Why the preeminence of SpaceX?

Virgin Orbit announced Thursday that they were suspending operations, and it looks likely the company will go under or be majorly restructured. In fact, with one exception private space companies seem to have been doing poorly over the years: big on hype and dreams, poor on substance. If it weren’t for one exception most observers would say that private enterprise in space is still, and perhaps might always be, a pipe dream.

That one exception is of course SpaceX, which has succeeded spectacularly better than anyone else in the running. Not only surviving as a company but actually advancing the state of the art while improving the economy of space launch. So what is SpaceX doing right that other companies can’t seem to duplicate?

Virgin and Blue Origin both came out of the gate pretty clearly fixated on the “space tourism” market, while SpaceX has always been focused on the more serious space launch market for actual satellites and doing stuff in orbit.

A few rich people taking short jaunts into space might attract a lot of news and social media attention, but it doesn’t really pay the bills for full-scale orbital operations.

I agree. Musk made the right call in that regard. I wouldn’t have thought so because they were competing against NASA, ESA, and the Russian space program, and the gov’t has more resources than well anybody. The real question is:

  1. How is SpaceX managing to keep cost overruns mainly in check? (or are they really?)
  2. How is SpaceX managing to have such lost launch costs? (or are they really?)

It seem unlikely to be solely the reusable rockets. And of course SpaceX is a private company so nobody really gets to see what’s going on. I know there has been speculation that the secret sauce is burning capital and they are not as profitable as some believe. I honestly have no idea. I know I don’t trust Musk because Tesla has been doing some creative accounting to make certain metrics look better than they are.

We’ll only know if/when the company goes public.


Just a clear upfront reminder. This is a thread about SpaceX and the other commercial rocket companies. Not about Musk himself. So don’t, just don’t hijack the thread with your hate or love of Musk. Keep it about Space Programs.

I will be fast to thread ban posters that ignore this instruction and hide such posts.
This will be in lieu of warnings though.

This isn’t quite right. Virgin Galactic is focused on space tourism (about which I share your skepticism, and not only in the scope of the suborbital ‘tourism’ market) but Virgin Orbit, referenced by the o.p. has always been a smallsat-to-LEO operation. Blue Origin was publicly focused on suborbital but they’ve been working a lot of other things on the downlow, including the BE-4 (a replacement for the RD-180 and now the first stage engine for the USA Vulcan) and the New Glenn launch system. Whether those will be successful is another question but they are not the space tourism-focused company many think they are.

As for SpaceX, I think it is pertinent to note that the company had some very significant fumbling before pivoting to what was essentially an Atlas IV replacement that was clearly intended to go after the lucrative government launch market (e.g. EELV/NSSL, NASA CRS and CCP) as a primary revenue stream instead of relying on the vagaries of the commercial satellite market. The teething issues with the Falcon 1 vehicle aside (which are typical of any new launch system), the original plan was to transition to a Falcon 5 vehicle, but at the point they were ready to start real development the financial analysis showed that it just wasn’t viable, so they upscaled to Falcon 9 and made an all out run at EELV. This was a prime time to do so because the United Launch Alliance (a joint venture that the government forced Boeing and Lockheed into so-as to actually prevent competition that was considered to be damaging to both companies) had a high per-launch cost and launch rate determined by contract, as well as having to adhere to Federal Acquisition Regulations (FAR) Part 15 documentation requirements that SpaceX, competing as a FAR Part 12, could ignore. (It is also the case that ULA was just overcharging because of their work structure, but that is another issue unrelated to FAR category).

In the smallsat launch market (which SpaceX has eschewed) the margins are even more slender, and as I’ve consistently said there is only room for one or two operated in that segment. RocketLabs came out of the gate strong on that, and there are a few other companies that may be able to at least get a few launches off, but Virgin Orbit has always struggled to deliver to their promised capability and demonstrate even adequate performance and reliability. It’s just a really tough business to be in even when you have guaranteed government contracts, and even moreso when you have to sell to a commercial market which will hold you to an on-time delivery.

