I want to say fuel is, but this Question Why isn't stainless steel used for rocket engines? has got me thinking about cost in the long run.

What part commonly found on most rockets cost the most to manufacture or gets replaced most frequently?

Technically the payload is not part of the rocket.

  • 2
    $\begingroup$ Do you include development costs? $\endgroup$
    – Polygnome
    Commented Apr 17, 2019 at 20:42
  • 1
    $\begingroup$ @Polygnome no. Maybe it should be measured in difficulty to make the part after it is developed? $\endgroup$
    – Muze
    Commented Apr 17, 2019 at 21:01
  • $\begingroup$ How do you measure that difficulty (in which unit)? $\endgroup$
    – Polygnome
    Commented Apr 17, 2019 at 21:04
  • 3
    $\begingroup$ A couple of hints for improving this question: With questions this broad, you need to specify you're looking for a general answer ("What is usually the most expensive part?"), otherwise people might get hung up on "which rocket do you mean?", as there is (presumably) quite a lot of variation between rockets. Also, if you want numbers, ask for Order of Magnitude approximations. What costs \$100,000? What costs \$10,000,000? The devil is often in the detail in these questions, and asking for approximate answers will reduce the difficulty for anyone trying to answer. $\endgroup$
    – Ingolifs
    Commented Apr 17, 2019 at 21:38
  • 2
    $\begingroup$ @Saiboogu from the question and comments I wasn't quite sure what was meant by "part." I guess re-reading the question now, my inference that production costs weren't supposed to be included was erroneous. Turning down the nitpick dial a bit. $\endgroup$
    – ben
    Commented Apr 19, 2019 at 19:16

2 Answers 2


I would say the turbopumps are the most expensive. They cost 3 million+ each, and are extremely vital in rocket engines. Falcon heavy uses 28 turbopumps. That is a whopping 84 million+ dollars, and falcon heavy launches cost 90 million+. So you can see why they want those rockets to be reusable to keep launch costs down. 1 turbopump is lost on every launch, obviously, but still with 81 million in turbopumps landing every flight that is a heck of a savings.

  • $\begingroup$ Why is a turbopump lost on every launch, even if the rocket is recovered? I think, if the rocket is not recovered, then all the turbopumps are lost. If it is, then only the economically not repairable ones, whose count can be between 0 and 28, but a near-0 is more likely. $\endgroup$
    – peterh
    Commented Apr 19, 2019 at 17:23
  • 7
    $\begingroup$ Second stage isn't recovered, and uses a single Merlin engine. I doubt Merlin turbopumps cost SpaceX $3M each, though; this thread floats numbers under $1M per engine. @Dan, what's your source for that? $\endgroup$ Commented Apr 19, 2019 at 17:29
  • $\begingroup$ You can't ask about typical/average rockets and then specify recovery - the two do not coexist at all at this time. There's the average expendable rocket, or there's the outliers at SpaceX. $\endgroup$
    – Saiboogu
    Commented Apr 19, 2019 at 18:00
  • 1
    $\begingroup$ And @RussellBorogove's link doesn't just float a number of that magnitude, it includes (if you listen or follow a link to the transcript) a quote from Elon, quoted by the VP of propulsion at SpaceX, putting a single Merlin at 'some fraction of a million dollars' - so the claims above about turbopump cost on SpaceX rockets is wrong. $\endgroup$
    – Saiboogu
    Commented Apr 19, 2019 at 18:16

The satellite

The most expensive thing is usually the satellite(s) source ranging in the hundreds of millions USD.

Rockets costs usually in the single hundreds of millions range.

The engines

The second most expensive thing is the engine(s). That's why plans have been made to recover them instead of the whole rocket source

  • 2
    $\begingroup$ Especially if the satellite is the JWST, at $ten billion plus. $\endgroup$ Commented Apr 18, 2019 at 14:22
  • $\begingroup$ @OrganicMarble Out of curiosity, do you have any idea how much it would cost to build a second JWST? $\endgroup$
    – gwally
    Commented Apr 18, 2019 at 19:38
  • $\begingroup$ No clue at all. $\endgroup$ Commented Apr 18, 2019 at 20:48
  • 3
    $\begingroup$ More generally, shall we say the payload is usually the most expensive, rather than the “satellite”? $\endgroup$
    – Paul
    Commented Apr 19, 2019 at 1:05
  • 3
    $\begingroup$ For aerospace development, the rule of thumb I'm familiar with is that initial R&D costs 10x the cost of building the first unit, suggesting that JWST #2 might cost only $1B, but that's when you intend to produce multiples from the outset, so maybe double the cost of #2 if you didn't plan for mass production. The other rule of thumb is that the 1000th unit costs 20% of the first unit cost, and that's probably roughly an exponential-decay curve. $\endgroup$ Commented Apr 19, 2019 at 17:27

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.