SpaceX has been launching and reusing their Falcon 9 for very cheaply quite often. Even though it is quite revolutionary, it is isn't the first time a rocket has been reused. The Space Shuttle side boosters (aka. solid boosters) were also reused and relaunched. Why is it the case that the Falcon 9 is known as one of the cheapest rockets ever because it reused the first stage and the Space Shuttle is known to be one of the most expensive rockets because it reused the side boosters. Isn't this contradictory?

  • 3
  • 1
    $\begingroup$ Fundamentally, the Shuttle was designed for the best performance feasible. The Falcon 9 is designed to the best price per pound to orbit. Very different optimization criteria. $\endgroup$ Apr 14 at 4:37
  • 1
    $\begingroup$ I'm not sure the recovery and reuse of the SRBs was ever actually worth the costs except as PR. It was only ever slightly more valuable to recover them, clean them out, and reset than to make brand new one. The Falcon 9 is much more economically sensible; it's far cheaper to refurbish and refuel a Falcon than to build a whole new liquid fueled rocket... but the problem is they had to give it a landing barge and work out controlled descent (you can't dunk a liquid fueled engine in seawater and expect it to work ever again). $\endgroup$ Apr 16 at 18:10

5 Answers 5


A few factors:

First, the shuttle had extreme capability and performance requirements that Falcon 9 didn't have, for example:

  • Recovery of a 100-ton vehicle from orbit, with a reusable thermal protective system that was extremely expensive to restore after each flight.
  • Delivery of 25 tons of payload to LEO in addition to a crewed vehicle with weeks of on-orbit endurance.
  • Ability to reach landing sites up to 1000 miles off its orbital path
  • The main engines were designed at the very edge of technical capability of the late 1970s, explicitly to push the state of the art in large rocket engine technology.

By comparison Falcon 9 doesn't recover its upper stage at all and it has a modest payload; it's about a quarter of the mass of the shuttle. Its engines are conservative in their basic design, though the execution of the design is sophisticated, and they have refined, simplified, and uprated the engine over several versions, increasing the capability of the rocket, instead of starting with the most aggressive possible design.

Second, the challenges of meeting the shuttle's capability goals were compounded by compromises in the funding; not enough money was slated up front for development, so instead of a fully reusable flyback booster, solid rocket boosters -- relatively cheap to develop and build, but almost not worth the expense of reusing/rebuilding -- were chosen. In comparison, SpaceX developed a much smaller and simpler rocket, and by making it economically viable in the expendable mode, they were able to get paying commercial customers to fund the development of reusability.

Third, the shuttle, for both political reasons and due to its size and complexity, had to be built and maintained by several companies, which increased management overhead and therefore cost. Rockwell built the orbiter, Rocketdyne built the engines, Thiokol built the boosters, Lockheed Martin built the external tank, with NASA having to coordinate everything. In contrast, SpaceX is one company, and they have to make Falcon 9 competitive in the commercial launch market -- more than competitive, in fact, since they need to compete with established launch providers. SpaceX, in addition, is privately held; unlike NASA's public-company contractors, they do not have to demonstrate high profit margins to shareholders in the short term (and in fact, they appear to have operated at a loss in 2022, though Starship development costs were certainly the major factor there).

  • 12
    $\begingroup$ Yes, your first point is a big part of it. Falcon is only a launch vehicle, and with only a reusable booster... the (relatively) easy part of re-use. Shuttle was vastly more ambitious. Of course, Starship is a whole different matter, but that's yet to prove itself. $\endgroup$ Apr 9 at 21:38
  • 6
    $\begingroup$ I remember at the time the Shuttle was meant to be a "cheap" and "routine" access to LEO like catching a commercial airliner - the whole point of the "Shuttle" name - up to 60 launches per year and replacing all current US launch systems. The intention was to be what SpaceX has become. Plus it was to be a Swiss Army knife, able to do practically any mission. Here's a good history of aspirations vs reality: spaceline.org/united-states-manned-space-flight/… $\endgroup$
    – Galerita
    Apr 10 at 4:49
  • $\begingroup$ And the strap on SRBs always made the Shuttle look like a ridiculous bastard $\endgroup$
    – Galerita
    Apr 10 at 6:44
  • 11
    $\begingroup$ The last factor is that once NASA had a barely working highly labor intensive vehicle, Congress declared victory and kept flying it as it for several decades. Had they been able to iteratively improve it operational costs probably would have gone down significantly while performance improved. The single biggest and most obvious change would have been, after the Air Force dropped out post-Challenger, to drop the extreme cross-range landing requirement and slash the wing size; with immediate savings in total orbiter weight and the size of the TPS to maintain. $\endgroup$ Apr 10 at 8:42
  • 3
    $\begingroup$ Ah, yes, the "steal Soviet satellite" requirement. $\endgroup$ Apr 10 at 19:19