Basically, SpaceX came in with the Falcon 9 to eat ULA’s lunch in the EELV contract as well as developing the operational tempo to support commercial launches at a price structure that Boeing and Lockheed couldn’t meet. Despite the claims that it was all internally developed, SpaceX received both development money and subsidies from both the DoD and NASA as well as various tax incentives to keep them afloat during development (so the next time someone tries to frame it as some real-world example of Randian ‘Objectivism’ in action, remind them that the company has sucked from the public teat just as much as any other government contractor. ) They aren’t actually as cheap as the advertised ‘manifest cost’ would have you believe unless you are flying a box of rocks into an easy orbit, but still significantly undercutting any commercial pricing from the standard players, and at a time that nobody else was ready to field a comparable-class launch vehicle. And they’ve been able to maintain that position because of the capital startup costs of developing a competing system, although to be clear because SpaceX is privately held nobody outside of the company actually knows just how financially sound the company is, although it has maintained a launch cadence with steady income that at least makes it plausible that they are in the black.

All of that being said, SpaceX was on the cusp of fiscal insolvency several times, and even now it isn’t clear that the gamble on Starlink and ‘Super Heavy’ is going to pan out. The basic capability of Falcon 9 is sound and demonstrated that the company can launch the system with high reliability; Falcon Heavy, on the other hand, hasn’t seen much demand in either the government or commercial segments, and aside from the risible notion of colonizing Mars it is unclear what Super Heavy is even supposed to support. ULA is allegedly up for sale (three guesses on who might buy it, and if you are thinking someone who looks like they should be a Bond villain you’re probably right al though that is a large swath of the billionaire would-be space magnate cohort) but whether anyone can convert that to a going commercial concern when McDonnell Douglas, Martin Marietta, Boeing, and Lockheed have failed in the past is a bet I would not stake real money upon.


Evidently SpaceX is betting on a “if you build it they will come” scenario for Starship, which isn’t necessarily as quixotic as it might sound. Falcon 9 has been invaluable as a test bed for the technology; not just the hardware of reusable engines and stages but the operational experience of actually recovering, refurbishing and reflying rockets, and doing so on a schedule. In theory at least the Super Heavy booster is little different than simply scaling up what SpaceX is already doing. The upper stage is more challenging; evidently the test flights of 2021 led to completely scrapping the Raptor 1 engines and starting over with the Raptor 2.

IF the upper stage works, even if only poorly at first (poor reusability and slow/expensive refurbishment), it will offer capabilities that might create its own market. The massive and wide diameter payload for starters. SpaceX claims that Starship will reduce the cost of completing the Starlink constellation, an application where large payloads can offer economy of scale. And if the cost goals can be met or even significantly improve on what’s available now it will make a difference. Even $300 a kilo to LEO would be huge.

SpaceX benefited from several things:

  1. No legacy cruft. They started with a blank sheet. Not just in rocket design, but in corporate structure. The other big rocket companies evolved in the era of cost-plus contracting, which is slow and inefficient. They’ve had a hard time adapting to a more cost-conscious and rapidly evolving environment.

In the case of NASA, they have been crippled by politics. They are forced to use incredibly expensive and old equipment such as the RS-25 from the Space Shuttle and the solid boosters. Fluctuating funding and changing political requirements with each administration have made coherent, efficient engineering very difficult.

  1. Vertical integration. SpaceX relies very little on an external supply chain for major components. Given that the supply chain has been distorted by cost-plus government contrcting, that has saved them a bundle.

  2. Iterative development. SpaceX blows up rockets. Lots of them. They are willing to try things, make the minimum viable rocket that moves their test program ahead, and if it blows up, learn from that and try again. In contrast, the big contractors would work on the design phase of a rocket for years, then build the finished design, This can be a very slow process because you can’t afford mistakes.

SpaceX has benefited from a lot of synergy that comes from their, “build fast, break things, try again” mentality. It can only work if the rocket can be built quickly and cheaply, so SpaceX focused on iteration speed, manufacturing, and low cost.

For example, the shift from composite to stainless steel in Starship allows them to knock out prototypes very quickly. And instead of needing carbon fiber specialists, they hired welders from a water tower company and worked with them to learn how to put together a rocket with steel. And now they knock gigantic rockets out regularly - they’ve already build seven Superheavy booster prototypes and 26 Starship prototypes.

