- Launch roughly on time, after a scrub yesterday. (Sounds like the scrub was due to ground equipment, most notably the water system.)
- Initial ascent was good, but then one engine on the booster went out.
- Relight of the booster's engines after stage separation for the boost back burn failed. Engines did light again for a landing burn, but seems to have hit the water harder than expected and was very off target.
- Starship lost one engine shortly after stage sep. Turned into an unintentional test of engine out capability. It made it to space.
- Some weird motion and lots of off-gassing after engine cut-off, with uncertainty about if it actually got a good orbital(ish) insertion. Seems to have been benign, with the motion being a weird slow flip to the orientation for payload deployment.
- Test deployment of dummy payloads was successful, including a couple with cameras to look back at Starship.
- An in space engine relight test was skipped, presumably due to the issues during launch.
- Re-entry to over the Indian Ocean seemed to go really well. Nothing obviously burning or falling off. The amazing views of the plasma during re-entry, something never seen live before starship, are now routine.
- Starship did a maneuver to simulate how they'll have to go out over the gulf and back to the landing site.
- Nailed the target, evidenced by views from drones and buoys. Soft landing before falling over and giving us a big (expected) boom.
As far as overall progress from previous test flights goes, they're at least treading water while making many large changes. I think they were hoping to try for a tower catch and actually going orbital for next flight, but I highly doubt that now. The boostback burn failing was the largest failure, with the engine failure on Starship being a close second. Good performance despite engine out seems to be an unintentional success.
Good summary. The booster appeared to hit the water at 1400 km/h (a bit under 900 mph) so not really survivable :-). Engine out on ship seems to left them with just enough fuel to land but not enough to do the hover thing (simulates being caught by chopsticks). They notched it down to two engines (vs planned 3) on the landing it seems?
Basically if they can figure out the engine issues, it looks like they should be able to do a full end to end flight. That's reasonable progress. Given the IPO this was a pretty important flight and I don't think they hurt it (like blowing up on the launch pad would have). So their one step closer it seems.
They've made great progress but have a bit left. It's always the last mile, isn't it?
It's so cool seeing progress in this space (sorry).
There's no materials science breakthrough needed -- the shuttle used ceramic tiles successfully its entire service life. What's needed is engineering work, and that's what SpaceX has been doing.
We’ve seen much larger holes than that in previous tests. Some of the control fins burned completely through.
Here is a decent summary.
The re-entry itself looks amazingly smooth compared to V2. TBD whether it's good enough for re-usability (much less rapid re-usability).
But Flight 12 was definitely forward progress.
The problem was never solved and turned what was supposed to be a few days into weeks or months. Every mission the shuttle had to go back into the assembly building and have all tiles inspected and potentially replaced.
Obviously doesn’t guarantee they’ll find solution, but fast iteration will definitely help.
Its turn around time is ridiculous, it has to be maintained with specialized equipment/hangers, along with external contractor assistance.
Compared to the Gripen, as an example, which can land on a freeway and be up in the air again in a few minutes.
One was designed to be used in war, in desperate scenarios, with no ability to coddle it. The other, the F-35? Is designed around milking the taxpayer as much as possible, and employing people in as many politician's states as possible.
The shuttle was like that, I think. Which is really sad.
The gripen is a much less capable non expeditonary platform designed to maximize asymmetric losses if sweden is invaded. As a small country sweden has to follow a porcupine strategy to deter invasion.
Presently the actual comparable to the F-35 is attritable drones, which is why every mid-size and major power is developing them.
It’s an extremely different design goal, the US doesn’t mind exotic weapons that require exquisite (and expensive) methods of servicing, they have the budget and the assumption that a well equipped air field will be immaculately maintained.
Meanwhile the Mig-29 designers assumed it’d operate from damaged/poorly maintained fields, so on the ground you can shut the primary air intakes and it uses ones on top of the plane to get air, drastically reducing the FOD risk on taxi/takeoff.
I do wonder how well the F-35 would fare in an actual shooting war against near peers when all the peacetime assumptions breakdown.
The F-35 was just in a war, in Iran. It performed as expected and was able to roll back Iran's air defense network in days.
The maintenance is the real difference - US specifically USAF gear is designed for nice air conditioned hangars to do regular maintenance, Gripen, MiG-29, and to way lower effect F-18 (when compared with F-16) - the first two assume forward bases without ability to do major maintenance, and even the latter (and other carrier adapted ones) promote things like quick swap engines because that's no space for hangar queen to have deep engine maintenance just so engine vendor can claim long time between overhauls
The engines did have lower overall hours, yes, but the suggestion they need whole overhaul after very few hours is because it looks so when looking at it from USAF doctrine where "removing engine and sending it to special facility" is only for rare complete overhauls, and local mechanics are supposed to do regular minor work all the time.