There has been about a 20-fold reduction in launch costs vis-a-vis NASA since SpaceX began using the Falcon 9. The Falcon Heavy is yet cheaper and Starship will be cheaper again in costs per kg to LEO.

Consider the following diagram: Recent changed in launch costs since SpaceX operations

The diagram is from The Recent Large Reduction in Space Launch Cost written by Harry Jones of NASA Ames Research Center to accompany a conference presentation in 2018. It's in "current dollars", but the paper is 6 years old, so the analysis may have changed slightly. NOTE the y-axis is in kilo-dollars per kg not dollars per kg.

(Also see this Our World in Data page.)

From the abstract of Jones' paper, "NASA’s space shuttle had a cost of about USD 1.5 billion to launch 27,500 kg to Low Earth Orbit (LEO), USD 54,500/kg. SpaceX’s Falcon 9 now advertises a cost of USD 62 million to launch 22,800 kg to LEO, USD 2,720/kg. Commercial launch has reduced the cost to LEO by a factor of 20."

The analysis is quite detailed so it's worth reading the entire paper, "The possible technical approaches to cut launch cost have been assessed as follows:

  1. Simplify the vehicle configuration
  2. Increase vehicle production and launch rates
  3. Use industrial design and production methods (cultural change)
  4. Optimize for minimum cost
  5. Reduce the parts count
  6. Increase simplicity and design margins
  7. Reduce instrumentation
  8. Design for production and operation"

In contrasts with NASA, "The technical causes of high launch cost have been assessed as follows:

  1. Goal of maximum performance and minimum weight, originally from ballistic missiles
  2. Higher cost of expendables versus reusables
  3. High cost of human spaceflight
  4. High cost of new technology, hardware, and software
  5. Low failure tolerance and consequent intense design effort and detailed oversight
  6. High system complexity, parts counts, and number of interfaces"

SpaceX itself attributes the low costs to four key factors:

  1. Smaller workforce
  2. Use of in-house development
  3. Fewer management layers and less infrastructure
  4. Commercial development culture

The most competitive launch systems with SpaceX are from countries not in the good books of Western democracies: Long March (China) and Proton/Angara (Russia); both of which cost around $4000/kg to LEO.

To me the most important reason for SpaceX's low cost is it is the ONLY orbital system to FULLY REUSE their boosters.

Other systems have been partially reusable. For example, but NASA's attempts with their solid rocket boosters have been clumsy and expensive. And complicating this they are mixing (reusable) SRBs with liquid fueled rockets for other stages, with the giant first stage being thrown away after each launch. They've continued with legacy systems rather than innovating. You would think they'd have learnt by now.

I'm sure there are more reasons than those given in the paper (above), but fundamentally NASA has become something of a dinosaur, set up for relatively few specialist missions (volume matters!), and now turns to SpaceX for many routine operations. The reduces NASA's overall costs but increases the cost of their own launches since there are so few.

It's questionable how much of a role NASA will have in launch and space travel in the future, although it will have a role in overall mission design and payload packages, such as landers and deep space probes.

