Just to be clear, I’ve never stated that SpaceX is doing these launches for free, or not making a profit on their national security (EELV/NSSL/NASA) launches, for which the actual contract value is substantially over their advertised manifest cost, which even their commercial customers don’t pay for all launch services. The cursory analysis (which I eclipsed for brevity) doesn’t really stand up to scrutiny without some supporting financial data from SpaceX which they have steadily refused to disclose but even if we take it as gospel it does not account for the debt load that SpaceX, like any large company, carries on its balance sheets.
I don’t think there is a real basis for estimating the actual costs for a Falcon 9 loss, but I have worked on studies for space launch vehicles in the same category (and no, not using NASA or DoD costing tools or methods), and even given the economies of scale of high launch rates, efficiencies in streamlining and automating processes, and making design choices to minimize manufacturing and processing costs, I still do not believe that they are launching a Falcon 9 Block 5 for build and operating costs of US$15M per launch. Even with Stage 1 reuse (which is not as ‘turnaround’ as enthusiasts would like to believe) that is just not a plausible figure given just the labor costs of the number of people involved in inspection, integration and testing, handling and transportation, mission planning and NRE (for non-repeat missions), and launch operations. I’m not going to make any specific estimate because there just isn’t a basis for anything but a wild ass guess but that figure is well below the floor of what their costs are.
The Space Transportation System (STS or colloquially ‘Shuttle’) shouldn’t really be part of that comparison because it had specific features and design drivers that are outside of what the Falcon 9 does or can do. Although it was originally expected to be a low cost “delivery truck” to orbit, it was pretty clear early on in the design that it was not either going to be be capable of the high launch availability or low cost estimates even before the failure of Challenger on STS-51-L in early 1986 caused a more than two and a half year pause in Shuttle operations and a series of operational changes that created more work and expense in addition to the refurbishment required in Shuttle Main Engines, rework and repair on the thermal protection system tiles, and a variety of other unanticipated costs and schedule drivers. I’m not sure where you got the $60k/kg cost listed for Shuttle but that is in the ballpark for costs in the late 'Nineties and early 'Aughts where flight rates were only 2-5 launches per year and the United Space Alliance, which performed all of the actual ‘wrench-turning’ and integration, was having difficultly recruiting and maintaining skilled personnel. Needless to say, the Shuttle never lived up to the promise of cheap and reliable access to space, and the program was mostly retained to complete the International Space Station (which itself was a massive boondoggle).
Atlas V and Delta IV (and though you don’t mention it, Delta II) are comparable vehicles, and in fact I worked on a study showing that the United Launch Alliance (a ‘joint venture’ forced upon Boeing and Lockheed Martin specifically to prevent competition as there was the belief that the companies would not be able to maintain profitability competing for launches) which concluded that they could perform the same number of launches with the same reliability for ~60% of EELV contract costs, or else perform twice the number of launches at the same overall contract costs, all without any significant changes in design and just by implementing modern data handling systems and greater automation in integration processes. There were certainly further costs savings in implementing design changes and supporting higher flight rates (albeit not ‘order of magnitude’ reductions) but there was really no incentives under EELV to do so, and in fact there were specific pressures to avoid implementing changes with costly requalification efforts and uncertain impact upon reliability for national security launches. At that time there was not a large market for commercial launches for vehicles with that capability and what there was was being serviced by foreign competition (specifically Russia and Ukraine). Boeing and Lockheed did make efforts to get involved in commercial space with the Delta III and the Sea Launch joint venture (for the former), and Athena (for Lockheed), both with negative results which doubtless diminished the interest from corporate management on investing internal research and development (IRAD) funding into uncertain space launch developments.
I want to be clear that I’m not trying to bash SpaceX, which through having a singular focus on space launch and a willingness to implement cost reduction measures and facilitate methods to increase launch tempos has actually served to engender a commercial space launch market with what are inarguably lower costs and higher launch availability, and after early stumbles with enthusiastic but inexperienced staff hired people experienced in engineering, manufacturing, and launch operations to find the right balance between innovation and reliability. But I don’t believe for a second that they have achieved any kind of order of magnitude, sub-US$1000/kg launch cost with the Falcon 9 even with reuse and their admittedly impressive launch tempo.
I don’t have enough insight to even guess at what the launch costs for ‘Starship’ are, and I’ve never said that “reusability is still of questionable importance…”, just that it is not the orders of magnitude cost savings that people believe it to be and that the comparison to airliners is not really Germaine, but I’ll heat my hat if SpaceX is actually “launching [Starship] a couple hundred times a year” in the next three or four years, or that there is a commercial market for such launches aside from SpaceX itself generating a massive amount of orbital e-waste.
First of all, this mythical and constantly inflating-in-value “$10,000-toilet-seat” does not and has never existed. That mangled factoid actually comes from a toilet enclosure that was custom fit to the lavatory in the P3 Orion aircraft, and the issue was over Lockheed margin charing $604.09 for a component that the US Navy argued should have cost $554.78, the result of which is that Lockheed refunded the Navy $4,606.74. More money was wasted in adjudicating and then responding to outrage than was actually recovered in the overcharge.
