Perspective on a Failure
Another Starship lost - is the program in trouble, or can SpaceX bounce back?
SpaceX has lost another Starship, Ship 36, to an accident during a static fire attempt. The vehicle burst, and the was engulfed in a fireball, during the propellant load operation.
There have been a string of failures in the Starship program this year - mostly relating to the new V2 iteration of the ship. The upper stage of Flight 7 was lost due to a resonance issue, Flight 8 was lost at roughly the same point in flight - but, contrary to what observers initially thought, due to a different issue related to the engines. Flight 9 completed its burn and reached its suborbital trajectory, but due to a leak lost attitude control and was destroyed on reentry. The booster on this flight was also lost when it exploded shortly after igniting the engines for its landing burn.
According to a statement by the company, the most recent explosion was due to a failure in a COPV (Composite Overwrapped Pressure Vessel) containing nitrogen gas. If so, this is a problem with the quality control in a particular component, and not - as some are keen to suggest today - an inherent design flaw in the vehicle. A full investigation will likely reveal more details.
There are many critics of Starship - those who dislike the company, its founder, or the very concept of space colonisation. There are also those who tout their engineering credentials and argue that the vehicle is doomed to failure. All of these people will no doubt be celebrating this event, and claiming vindication of their beliefs. Such crowing is entirely premature, and those indulging in it will just make themselves look foolish later on.
Make no mistake though; this is a serious setback. A failure of this kind should not be happening at this stage in the program, and it’s no good glossing over it with references to “fail fast”. SpaceX needs to get Starship back on track - but based on past performance, we can expect that they will.
History
SpaceX has had failures before, and recovered quickly. There was a Falcon 9 lost in flight in 2015, causing the program to be stood down for several months - but the return to flight was the first successful landing of a booster. The loss of Amos-6 the following year (due to a COPV issue, like the current anomaly) likewise did not stop the Falcon 9 program.
And of course, the company began with a string of failures - the first three Falcon 1 rockets they launched all failed. There were so many failed booster landings before it was perfected that they made a blooper reel of them.
The peculiar culture of this company seems to feed off failure, and through it achieve long term success. We’ve seen more recent examples of recovering from setbacks we well. For instance, after the first flight of a full stack Starship, there was considerable damage to the launch mount - which critics suggested would take much longer to fix than it in fact did.
Anyone who has been paying attention to the company for a while knows about these things, and this is the source of optimism in the face of setbacks. Hopefully, if the damage at Massey’s is not too bad, the time it takes to repair the damage there will surprise us again.
The Next Mars Window
The next launch window to Mars opens at the beginning of November 2026, and lasts through to January. SpaceX will be doing everything possible to launch something towards Mars in this window, as each window represents a once every 26 month opportunity to test the interplanetary capabilities of Starship, and in the rapid development favoured by SpaceX, 26 months is an eternity.
There are many things to be learned from the first Mars attempt - how to navigate Starship to Mars and make course corrections if needed. How to communicate with it, and how to maintain power and thermal management during the trip. None of these are very novel things in space travel, but they are capabilities that need to be developed and debugged sooner rather than later.
So what do SpaceX need to do between now and the next window? Starship needs to get past the current failures, test in space relight, and at that point will be considered ready to go to orbit (as there will then be confidence in a controlled deorbit). Then, a demonstration of on-orbit propellant transfer will have to be done. Once this is successful, the flight rate will have to increase to such a point where sufficient propellant can be delivered to the Mars-bound Starship before it boils away. Then they are good to go. If they can relight the engines on orbit, and have enough propellant in the tanks, then as a simple matter of the rocket equation that can put the craft on a Mars transfer orbit. What happens on the way or when it gets there is another matter - but they can make the attempt.
My best estimate is that there would have to be 2-3 tanker flights, and thus 3-4 flights in total, to launch a mission towards Mars with minimal payload. In his recent presentation at Starbase, Elon Musk said the target was 5 ships with 10 tonnes of payload each - this to me was aspirational even before the current setback, but the fact that the choice was to fly more Starships with less payload, rather than concentrating the propellant deliveries to allow fewer flights with larger payload, indicates the priority for this first attempt.
Data is the main payload, and more data is obtained by having multiple shots on the target. If multiple Starships are launched in the 2026 window, they will likely arrive separately, one after another, meaning that engineers on the ground will have a chance to send software updates to the next ship based on the performance of the previous ship. The last ship will have departed Earth before the first arrives at Mars, though, so hardware updates cannot be iterated within a single window. This just underscores the importance of sending something as early as possible - they need to know if any hardware modifications need to be made as soon as possible.
Predicting the Future
2026 is still in play, despite those confidently asserting that this single event means it is now lost - which, incidentally, they have claimed after every incident. Often such people will make claims about Musk promising Starship flights in the 2018 window (he didn’t) and how the 2020, 2022 and 2024 windows were missed. All true of course, but there are explanations far better than “Its all a fraud and I’m the only one clever enough to see through it”.
Starship is a complex program, that has suffered substantial setbacks, but is now approaching maturity. A lot of the risk and delay of the program is already now behind, and there are not actually that many milestones before unmanned flights to Mars can begin. The simplistic analysis of extrapolating previous delays is not correct, no matter how good it feels for pessimists or how much clout it earns them online.
While the explosion at the test site does block further static fires of ships, it does not block other parts of path to 2026. Construction of a more robust second launch pad at Boca Chica is proceeding, as well as work on three more pads in Florida. The Super Heavy booster program is performing better than the Ship, with three successful catches out of three attempts on the tower, and one reflight of a caught booster. Many pessimists were quite sure that catching was ridiculous not that long ago, and the ability to refly the booster is a critical step forward for increasing cadence, as it substantially decouples flight rate from the rate at which engines can be manufactured.
Sometimes the smartest thing is to say “I don’t know”. Unlike people who are now absolutely sure, on the basis of their armchair engineering, that the 2026 window cannot be met, I myself don’t know. I suspect a lot of people at SpaceX don’t know either. But so long as its not impossible, I am sure they will give it their best try.
This is the end of the article. This one is free for all readers, but some of my work is for paid subscribers only. Upgrading gets you full access to articles like these:
The history and current state of the art of partial gravity research, important for building human colonies on Mars or the Moon.
How the threat of ASAT weapons and satellite constellation technology could lead to a militarisation of space.
…and many more, including my regular monthly Mass Value Report where I apply a unique analytical lens I developed in this paper to the launch and space industries.
To subscribe costs about the same as a cup of coffee every month. Would you buy me a coffee to hear a complete set of my insights each month?
I actually like it when SpaceX has accidents because these enable them to discover problems fast and without human deaths. Contrast this with Boeing, where so many have died in operational incidents.
Comparison class: the annual International Rocket Engineering Competition https://www.soundingrocket.org/what-is-irec.html. Lots of them exploded this year -- as usual. Students learned, as usual, aided by a team of experts who analyzed their hardware prior to test and participated in the post-kaboom analyses with them. It's a huge volunteer effort.
Rocketry is hard. I wish them luck.