What is the -main reason- for Falcon 9 being more cost-effective?

[Pipcard]

The expendable Falcon 9 is already very cost-effective compared to other launchers in its class. But what is the primary reason for this?

- Is it because of vertical integration/ (most) parts being manufactured "in-house"?

- Is it because of commonality in fuel types, engines (mostly), and stage manufacturing? Musk claimed that "to a first-order approximation," three different fuel types "triples your factory costs and all your operational costs." But how accurate is that?

- Or is it because of cheap labor? (i.e. longer hours for the same or less pay compared to other launch vehicle manufacturers in countries with similar wages)

- (or something else?)

[QuantumG]

It's because SpaceX made lowering the cost of spaceflight the primary goal.

[Kabloona]

I'm not convinced cheap labor is the "main" factor. There are other launch vehicle companies where engineers work crazy unpaid overtime too (Orbital was that way when I worked there.) And the LA area is extremely expensive to live in, so there's a limit as to how little SpaceX can pay their workers.

IMO, the following quote from former SXer Max Vozoff is more relevant:

There's a YouTube video of Elon speaking somewhere in 2003 saying ...  "we're really just a systems integrator, we're buying things from other people", but by the time I showed up in 2005 that had completely turned around and pretty much everything was getting done in-house. 

And you can see why when you see the interactions with these suppliers, particularly the ones in the space industry.  They think they're the only ones who can make this widget or who have the secret sauce, and when you say "no, you're too expensive", they say "well, that's what it is". And they're used to customers who, if they slip the schedule and double the price, the customer shrugs and goes back to headquarters and says, "well, it's gonna take twice as long and it's gonna cost twice as much", and that's how things go in a traditional government run program.

But SpaceX would say "no, that's not acceptable", and they'd cancel the contract.  And sometimes these suppliers were literally scoffing on the phone as you hung up, and call you back a few months later saying "so, have you changed your mind yet?"  And being able to say to them that "no, if you can do it, then maybe somebody else can do it too", like either SpaceX figured out how to do it themselves, because they hired some smart people and gave them the resources and tools, or you find another supplier with maybe a non-space version and you upgrade and qualify it for space.

And now what you've done, this backward supplier has bred a competitor for themselves, where they're not used to competition.  I mean, many of the suppliers in this industry would just go out of business in a heartbeat if competition were actually introduced.

So really that's the game changing stuff that SpaceX has been doing: bringing stuff in-house, not just because it gives them control of cost and schedule, but because the space suppliers, traditional suppliers just don't get it.  They're not used to being held to schedules and budgets.

And that's not true of everybody, but there is list of anecdotes I could tell you about suppliers with this attitude.  And in each case either SpaceX brings it in-house and makes it successfully, or they find another supplier and upgrade it, and that supplier is usually thrilled to have a whole new market opened up for them.

The above quote was posted in another thread by Dave G from an interview of Max Vozoff here:

http://thespaceshow.wordpress.com/2011/03/05/max-vozoff-friday-3-4-11/

Having worked on launch vehicles myself, I can verify the truth of what Vozoff said about aerospace suppliers. It can be extremely costly to buy a batch of specialized components because of the "secret sauce" trap. There may be only a few suppliers of the specialized item, and it's extremely costly for a competitor to develop and qualify a similar product, so there isn't a lot of competition to drive costs down. And the fewer components you buy, the more expensive they are, because production setup and lot acceptance test costs get amortized over fewer units.

SpaceX's ability to design, manufacture, and test so many of their components in-house at a much lower cost than a traditional supplier can do it for is, IMO, their biggest leverage in reducing costs (aside from reusability, if they can make that work).

[Comga]

How about a bit of each of these?

They have a relatively simple design.

They may pay less for similar work.

They may or may not be selling at a price that will make a profit at a sales level they may or may not achieve.[/size]

They have definitely developed a workable engine for much less than other domestic sources. Partly that's because it's relatively simple, LOX / RP1 turbopumped, chosen for simplicity over absolute performance.

Whether building everything in-house saves money is a question we cannot answer.  It runs against much of the business mantra of the current era, where outsourcing everything possible is seen as the way to minimize investment and maximize utilization.  However, it does cut down on interfaces, the writing of contracts, the paying of bills, the shipping, etc.

And there is the stuff Vozoff says. I have done it myself, finding commercial products and working the qualifications to get product that wouldn't normally be considered adequate for space but uprating it the minimum amount necessary.  I drove a multi-million dollar mechanism plan down to a few tens of thousands of dollars of solid state hardware.  It can be done.

Personally, I am going to bet the farm (professionally) on them achieving some good fraction of their immediate (Earth orbit) goals.