The Raptor assembly process is amazing. Raptor is one of the most complex engines you can build - a full flow, staged combustion engine. SpaceX is now building two Raptors every day. The Raptors cost under $1 million each, and Musk thinks they can get the cost down to $250,000.

In comparison, Rocketdyne/Aerojet just got a contract to make RS-25’s,a similar class of engine. They will be making only four per year, at a cost to NASA of $100 million per engine. SpaceX will make 500-600 Raptors in that same period, for not much more money.

My favorite Musk quote came when he was asked, “What’s one piece of advice you’d give an engineer working on rockets?” His answer was, “Never accept the constraints you were given if you can see a better way. Push back.”

That’s what happened with the switch to stainless steel. One of his engineers was working on part of the rocket which was then composite, and he realized that Stainless would be better in this application for a bunch of reasons. He took his idea to Musk, and Musk agreed and with a phone call scrapped $20 million in carbon fiber tooling and switched production to stainless. A whole lot of good secondary outcomes flowed from that, such as the rapid prototyping ability.

This sounds like elementary advice, but as someone who worked for a huge corporation, I can tell you that it can be almost impossible in practice. If you are a lower level engineer, by the time you get a problem to solve the constraints are baked in, and if you see a better way well, that ship has sailed. Try to push your idea, and you will get stomped on somewhere along the bureaucratic chain.

SpaceX makes it work because there is one guy who can call the shots, and he’s willing to listen to anyone. If SpaceX had been structured as a typical public company with professional managers and a board with fiduciary responsibility, risk aversion and a large bureaucracy will force a very methodological, careful design process with toll gates, requirements docs that can not be questioned by engineering, etc. Blowing up a rocket means pressure from shareholders and executives, and maybe your job. No person has the authority to do something radical like scrap $20 million in tooling because a low-level engineer had a better idea, and a board level decision is often a compromise.

For another example, look at why Lockheed-Martin spun off its ‘skunkworks’. They recognized that their management structure could not deal woth rapid, high-risk development. So they gave an engineer, Kelly Johnson, complete free reign to do whatever he wanted with minimal management oversight and they were shocked by how fast his group was able to develop new aircraft.

So in short, the established aerospace companies grew fat and inefficient on government cost-plus contracting. NASA was crippled by inconsistent funding and a set of requirements they didn’t even want that were driven by politics rather than engineer8ng, and SpaceX was free to innovate and try anything at all due to its unique funding source, vision, flat management, and a Silicon Valley ‘move fast and break things’ mindset.

How not to land a rocket booster:

How many of those crashes would have been allowed at NASA before heads rolled and politicians started pontificating? NASA was ahead of SpaceX in vertically landing a rocket. But they crashed the DC-X ONCE after multiple successful flights, and that was used as an excuse to kill the program.

It is if there is no market for it, and I haven’t found anyone arguing that there is a current or foreseeable market demand for a 150 metric ton launch capability. I’m not saying that something couldn’t emerge, but there is currently zero demand for this beyond the $2.9B that NASA is throwing at SpaceX for the giant combination lunar lander/pizza oven. Even the various commercial space station concepts (of which only Sierra Space has really made much progress) aren’t relying on more than about a 40 metric ton lift capability.

This demonization of “cost-plus contracting” comes from a place of not actually understanding the problems of acquisition strategy and contracting. For the record, the shift to “fixed cost” contracts has not been a resounding success, is basically the definition of “slow and inefficient”, and in many ways a complete disaster as companies like Lockheed and Raytheon have played the “out of scope” game, forcing the government to issue new contracts to do work that wasn’t explicitly included in the original “fixed price” RFP. Cost Plus Incentive Fee (CPIF), Cost Plus Award Fee (CPAF), and Cost Plus Percentage are all viable contract instruments which offer the flexibility to expand to unplanned but necessary work and function well provided they are competently managed by the contracting authority, i.e. they don’t just hand over requirements and expect the contractor to perform but hold to milestone deliveries and objective evidence of performance to well-defined contract goals, and most importantly are willing to cancel or withhold payments until milestones are met. The contracting authorities have become so adverse to any appearance of conflict of interest or subversive corruption that they end up being very permissive to very open and entirely legal exploitation by forcing the government to pay for work that it didn’t know needed to be done.