  • 8
    $\begingroup$ Oof, that chart is bad. $\endgroup$ Apr 9 at 20:33
  • 3
    $\begingroup$ @Russell Borogove Yes the plot is awful. I'm in the research business myself. I would have expected better from a NASA researcher. The Our World in Data plot is better, but it doesn't include the analysis in the paper. ourworldindata.org/grapher/cost-space-launches-low-earth-orbit $\endgroup$
    – Galerita
    Apr 10 at 4:31
  • 7
    $\begingroup$ @Michael It's bad because there are a lot of dots without labels. And what is the point of the first two regression lines? One going steeply from Vanguard to Saturn V. (And which dot is Saturn V?) And another flat line below the Space Shuttle. This makes no sense because there seems to be progress from 1990 to 2000. It's just SpaceX suddenly leapt ahead of the competition. The 3rd line links Falcon 9 to Falcon Heavy, so OK. The point made by the graph is clear, but context is lost in a mess of other information which is not explained. $\endgroup$
    – Galerita
    Apr 10 at 14:58
  • 3
    $\begingroup$ The paper that you referenced in this answer managed to capture the attention of Thunderf00t because its accuracy is debatable. There are several other answers on this site with the cost-to-orbit tag that reference more reputable sources of information. $\endgroup$
    – phil1008
    Apr 10 at 18:17
  • 2
    $\begingroup$ Oh, the Our World In Data plot is much better. It's somewhat remarkable that the Saturn V wound up being one of the most cost-effective launchers for such a long time; makes the alternate timeline with INT-19/20/21 and a flyback booster version of the first stage even more interesting. $\endgroup$ Apr 10 at 20:31

The Space Transportation System (Space Shuttle) or STS as I'll refer it to in this post, is more expensive than the Falcon 9 in recovery for a bunch of reasons. Here are some that I think influence the price the most.

  1. It's not Reused, it's REFURBISHED. The STS is technically not reusable if you account into how much maintenance goes into the orbiter every time it flies. This is about 600,000 hours of maintenance for 1 flight. Includes inspecting heat shield tiles, engine damage, etc.
  2. Cost-plus contracting. Since the space shuttle is built by contractors and assembled by NASA, it increases the refurbishing cost significantly as with cost-plus contracting, the manufacturers don't need to save money. On the Falcon 9 side however, SpaceX is a private company owned by Elon Musk, which means it needs to generate revenue, because of this, they try to save money in every way, whereas NASA, is like a guzzler of money. Even today, NASA's SLS rocket (Space Launch System) is way over budget already. Just look at the engine costs alone: The spaceX Falcon 9 uses 9 Merlin Engines, each costing about 2.2 Million USD (according to Google) . The STS orbiter engines alone (the RS-25) cost over 100 Million USD each (according to everyday astronaut). This significantly increases overall cost and in terms of spare parts, to replace a fuel valve on the Merlin 1D engine, it costs about 10,000 USD (according to Elon Musk's biography) . But for the STS RS-25 engine, it cost around 150,000 USD (according to aerospace parts suppliers). This also increases the refurbishing cost.
  3. The SRBs. When saying how much it costs to reuse the STS, most people forget about the SRBs, and there is no way to reuse them as you have to scrape all of the aluminum deposits (As the SRBs are APCP solid fueled motors), that is already a intensive process costing a lot of time and money. Next is separating the segments. Over 50 bolts are removed per segment to separate them. Which is also a labour intensive process. The most expensive process is loading the solid fuel into literally empty steel tubes (yes, there is nothing else in the segments) and also refurbishing the nozzle. As demonstrated very well by a lot of Amateur rocket motors, even steel nozzled on a APCP motor, the nozzle melts slightly and deposited of aluminum weld themselves onto the nozzle. (That explains for the nozzle of the SRBs after launch when recovering looking like they have been tampered as there not smooth anymore. ) Refurbishing the nozzle also costs a lot, as well as all the gimbal units (Yes, the STS SRBs have gimbal) need to be inspected in the process. Also, since it lands in the sea, inspecting for seawater damage is also crucial, this is also a labour intensive process.
  4. It's cheaper to expend the SRBs instead. As shown by documents about the total cost for refurbishing the SRBs once, it exceeds the cost of just buying a new one by a lot. This is one of the main reasons why ESA only recovers the SRBs but does not refurbish them.
  5. The Falcon 9 does not use heat shield tiles. If you know the orbiter well enough, every flight, on average about 700 heat shield tiles fall off. These have to be inspected and replaced which is expensive and cost a lot of labour. As for the Falcon 9, it comes in engine first and does a re-entry burn (which looks beautiful from the ground like a colorful jellyfish) which protects it from re-entry heating.
  6. Re-entry velocity. If you know re-entry heating well, the faster you re-enter, the more heat you generate. Since the Falcon 9 booster or first stage (I'll refer it to as booster as SpaceX calls it) comes in at around 2000m/s (assuming ASDS ) . Compare that to the orbiters 7800 m/s (entering at orbital velocity) will mean more maintenance is needed, this in increasing costs.