[daveklingler]

The expendable Falcon 9 is already very cost-effective compared to other launchers in its class. But what is the primary reason for this?

- Is it because of vertical integration/ (most) parts being manufactured "in-house"?

- Is it because of commonality in fuel types, engines (mostly), and stage manufacturing? Musk claimed that "to a first-order approximation," three different fuel types "triples your factory costs and all your operational costs." But how accurate is that?

- Or is it because of cheap labor? (i.e. longer hours for the same or less pay compared to other launch vehicle manufacturers in countries with similar wages)

- (or something else?)

You could start with the fact that the preponderance of space contractors often boast of how many different states their projects come from, and how many separate subcontractors.  Big contractors help sell projects to Congress by spreading the money around as many different congressional districts as possible.  I've seen ads in which contractors talk about a project being spread between 47 separate states.  I am not exaggerating.

Much ridicule was generated during the nineties when the DoD could not conduct an audit (they still can't), but Lockheed Martin regularly told every member of Congress much they were spending in each district down to the penny, and on what project.  More recently, Sierra Nevada actively advertised how many subcontractors they were hiring for Dream Chaser, and in how many states.  That's a feature for a Congressman who wonders what a contractor will do for his reelection.

I had personal experience with a rocket project which required a total of 5 people, one of them working part-time, for construction and flight qualification, on top of which the contractor placed 13 levels of mid-level management, driving the price from $30,000 to over $1M.

Eliminating the job provider function from government contracts can really make an enormous difference, in SpaceX's case about a factor of 10.

[Machdiamond]

They key is that SpaceX is a startup.
I have experienced both startup and large corporation environments, in both cases working long hours, and identical engineering tasks cost ratio could easily reach the ludicrous(tm) factor of 100 if not way more - I kid you not.
Cheap labor and long hours has nothing to do with it. Actually it is the large corporation that gets the cheap labor rates because of the job risk factor, initially.
Now SpaceX has the added advantage that they are a talent magnet, this is also extremely cost effective.

[Coastal Ron]

I disagree.  Direct labor costs in manufacturing are not a big factor these days in general due to a number of reasons, and the factory SpaceX has set up for Falcon 9 is not excessively labor intensive.

If anything because of the commonality between the 1st and 2nd stages (i.e. diameters, tooling, fuel, engines, etc.), it could be argued that a Falcon 9 requires far less touch labor for the same assemblies than a comparable launcher like Atlas V.

[Dante80]

They key is that SpaceX is a startup.

SpaceX has been alive for more than 5,000 days now.
Unless I misunderstood the post and you are trying to describe the silicon valley mentality the company tries to keep.

[Jim]

I disagree.  Direct labor costs in manufacturing are not a big factor these days in general due to a number of reasons, and the factory SpaceX has set up for Falcon 9 is not excessively labor intensive.

If anything because of the commonality between the 1st and 2nd stages (i.e. diameters, tooling, fuel, engines, etc.), it could be argued that a Falcon 9 requires far less touch labor for the same assemblies than a comparable launcher like Atlas V.

Not true.  There isn't much difference in the vehicles.  The common diameter is a minor contributor.

It is the age of the workforce and pay.  I know an analyst (a non technical support person) who is heritage McDonnell Douglas that is making over $80k, will get a pension with a 401k plan. 

[pathfinder_01]

Not true.  There isn't much difference in the vehicles.  The common diameter is a minor contributor.

It is the age of the workforce and pay.  I know an analyst (a non technical support person) who is heritage McDonnell Douglas that is making over $80k, will get a pension with a 401k plan.

I would agree with this one, but I think it is more than just labor. China and Russia are cheaper sources of labor and ESA's rocket is an bit more subsidized but somehow despite all of this Space X can come up on top. Not to make an point but Atlas and Delta have an hard time getting non governmental launches.

[Nomadd]

 How much would SpaceX pay a contractor for the turbopump in the Merlin? Russia is cheap labor.

[Coastal Ron]

I disagree.  Direct labor costs in manufacturing are not a big factor these days in general due to a number of reasons, and the factory SpaceX has set up for Falcon 9 is not excessively labor intensive.

If anything because of the commonality between the 1st and 2nd stages (i.e. diameters, tooling, fuel, engines, etc.), it could be argued that a Falcon 9 requires far less touch labor for the same assemblies than a comparable launcher like Atlas V.

Not true.  There isn't much difference in the vehicles.  The common diameter is a minor contributor.

Here is a picture of the aft end of an Atlas V:



If you look closely at the aft end you'll see how much labor is expended doing riveting and what looks like bolts or some other type of attachment hardware.  That is a LOT of manual labor to build that section.  The Falcon 9 aft end is not as complicated.