As far as ‘corporate structure’, it is true that SpaceX is very focused on space launch versus companies like Lockheed and Boeing with much larger scope of defense contracting and the shadiness that comes with that, but it really isn’t much different from the progenitors which started the entire space launch industry. SpaceX is very much a throwback to the early days of Convair, The Martin Company, Grumman Aerospace, et cetera, that were on the forefront of space launch development in the ‘Sixties. If the complaint is regarding the complexity of current space launch contractors, one can really point back to the Apollo program and subsequent Space Transportation System (“Shuttle”), which imposed systems engineering requirements that created a lot of complexity as a result of having so many different contractors providing different parts of the launch system. ULA is actually the result of goverment-mandated consolidation to eliminate supposedly destructive competition, so complaining about how it isn’t competitive or has such a complex bureaucracy is really a complaint at the DoD creating a monopoly.

The use of the RS-25 and Shuttle SRM was supposed to be an interim design allowing NASA to continue to use existing ground support hardware, systems, personnel, and production while developing a true next generation launch system. Because of political interference by the US Senate, the Space Launch System has become an end onto itself with enough changes that it no longer uses much of the ground support equipment and is essentially a very different system from what the JUPITER proposal writers envisioned.

This is largely nonsense. SpaceX has brought the development of certain components, such as IMUs and other avionics components, in house because it made business and technical sense to do so, but they procure tubing, actuators, sheet steel, aluminum forgings, termination ordnance, those shitty Marotta valves everybody in the industry uses, et cetera out from industry because it would never make sense to try to produce those in house. SpaceX is just as dependent upon external supply chains as anyone else in the industry for a vast array of components.

There is an element of truth here, because when a program has a lot of oversight or fixed contract, every ‘failure’ is regarded as a problem rather than a learning opportunity. That being said, when SpaceX has had launch failures or significant anomalies, they’ve had to stop and go through the same failure analysis and fault tree cycle that everyone else does. I would say what SpaceX does well here is set expectations that failures can occur rather than that success is the default, which is frankly just a realistic expectation when you are developing a new system and testing the limits of capability, which other players both large and small are often unwilling to do.

But mostly what SpaceX has done to be successful is focused on the low cost and high impact changes that allow rapid processing and integration throughput. They have carved away things that take a lot of time but shouldn’t, streamlined processing by virtually eliminating all physical paperwork and signoffs by people who don’t need to have approvals, granted authority to people working on the ground to propose and push through changes to improve integration and processing efficiency, and generally done a lot of things that the space launch industry as a whole should have been doing for the last couple of decades except there has never been an impetus for throughput to push for those changes. The supposed advantage of (partial) reusability aren’t really a cost savings per see (as the major cost is in the labor of integration rather than cost of the physical hardware) but by being able to use stages repeatedly they get a higher flight tempo than they could by building all new Stage 1 systems for every mission; even if they have to swap out engines or IMUs, if they can achieve three or four launches in the time it would take to build and test the stages for a single expendable rocket, they can amortize their fixed costs and use labor more efficiently, which is a cost saving that only an accountant would appreciate.

Can anyone else replicate that to become competitive? It seems unlikely, not because there is some special sauce to it—this is basic business practice that the automotive and consumer electronics industries developed decades ago—but just because there really isn’t currently any space for another competitor, and someone trying to get in now would have a huge hurdle of capital expenditure and development before they would even be technically competent. Blue Origin is obviously striking for a middle ground between Falcon 9 and Super Heavy with their New Glenn system, but even if it turns out to be workable and cost effective it will require a bunch of other elements to provide the ‘LEO business park’ business model to be viable without government subsidy or contract.


The bigger effect here, IMO, is the low volume in the supply chain.

My direct experience is limited to cubesats, but I can see the same effect writ large elsewhere. You have one subcontractor that can only serve a few customers. Because the volumes are so low, they can’t get any reasonable level of amortization. The prices include all the overhead you’d expect from having a completely separate business. And the product you get isn’t really what you want anyway, because you still aren’t buying a custom part.