MiG-29 instead was done under doctrine that the airbase does not have mechanics capable of doing such overhaul, nor the facilities to do so, and instead you swap the engine and send the used one to maintenance facilities further away from the front, same with other aggregates.
Not just space-potatoes… but missed the separation shot on the live feed. How in the hell!?
Also I think Ship now has methane thrusters on it. They were operating with a clean blue flame in short purposeful bursts.
A good cold gas thruster produces a lower density, more expanded flow, which looks blue for the same the reason the sky looks blue.
One can compare this to the exhaust from various Falcon-9 engines and thrusters when it is illuminated by the sun on the backdrop of the night sky: https://youtu.be/JRzZl_nq6fk?t=193
This was as good, if not better, than the livestream itself!
SpaceX’s people were saying it was on target, and it seems to have landed in about the same position relative to the camera buoy as previous flights. I don’t think there’s any evidence to call it off target. The landing and toppling looked the same as previous flights too.
EDIT: I cannot reply further in this thread, but my understanding is that the non vacuum engines are not intended to stay lit throughout the orbital flight in a typical mission. If they are, they can gimbal and compensate.
I said some raptor engines are on a gimbal, not vacuum engines.
To be precise, the three central engines can gimbal up to 15 degrees. That can control the thrust vectoring when an engine fails, and that’s what happens in the last flight.
Since the flight already happened and we know it didn’t spin out of control (unless you imply their diagnostic and telemetry was completely off and the engine was actually on) something must have compensated for the failure. It wasn’t magic, it was in fact the central 3 engines that did that.
You may be confused because those are called sea level engines, but that doesn’t mean they can’t work in vacuum.
The upper stage has six engines. The outer three engines are “vacuum engines” (optimized for operation in space). The inner three engines are “non vacuum engines” (optimized for operation in the atmosphere, at sea level).
The outer three vacuum engines are not gimbaled, but the inner three sea level engines are. Thus, it is completely accurate to say that they gimbaled some of the engines to compensate for the engine failure.
The word "live" is doing a lot of work here. Astronauts used to film the plasma going past the windows of Shuttle.
I remember as a kid my science textbook had a still of it to illustrate plasma.
A latency of a few seconds for streaming video compared to several months for a still photo from the Shuttle seems an entirely valid use of 'live'.
The booster not completing the return part of the flight was disappointing. They had a similar incident in one of the previous flights, when they tried to maneuver the booster too aggressively immediately after stage separation which caused problems with the fuel supply. If it was something similar this time, it might be solvable by changing just a few details of the maneuver. So, maybe it is not that huge of a deal.
There were many cool things in the webcast, from them showing the catamarans that are deployed at the landing site, to the views form the cameras on-board of the "satellites". The first few minutes after liftoff were just amazing visually.
The actual mission was not in doubt
They are willing to have "negative outcome learning experiences" to gather data quickly. and, of course, data, data, data.
I like it because I know what insane amount of red tape has built up to do anything similar in a Gov (any Gov).
Absolutely true, but ignorant stock traders making irrational trades only matters if company management pays attention to them. Musk will maintain complete control of SpaceX even after the IPO, so he can focus on long-term value rather than short-term ups and downs.
Of course, over time, if more shares are issued, this may change.
https://scholar.google.com/scholar?q=NASA+Data+Acquisition+S...
https://scholar.google.com/scholar?q=Design+of+Electrical+Sy...
A typical test stand would have maybe a thousand channels of relatively slow data (pressures, temperatures, flow rates, valve states, etc), and maybe up to a few hundred of channels for essentially audio data from vibration sensors. This amounts to sub-gigabit per second data rate overall.
If very high speed video / multiple video cameras are used, this could generate massive data rates, but unless something interesting happens it is not clear how important this data is.
In flight, the telemetry data rate from the entire Falcon-9 used to be measured in megabits per second per stage, plus the video stream. It was not a huge amount of data. Presumably now with Starlink they send a lot more telemetry from Starship, but in flight the engines typically have far, far fewer sensors compared to the ground testing.
Is this enough progress to keep a 2028 crewed landing? Don't know.
I'm curious whether they are going to try to recover a Starship before trying for in-space refueling (or the reverse). Either way, I think both have to work before they can try for an uncrewed lunar landing (presumably in 2027).