It is the age of the workforce and pay.  I know an analyst (a non technical support person) who is heritage McDonnell Douglas that is making over $80k, will get a pension with a 401k plan.

What is the salary of the equivalent person at SpaceX?  And are you implying that SpaceX does offer a 401K plan?

And how does that equate to $Millions of dollars in manufacturing costs difference between Atlas V and Falcon 9?

[meekGee]

Whenever a new company does something better, the established players invariably and without fail label it as a "sweatshop".  Instead of looking inwards, place blame elsewhere.

This particular "sweatshop" happens to be located in California, where cost of living is high, as are salary expectations.

The main factors were:

- Vertical integration, so direct supervision over costs and elimination of profit margins three layers away.
- Vertical integration, so agile development.  No quarreling over who has to pay for which change order. The proof is in the pace of development.
- Vertical integration, so system engineering extends from top-level design to individual components.  No political decisions.
- Smart engineering.  VTVL, size and number of engines, type of fuel, reuse architecture.  Folks can claim SpaceX got lucky, but if you make the right choice time after time after time, there may be a method to your madness.
- Smart business strategy.  Getting development flights out of commercial flights - less cost to amortize into the per-flight price
- Cheap infrastructure.  Again - less fixed costs to amortize into the per-flight price.

[Coastal Ron]

The expendable Falcon 9 is already very cost-effective compared to other launchers in its class. But what is the primary reason for this?

- Is it because of vertical integration/ (most) parts being manufactured "in-house"?

Depending on the design, building what you need in-house can be a big savings.  At one company I worked at we won a large DoD contract for portable electronic systems, and we designed a small factory to build them.  One of our consultants advocated that we outsource the assembly of the unit to the company that was building the housing, but their bid came back significantly higher than what we knew we could do it for.  Part of the difference, obviously, was that they had to tack on 15%+ for profit, but they also didn't own the design so they have to bid conservatively.

And as others have mentioned, when you only buy in small quantities it doesn't make sense to buy from more than one supplier, so that one supplier (and the other competitors) can bid high.  Plus there are cost advantages to having quicker turnaround when you have design issues during development, or issues during production.  But volume really dictates whether this makes sense overall, and SpaceX does plan to have good volume.

- Is it because of commonality in fuel types, engines (mostly), and stage manufacturing? Musk claimed that "to a first-order approximation," three different fuel types "triples your factory costs and all your operational costs." But how accurate is that?

One component of that is the procurement cost of all those engines, and the overhead cost of maintaining the inventory for all the supporting parts.  Only having basically one engine with two variations greatly reduces overhead and carrying costs.

- Or is it because of cheap labor? (i.e. longer hours for the same or less pay compared to other launch vehicle manufacturers in countries with similar wages)

Los Angeles, which has more manufacturing than any other region of the U.S., is not a low cost location.  On BestPlaces.net, they rank the cost of living rating for Alabama as 89.5 against a U.S. average of 100.0, so less costly than the average across the U.S.  For California the cost of living rating is 198.4, or twice as high.

So it's unlikely that SpaceX is paying significantly less for labor than ULA is.

- (or something else?)

More so than any of the above, QuantumG put it nicely - SpaceX made lowering the cost to access space as their goal.  Removing or reducing costs has to happen during the design phase, because once you've designed your product the costs are pretty much locked in.

Elon Musk did a great job of understanding the cost factors of rockets, and was able to create a solution that addressed those cost factors.  He didn't succeed at first, and there is still work to be done, but he understood what needed to happen.  Plus, and I think this is a very important factor too, Musk could control everything within SpaceX, so they had clarity of purpose.  I'm not sure ULA has the same ability in that department, since Tory Bruno doesn't control his own destiny.

[TomH]

Not saying these are the most important, but they are a couple of factors, and numerous factors all add up.

-A lot of work had already been done on the Merlin design. SX refined it into a fairly simple design that also has the highest T/W

-In that Merlin is relatively small and many engines are needed, the production line stays busy and productive.

-F9 S1 is the largest stage that can possibly be transported by truck, avoiding higher cost transport by other methods.

When every last thing you do is designed for cost efficiency, it tends to make you competitive.

[deltaV]

- Smart engineering.  VTVL, size and number of engines, type of fuel, reuse architecture.  Folks can claim SpaceX got lucky, but if you make the right choice time after time after time, there may be a method to your madness.