I hope this has given you a good explanation for why the STS is more expensive to reuse than the Falcon 9. This is just a few of the many explanations.

  • 3
    $\begingroup$ Please could you provide sources for the numbers in your answer. $\endgroup$ Apr 10 at 7:59
  • $\begingroup$ Point 2 (cost-plus contracting) is important. Same principle as with (purely) military contracts and certain health care costs. $\endgroup$ Apr 10 at 23:46
  • 2
    $\begingroup$ "(That explains for the nozzle of the SRBs after launch when recovering looking like they have been tampered as there not smooth anymore. ) " Um, the nozzle was blown in half by explosive charges before water impact. $\endgroup$ Apr 11 at 1:41
  • 1
    $\begingroup$ "As shown by documents about the total cost for refurbishing the SRBs once, it exceeds the cost of just buying a new one by a lot. " Reference? $\endgroup$ Apr 11 at 1:42

The other answers provided all the fine details as to the exact differences but I'd like to focus on WHY the space shuttle was designed to cost more than necessary:

The spaceshuttle was a breakthrough research project in several areas built to (not commercially viable) military specs designed by hand. The Falcon 9 is an idea from the 70s finally implemented with modern CAD systems, computers and materials**(and specs designed to be economic from the getgo)

A lot of people forget that the SpaceShuttle program was not primarily a research (and even less a commercial) project. It was a requirement from the military to provide unprecedented launch cadence and carry mass (amongst other requirements). [well actually the space shuttle was a research idea already but the air force requirements were what kept it from cancellation and increased the specs heavily] Therefore economics weren't a major consideration in the design process. Which is why - as Lawn Hollander's answer makes clear - it's not reused but refurbished at insane man-hour costs. But that didn't matter because the goal was a high launch cadence, not to save money (which also fell through, but I'm focusing on the design aspects).

This focus on military uses (and then research) also meant that it was supposed to be highly configurable with a lot of functionality that the Falcon9 does not have (and for economic reasons absolutely does not want) like multi-week crew support, capabilities to collect satellites from orbit, high configurability and so on. The Space shuttle was a multipurpose orbital vehicle. The Falcon 9 is almost a pure mass-delivery platform. And all of these capabilities (even if it's just a modular mounting rail) have weight, design- and manufacturing costs attached to them.

On top of that: even with those requirements, designing (and building) a spaceshuttle equivalent today would be much much cheaper. Not only because of computer assisted design, but also because material science, computer simulations (requiring far less safety margins) and manufacturing capabilities massively improved. It would still not be able to compete with the Falcon 9, but could get a lot closer. And that ignores the massive amounts of research that the shuttle programs bill contained that is available for free to SpaceX engineers.

Another factor was that the space shuttle (at least in the design phase before all the issues cropped up) was designed to be quite economical already (for the time). Optimizing more and more (when you already think you're way ahead of the competition) would likely lead to delays and problems without any expectation of a relevant payoff.

So from multiple angles, there was no or little incentive to be especially economical. The Falcon 9 on the other hand was from start to finish meant to make money, to launch as cheaply as possible, to be an economic enterprise.

  • $\begingroup$ Do you have references for your claims? I'd especially be fascinated to see a reference for "It was a requirement from the military to maximize launch cadence in the case that an orbital arms race with the soviets could start." $\endgroup$ Apr 11 at 1:39
  • 1
    $\begingroup$ @OrganicMarble I overexaggerated the space-war aspect, but the air force reconnaissance requirements were a major driving factor (in enlarging the carrying capacity) history.nasa.gov/SP-4221/contents.htm in a sense having (a lot of) military stuff in space was still the goal [I also edited the post to reflect the facts more accurately, just in case you're wondering as I couldn't find that clear evidence for space-arms-race motivations] $\endgroup$
    – Hobbamok
    Apr 11 at 14:05
  • $\begingroup$ This answer brings to the front the question: So if a principal goal was higher launch cadence, and the design catered to that: Why didn't it happen? Looks like the US bought a worse-than-usual cadence at a very high price. $\endgroup$
    – davidbak
    Apr 12 at 20:19
  • $\begingroup$ @davidbak the refurbish time, (especially of the heat protective tiling) was MASSIVELY underestimated, so shuttle turnaround times were significantly above expectations. Also there was - in the end - no actual need for the high launch cadence, so it wasn't pushed to the limits. Or at least the military usecase (where money would've been largely irrelevant) wasn't fully utilized and NASA had to make do with their (significantly smaller) budget. On top of that, regular rockets remained cheaper so the shuttle was only used where necessary $\endgroup$
    – Hobbamok
    Apr 12 at 20:56
  • $\begingroup$ This increased refurbish time (or better: the maintenance that required the time) also drove up costs significantly above expectation) $\endgroup$
    – Hobbamok
    Apr 12 at 21:01