As a specific example, we bought cubesat motherboards (“CD&H”, as they’re called) from a little one-man outfit. They cost a few thousand each for a board less featureful than a $10 Arduino. But they had space-rated connectors, and were designed for the cubesat form factor, and some other things.

There’s no way they could have cost less than a few thousand, because the guy probably sold only a few dozen a year, along with a handful of other products. That barely pays for one engineer’s salary, let alone all the other overhead.

And it was a crappy product and a poor fit for our cubesat. We eventually had so many issues that we designed our own CD&H in-house. It took three weeks from design to final bringup, and worked great. Did exactly what we needed it to, no more and no less.

We have a number of stories from SpaceX about exactly the same kind of uneconomic supplier relationship. They wanted some unit for X dollars in Y months and the quote was far higher and father out than they wanted. It’s not that the supplier was reaming them. It’s that the entire relationship structure was inefficient and made it impossible to charge less. So SpaceX brought these things in-house.

Where this intersects cost-plus accounting is that contractors may have very little interest in rooting out these inefficient relationships. Much worse is when there is political influence, and in those cases there is an explicit incentive to maintain these long supply chains, because this allows them to spread the suppliers over many states. Consider Blue Origin’s National Team for instance, where they are very explicit about how their suppliers are spread out among the states.

That is, of course, what wins senate and congressional votes. SpaceX is no stranger to lobbying, but they have managed not to put political goals ahead of engineering and economic goals–unlike most of the legacy contractors.

One thing Starship can offer is the ability to launch physically larger, not just heavier, payloads. Currently there is high interest in inflatables such as heat shields and habitat modules because existing rockets just don’t have wide enough payload fairings for some of the things people would like to launch. A 9-meter diameter means for example that a Starship could have launched the James Webb Space Telescope with its mirror segments already unfolded and in place.
A big unknown is cost. You’re correct that no one is currently lining up to buy 150 tonne payloads at current prices. But if Starship is cheaper- if it turns out to be a LOT cheaper- that could change.

Or the same folding design scaled up to such a size that its prospective findings might result in Dr. Becky’s level of giddy excitement reaching the point of causing an enthusiasm supernova that would level half of Oxford.

First of all, “Cost-Plus” is a contracting mechanism, not an accounting method. Second, while the statement that contractors “have very little interest in rooting out these inefficient relationships” is true, but that is because it actually isn’t worth pursuing cost reductions on a low volume system; when there is it is often the contracting authority driving those changes through a CPAF/CPIF that provides award fee criteria for cost reductions. The whole idea the that Fixed Fee contracting generally reduces costs is just not borne out by evidence. Each contracting mechanism is really suited to a particular set of programs, and no contracting mechanism is some kind of magical replacement for a well drafted Request for Proposal (RFP), contractor selection process, and good contract management practices after award. If, as a contracting authority, you can’t manage a CPAF/CPIF, where the contract authority has wide latitude to assess contractor performance, you certainly can’t control a Fixed Price or Cost Reimbursement effort where contractors can made endless arguments for out of scope work that grinds any progress to a halt.

The problems with legal ‘corruption’ in the form of political offerings that are distributed across Congressional districts is something that no contract structure or acquisition strategy can really abate. And as much as SpaceX and its principal wants to portray it as the scrappy underdog fighting against the politically entrenched forces of institutional malfeasance, the not-to-be-named proprietor has been heavily invested in both lobbying and ‘dark money’ contributions to candidates across the board, in addition to well-advertised above board involvement with previous presidential administrations. And regardless, the notion that SpaceX is so highly ‘vertically integrated’ as to not be heavily dependent upon external suppliers is ludicrous upon the face.

Just…no. Even if you could fit the deployed James Webb Space Telescope within the fairing of a Starship, it still would not be able to survive the loads and environments during launch in the deployed configuration because it is way too delicate, even if it were ‘beefed up’ with additional structure. And before we go there, the notion that you could pack up a ground-based observatory and launch it into space is also nonsensical. Space-based telescopes are very much purpose designed instruments that are quite unlike their terrestrial counterparts and will remain expensive, difficult-to-design articles for the foreseeable future. This is not to say that there are not certain capabilities that could be developed to utilize the capacity of a super-heavy lift vehicle like Starship (if it actually flies and is reliable) but there is no market in existence or on the foreseeable horizon that would fiscally justify the capability, and the space astronomy ‘market’, just as it is, will never be more than a rider on whatever launch capability might exist to support other applications.


