figure 2.

Other corroborating data includes an Aug 31st, 2022 press release by NASA where they announced that they had awarded five additional missions to SpaceX at a cost of $\$$1.436 billion or $\$$287 million per mission. This places the cost of future resupply missions 10 through 14 at 1,436,438,446 USD, which works out to 86,794 USD/kg if we assume that each mission delivers the maximum payload of 3307kg to the ISS. On March 8th, 2023, Robyn Gatens, Director of the International Space Station at NASA, stated informally during a Q&A session at the IEEE Aerospace Conference that half of the ISS budget goes to launch costs. As NASA spends roughly 3B per year on the ISS, and resupply runs deliver people and cargo at a rate of roughly 20,000kg per year, this works out to...

$$3,000,000,000 USD / 2 / 20,000kg = 75,000 USD/kg$$

The paper also traced the source of the Falcon 9 information on the Visual Capitalist chart (published 2 years ago on January 27, 2022) to an estimate made back in 2016 by The Tauri Group. If you're interested in those details, you can find the full paper by searching for it on Google Scholar. But the Falcon 9 information was also for launching satellites, not resupplying the ISS.

Regarding the Falcon Heavy data point, the article makes the following observation...

If Falcon Heavy were cheaper than Falcon9 as the chart suggests, then one would expect SpaceX to use Falcon Heavy to launch its Starlink satellites. In practice, the Falcon Heavy may be less reusable than the Falcon9 because the Falcon Heavy has not yet demonstrated the ability to land its core stage downrange on a drone ship. SpaceX’s actions may reflect an internally held belief that using Falcon Heavy would be more expensive than using Falcon 9 for launching Starlink satellites.

So far as the US government and its taxpayers are concerned, we are not observing that spaceflight costs are trending down.

Of course, the Space Shuttle is getting creamed by SpaceX's rockets in the court of public opinion. However, a more considered engineering analysis shows that if you consider the class of missions that involve sending a vehicle to a LEO rendezvous and having that vehicle return to Earth safely, the Shuttle was a cost-effective launch system.

  • $\begingroup$ I have a couple points of criticism. First of all, the graphic presented by the Asker does not concern itself with cargo delivered to the ISS, but rather just into orbit. The end-to-end cargo Dragon service offered by SpaceX doesn't only have the launch to space, but also many other services that SpaceX provides. Such as the capsule, reentry/recovery, on-orbit-monitoring, etc. Shuttle was also used to launch satellites, which F9 can do without Dragon, eg the GPS III sats. Here, the gov't paid \$82.7 million to put a 3880 kg satellite up, so at \$21,314 / kg. $\endgroup$
    – Dragongeek
    Commented Feb 26 at 12:25
  • 1
    $\begingroup$ My second point of criticism is you are comparing a "cost" (Space Shuttle) to a "price" (Falcon 9). This isn't really a proper comparison, because the cost is what it literally cost NASA to design/build/launch/operate the SS, while the price reflects more "how much money can I squeeze out of NASA?" on the side of some SpaceX accountant. There's a healthy profit margin in there, which SS didn't have. If you are a private customer, you can purchase a F9 to orbit (and even back) for much less than NASA is willing to shell out, and people have done this (eg Isaacman). $\endgroup$
    – Dragongeek
    Commented Feb 26 at 12:33
  • $\begingroup$ The number I always heard thrown around when I worked at NASA in 2019 was $10K/lb. This article from 2022 (much more recent than any Shuttle flight) has numbers more in line with that: www.nbcnews.com/news/amp/rcna23488 $\endgroup$
    – Doresoom
    Commented Feb 26 at 12:41
  • $\begingroup$ @Doresoom There's a big difference in delivered payload between launching a satellite and sending payload in a vehicle with return-to-Earth capability. This deleted answer covers several things to be aware of when using cost-per-kg as a performance metric. $\endgroup$
    – phil1008
    Commented Feb 26 at 16:11
  • $\begingroup$ To clarify, the $10K/lb number was cargo delivered to the ISS via SpaceX Dragon or Orbital ATK Cygnus. That didn't include return to earth via Dragon. $\endgroup$
    – Doresoom
    Commented Feb 26 at 16:25

Before I start writing about why spaceflight is getting cheaper, I think it should become clear where the price is coming from. NASA has been getting billions of dollars to build and launch rockets, but where does that money go?

The main cost is to build the rocket. According to this site https://www.eclipseaviation.com/how-much-does-nasa-pay-for-rocket-fuel/#0 NASA only spends around 200 thousand on rocket fuel per launch. The remaining money is used to build the rocket. This site https://mashable.com/article/nasa-rocket-cost-sls-artemis-space here claims that it took around 3.8 billion to build each Saturn V rocket for the final Apollo missions. The Apollo program cost around 200 billion dollars (The 200 billion wasn't just for the rocket).

SpaceX has been reusing rocket parts. On their webpage https://www.spacex.com/launches/ it says that they relaunched their rockets about 110 times. Currently a single launch with the Falcon 9 rocket costs about 60 million dollars. However the second stage is not being reused so it still costs a fair bit. Starship is fully reusable and is expected to cost around 1 million per launch. Theoretically it could cost only 200 thousand per launch if you only count the price of the rocket fuel. Starship will be carrying up to 1000 people and meaning that a ticket on Starship will probably be a bit more than a 1000 dollars.

Newer technologies are making things smaller and lighter, however the main reason is that spaceflight is becoming cheaper is that the rockets are being reused.

  • 1
    $\begingroup$ Elon Musk always uses the airline industry as an example: imagine, you threw away the airplane and built a new one from scratch every time you land. That would be insanely expensive. And now imagine that instead of having tens of thousands of flights per day, you only had a couple of flights per year, so you cannot even achieve the economic benefits of mass production. The only way to truly bring cost down is full and rapid reusability. (Where "rapid" is a clear reaction to the Space Shuttle program.) $\endgroup$ Commented Aug 5, 2022 at 13:04
  • $\begingroup$ I guess it's not that clear, but I'd like to focus more on history than on current events. I understand the reasons for reusable vehicles being cheaper than expendable ones, but I'd still like to compare how similar vehicles prices' evolve; and with SpaceX financial data being their private business it's hard to evaluate Falcon launches' cost. $\endgroup$ Commented Aug 5, 2022 at 15:50
  • $\begingroup$ @DanilaSmirnov There are links to reputable sources of data on Falcon9 plus Crew Dragon launch costs in this answer: space.stackexchange.com/questions/10350/… $\endgroup$
    – phil1008
    Commented Dec 14, 2023 at 9:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.