One of the main reasons for the cost difference is that when the mission requires the use of an orbiter, this adds cost and reduces the payload capacity to orbit. When the mission requirements can be met without using an orbiter, it is possible to reduce the cost-per-kg to orbit significantly.

To illustrate this better let's first let's consider some launch systems that enclose the payload within fairings which are jettisoned on the way up. Examples of such systems are Falcon 9 (when configured to launch a stack of Starlink satellites), or Falcon Heavy.

Next, we'll consider launch systems that enclose the payload within a spacecraft that travels to the destination and then deorbits and reenters the atmosphere to return to Earth. Examples of this kind of system include the Space Shuttle and Falcon 9 + Crew Dragon.

Enclosing The Payload Within Fairings

This earlier answer estimates the cost-per-kg to LEO, with Falcon-9-plus-fairings, in its autonomous-drone-ship-landing reusable configuration, at 7,680 USD-per-kg (before taxes and fees; 2024 USD).

Using an Orbiter

This answer (note: It is my earlier answer and it references a paper that I wrote, which in turn references reputable sources of data) provides a chart of the Space Shuttle's cost-per-kg versus year. Note that the chart assumes that the Space Shuttle was always loaded to its to-the-ISS payload capacity of 16,050 kg.

enter image description here

If we were to redraw the orange Space Shuttle's curve on the above chart with the to-LEO-payload capacity of 27,500 kg instead of the to-ISS-payload capacity of 16,050 kg, then the Space Shuttle's lowest cost to orbit would be \$58,920 * 16,050 / 27,500 = \$34,388 per-kg. If we correct for two years of inflation this is \$34,388*1.097 = \$37,724 in 2024 USD.

"With-Fairings" to "With-Orbiter" Comparison

So, Falcon 9 (with fairings) is 7680/37724*100% ~= 20% the cost of using the Space Shuttle to launch payloads to LEO. (Note: There are several reasons why people sometimes report numbers such as "20X cheaper" instead of "5X cheaper". If you read through the referenced answer it should help to explain why the 20X versus 5X discrepancy exists.)

"With-Orbiter" to "With-Orbiter" Comparison

Using the above chart, we can also compare the Space Shuttle to Falcon 9 plus Dragon (and other companies' systems) for the mission of delivering payloads to the ISS. This mission requires the use of an orbiter, since science payloads and sometimes crew need to be returned to Earth.

Now the Space Shuttle's cost varied due to startup costs and because it was temporarily taken out of service after the Challenger and Columbia tragedies, but at the best of times it achieved a per-kg cost as low as \$58,920 per-kg (2022 USD). Falcon 9 plus Dragon's lowest cost so far has been ~\$70,000 2022 USD). So, when you do more of an apples-to-apples comparison, the costs have not changed much. In fact, for the ISS resupply mission they've gone up slightly.

Hopefully this data helps to illustrate how much cost-per-kg can be reduced by deleting the orbiter and using faring instead. The orbiter increases mission complexity and reduces the amount of payload that can be delivered to the destination.

As for your question...

Why is it the case that the Falcon 9 is known as one of the cheapest rockets ever?

... this is an interesting question. If you simply want to be better informed on the topic of launch costs, I recommend this article. It does a very good job of examining and discussing many of the reasons that people have come up with to explain why launch costs are high. I would recommend reading it to become better educated on the topic. After that, I would recommend searching for questions on this site with the tag, and reading through some of the answers and the discussions in the comments.


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.