It is true that ‘mil-spec’ hardware costs more than a dimensionally and functionally equivalent purely commercial component but that isn’t due to corruption or malfeasance but because those components, intended to be used in safety- and mission-critical applications, have traceability back to the original manufacturer and material certifications showing that they meet the specified minimum strength, corrosion, coatings, qualification, and other qualification. In many cases with things like fasteners and adhesives, these are not special parts built for the space launch industry but actually commercial grade hardware also used in commercial aviation, which also pay the large costs for this information. Having worked on investigations involving suspected counterfeit parts that failed unexpectedly, I can say that having this information is crucial in coming to root cause and avoiding defects that compromise the design.
There is certainly corruption and waste in aerospace and defense contracting, but it is not down at the hardware level, and I have never seen any engineer or procurement person deciding to add cost to a design or select the most expensive choice just to make the company more money (and in fact proposed cost reductions are often nixed by government contracting for not ostensibly meeting requirements even though they are technically acceptable or even superior). The real egregious wastes are at the program and even legislative level, where contractor proposals go through technical evaluations without any independent engineering review only later to discover that the contractor never proposed providing obviously necessary parts of a system (leading to ‘plus-ups’ or additional contracts at a premium because of schedule) or programs imposed by Congressional. fiat even though it is obvious to everyone involved that the program objectives are infeasible. Forget $50 hammers or exotic alloys in a place where 5052AL will do; it is the hundreds of millions or billions sunk into programs that never see the light of day, or experience cost growth because of shifting or conflicting high level requirements and waves of politically-induced slowdowns and speed ups.
I’m not going to defend SLS which is widely regarded in the aerospace industry as a massive boondoggle that is kind of a tool looking for a nail, and has a projected launch rate so low (at most once a year, and currently on manifest only about once every other year) that it will never be financially viable. That being said, the ‘go fast and break things’ approach is not a great way to work on that level, because while you learn more from failure than you do success, you also get more information about failure mechanisms from well-instrumented ground tests where you can control conditions and reproduce a failure than you do in flight where loss of telemetry and a lack of control can lead to uncertain root cause and effective corrective action.
Certainly, the first failure of the ‘Super Heavy’ wasn’t due to a need for design improvements on the vehicle but a predictable occurrence of unreinforced concrete being blasted back up into the engine bay. That wasn’t a ‘successful failure’ in learning something valuable; it was a complete waste of a test launch causing easily avoidable damage because of an impetus to move too fast and not listen to the experience of those saying it was a bad idea. Elon Musk is really good at hype-spinning these kinds of stupid failures into celebrations but it was really just fucking stupid, not to mention the damage done to Brazos Island Stage Park and the ecologically sensitive Boca Chica Bay. But fuck 'em, it’s just a bunch of birds and a species of stupid ocelot, amirite?
This is just…not true. The dynamics of the space launch environment (vibration, shock, acoustics) are higher than essentially anything on Earth outside of a blasting pit, and the combination of stressing space environments such as thermal vacuum, thermal cycling, and ionizing radiation are just not conditions seen by any other engineered system outside of the core of a nuclear reactor. The “constant significant acceleration” is in generally not a stressing condition to design products to survive, and even if that were a major constraint payloads and components experiencing a space launch will see loads due to thrust of 3-15 gees of equivalent acceleration.
Most space contractors–including, in its early years, SpaceX–have tried to adopt a broad use of commercial of the shelf (COTS) hardware for the purported costs savings and just delta-qual it for the launch and space environment. This almost always results in problems that require redesigns or ad hoc ‘fixes’ that often end up costing more than just buying a purpose-designed component which has been qualified to meet requirements in the first place, which is why the common retort to COTS in the aerospace community is “Crap Off The Shelf”. It is certainly a case of “buy once, cry once” versus spending thousands of hours of engineering labor to try to make a square peg fit in a round hole and still retain its shape.
I really wouldn’t compare Shuttle to SpaceX (at least, not the Falcon 9); Shuttle just got more and more expensive and less available as the program went along, and because of the fixed requirements of the design there was very little design space to make cost-saving changes and improvements that weren’t forced by obsolescence. SpaceX has certainly turned the Falcon 9 into a viable and reliable launcher, and I don’t question that they’ve achieved some savings in terms of efficiencies and increased revenue through reuse; I just question the unfounded assumption that they are making dramatic, order-of-magnitude reductions in cost and then holding all of the profits while still somehow undercutting everybody else in the industry. Shuttle was always a political program with competing objectives and unrealistic expectations given how complex the system was. The Falcon 9 is a mostly conventional rocket build by a company doing some things that are innovative (albeit, not as novel as many people seem to believe), and with the willingness to take some risks and be focused on continuous improvement instead of locking down a ‘working’ design.
Stranger