That attitude is how systems get expensive. Progress often happens a tenth of a percent at a time. Negligible improvements are not negligible when multiplied by a hundred or a thousand.

Furthermore, there are knock-on effects to in-sourcing. You keep knowledge in-house. Feedback is faster. You can get a product that is more optimized toward your use case. And–possibly most important of all–you are no longer constrained by the boundary of the contractor relationship. This is a specific case of the more general Conway’s Law:

Any organization that designs a system (defined broadly) will produce a design whose structure is a copy of the organization’s communication structure.

Since the communication between a company and a subcontractor is even narrower and less flexible than that between the divisions inside a company, that relationship forces even more constraints on the final product.

It is absolutely crucial that divisions within a company are able to “trade” with each other. That might be mass, cost, or something else entirely. Rigid boundaries–especially those imposed for political reasons–make those trades impossible.

It doesn’t have to be demand for 150 ton capacity. If Starship has a lower per kg launch price than others even for partial payloads, it can be used to replace smaller rockets as well. It can also combine satellites in one mission or do ride-share with smaller sats.

There is currently huge demand for Starship - Starlink. The big Starlink satellites were designed for Starship. The full planned constellation is 42,000 satellites. They have a lifespan of five years, so that’s a whole lot of launches, plus enough launches to replace 8,400 satellites per year indefinitely. That’s about 70 Starship launches per year, just to maintain Starlink.

Then there are the other constellations from OneWeb, Kuiper, and who knows what else. There currently isn’t enough launch capacity at a low enough cost to really put large constellations in orbit. SpaceX is even launching OneWeb satellites.

SpaceX also plans to retire Falcon 9 and Falcon Heavy at some point and move all those payloads to Starship.

Then there are the Artemis launches, and the inevitable defense launches. And for missions past GTO, you need lots of Starships for refueling. Seven launches per mission.

If all this works, that’s a pretty solid transportation infrastructure. Costs will,come down more, and hopefully enable all kinds of new payloads. If we get really lucky, we might get a virtuous cycle where satellite and space probe construction costs come down to reflect the lower launch cost and therefore lower risk, plus more volume manufacturing.

SpaceX is hoping that this will grow markets for space lift, but Starship can probably be sustainable just as the transport system for Starlink, which is a potentially huge profit center for SpaceX.

One real problem is that they plan to run an assembly line of these things, and they are reusable. So the inventory of rockets will grow to the point where they aren’t going to need more, and they will have a lot of excess production at that point.

Maybe at some point they might just start expending Starships if they can make them cheap enough. Inthink an expended Starship can lift something like 250 tons to LEO. Or you could outfit one with 250 tons of gear and life support and turn it into an orbital research station.

The other possibility is that Musk will really try to start his Mars colony, using funds from Starlink to fund his dream and use a lot of rockets.

Reversed thread ban. Apologies.

Reading through the various posts, and I sincerely hope that SpaceX continues to succeed and Starship ends up being everything it can be. There are a lot of benefits to having humanity being able to operate further and more cheaply into space. While in the extreme long-term humanity must become an interplanetary and interstellar species if we want to survive, in the shorter term, it needs to be about resources. And I think that should be the focus. We don’t need a moon base (except for cheaper launches) or a Mars base. What we need is the ability to capture an asteroid, carve it up and bring the resources down to Earth. And then of course space-based solar panels could be a great boon. Or space-based nuclear as it gets rid of the NIMBY problem.

@What_Exit, I don’t think @Sam_Stone’s post was inappropriate. He was responding to the claim that Starship simply wouldn’t have enough customer base, and he pointed out that if there really will be a Mars colonization effort that that was/is the raison d’etre of Starship in the first place. That doesn’t seem to me to be in violation of the warning; are we really forbidden to even mention E.M.'s name?

Agreed. Sorry for the problem.