All of the design decisions you quoted differ from SpaceX's previous plans. They initially intended to sell Falcon 1 and Falcon 5 but ended up with Falcon 9. Their next big rocket was initially planned to use kerosene or hydrogen before they settled on methane. They initially considered parachute recovery of first stages before switching to VTVL. Evidently they aren't particularly good at making the right choice the first time. Their strength is the courage to (1) try new things and (2) abandon what didn't work.

[meekGee]

- Smart engineering.  VTVL, size and number of engines, type of fuel, reuse architecture.  Folks can claim SpaceX got lucky, but if you make the right choice time after time after time, there may be a method to your madness.

All of the design decisions you quoted differ from SpaceX's previous plans. They initially intended to sell Falcon 1 and Falcon 5 but ended up with Falcon 9. Their next big rocket was initially planned to use kerosene or hydrogen before they settled on methane. They initially considered parachute recovery of first stages before switching to VTVL. Evidently they aren't particularly good at making the right choice the first time. Their strength is the courage to (1) try new things and (2) abandon what didn't work.

Actually, choosing an engine size that allowed them to develop it on  F1, then progress to F5 and F9 is exactly the kind of smart choice I'm talking about, and it gave them the ability to develop on the cheap, and remain agile afterwards.

For BFR, they looked at several choices, but made the right decision.  If there were no alternatives, it wouldn't be a choice, now would it.

Propulsive RTLS was the plan from a long time before you realize it. Parachutes were something they wanted on early F9s in order to prove (maybe to themselves as well) that the airframe can be reused.  It never worked, and they sure dropped it fast enough.  Don't think for a minute that propulsive RTLS came up only after they gave up on parachutes.

So yeah, smart choices, and then agility.

[Darkseraph]

It is hard to say since SpaceX is not a public company and is under no obligation to publish its financials. For all we know they have been operating under razor thin profit margins. We don't know how profitable they really are. The Chinese may in fact be telling the truth in that they couldn't believe SpaceX prices. These very well may be too good to be true and part of their current business model is to undercharge to gain market share as fast as possible then jack up prices later. The fact they went to so much effort to gain a toehold in the lucrative millitary launch contracts business even though the launch rates are not that high tells me they feel it is very important for their future profitability. The military pays more than commercial commsat operators and international competitors are not able to compete for those contracts. It seems naive to believe they got into a very public spat with ULA because of a deep held conviction to save the tax payer billions. They outspent ULA in lobbying last year according to a SpaceNews article.

Overworking and burning out young staff has been said to be a factor by several sources and there are on-going lawsuits about this off-the-clock work.


Vertical integration has become almost a mantra as an explanation for their success but I am not convinced it is the critical reason. Though you can cut out price goughing from suppliers by making your own components, it comes with very high capital costs and if you are in a low volume business, this means you're running expensive equipment under capacity and not spreading costs among multiple users. It is not a fix all solution and the fact OrbitalATK charge similar prices for their cargo contract despite pursuing the exact opposite strategy (outsource almost everything) leads me to believe it is not the most important factor.  Lacking access to their books, it is hard to evaluate how important in-house production is to their internal costs.


I don't want to sound overly negative or like I am bashing SpaceX in favor other companies. I'm incredibly impressed by their achievements so far and can only cheer on these accomplishments. I don't think Elon got into the rocket business just to make a buck..there are more lucrative businesses to get into. It sounds to me that he's being honest that his vision is to lower the cost of spaceflight and enable a city on Mars. But I think it is possible he wants to achieve that so much he is willing to be ruthless, dishonest and underhanded to achieve that end. He is not a saint and you wouldn't survive very long in this business if you were!

[BobHk]

- Smart engineering.  VTVL, size and number of engines, type of fuel, reuse architecture.  Folks can claim SpaceX got lucky, but if you make the right choice time after time after time, there may be a method to your madness.

All of the design decisions you quoted differ from SpaceX's previous plans. They initially intended to sell Falcon 1 and Falcon 5 but ended up with Falcon 9. Their next big rocket was initially planned to use kerosene or hydrogen before they settled on methane. They initially considered parachute recovery of first stages before switching to VTVL. Evidently they aren't particularly good at making the right choice the first time. Their strength is the courage to (1) try new things and (2) abandon what didn't work.

And if they had been old space theyd have done EXACTLY what they initially intended and been out of business by now.

Thank goodness they went the way they did operating from first principles.

[QuantumG]

Propulsive RTLS was the plan from a long time before you realize it.

Dunno what your information is, but both Jon and I have first hand accounts of SpaceX engineers saying they would never do vertical landing as late as the second half of 2010. Rumor has it that Masten's in-air relight of Xombie (May 2010) was the catalyst for Grasshopper and skepticism about vertical landing continued internally until the first flights in 2012.