The big question is re-usability. How close are they to relaunching a Starship? They may not know for sure until they can get one back intact. If they can launch at least once a month, maybe they'll make it.
If they can re-fly a Starship this year AND demonstrate in-space refueling, then 2027 can be all about an uncrewed landing attempt. That would make me feel good about a 2028 crewed landing on the moon.
I'd bet that they'll not try in-space refueling before they demonstrated in-space relight of an engine. So they need to fly at least twice. Or even thrice because to demonstrate refueling you need two Starships in orbit.
At a minute in you can see the satellites being ejected out one by one.
> Sprint accelerated at 100 g, reaching a speed of Mach 10 (12,000 km/h; 7,600 mph) in 5 seconds. Such a high velocity at relatively low altitudes created skin temperatures up to 6,200 °F (3,400 °C), requiring an ablative shield to dissipate the heat. The high temperature caused a plasma to form around the missile, requiring extremely powerful radio signals to reach it for guidance. The missile glowed bright white as it flew.
https://en.wikipedia.org/wiki/Sprint_(missile)
Maybe v4.
Such accelleration! Much wow!
It seems to give the booster a real kick - what's that do to turbo's and fuel movement?
You've got hot exhaust onto cold cryo fuel tank header?
You've got to carry more mass in terms of protection for the tank?
Is doing MECO and then push and then get 100 yards apart or something before second stage / ship engines kick on a big enough penalty to justify all the extra complexity?
It's a major overhaul of the design they've been working on for a long time. There was talk of v3 fixing the problems in early v2 test flights. The booster is v3 as well which presumably is why they had some problems. I believe this is also the first time they flew the v3 engines with the plumbing fully integrated in a single piece housing they 3D printed.
Is it disappointing that they had a couple of engine outs, and also trouble with the booster relight? Sure. Do I have even a little doubt by now that they can fix these problems? None whatsoever.
The success of Ship 39 today was a big, big deal.
It might, but it certainly helps having a ton of them around. Given that they used 42 of them today and 2 failed in some fashion, we'll call that a 1:21 failure rate. On a more typical rocket with say 10 engines (eg falcon 9), there's a good chance they wouldn't have seen the same failure till flight 3.
20+10+3=33 on the booster, 3+3=6 on the Ship, total 39.
I remember Elon said they want to add 2 engines to the first stage, but that still would be 41. Where's the 42th supposed to be?
The whole point of Starship is that it's a reusable vehicle with easy turnaround and quick maintenance. And in particular it's supposed to be different than the other reusable vehicle with easy turnaround and quick maintenance, which turned out to be sort of a boondoggle.
Yet, they've now hand-built and destroyed twelve of these things across multiple redesigns, and it still hasn't completed its design mission once. In fact basically every launch has unexpected major failures.
As poor as its safety record ultimately ended up being, the shuttle launched successfully on its very first try. And we only had to hand-build five of them. And lost two, sure, which is still a lot less than twelve.
Yes yes, I understand that iterative design has merits and that the ability to rapidly prototype and try things in the stratosphere allows for less conservative tolerances and better ultimate performance.
But does it really take 13+ tries?! At what point to we start wondering if we have another boondoggle on our hands?
I think blowing up a handful of rockets is a fine idea. But at some point you have to ask yourself if it will ever work? Why are we on a another engine redesign? Why is this the third iteration of the second stage? How many more?
And what number is that point? Six? Nine? I'm thinking thirteen may be getting into the danger zone.
Here we have a cutting edge rocket design - scale, sophistication of engines, design goals - and a commercial evaluation, which path would get to the intended success cheaper. NASA doesn't like public embarrassments, and, as Henry Spencer reminds us, when failure is not an option, the success could be quite costly. So NASA spends billions and many years for a fragile system. If the goal is an airline-like operations, the design should be thoroughly shaken up. It's known that no simulation, no static testing can equate the actual flights in the ability to get the data best describing what conditions the system will encounter in real use. And also, given the industrial scale of Starship production, each flight hardware costs way less than if we'd built them manually, in quantities justifying naming each unit separately.
In Soviet Union, where rocket departments were part of artillery, the testing with actual launches seemed logical. In this case the approach to run a massive test flight program seems logical too, and we can't complain about the lack of progress - first Starship had way less capabilities and performed way worse. In USA we had more than 1000 tests for injector head for F-1 engine in Apollo program, and this number was justified at that time. Starship is way bigger - but the progress is also undeniable, and it would be odd to stop test flights now, when the 3rd iteration of design looks promising.
So, while we can't pin a particular number of tests, I don't think we should worry yet. This year and the next one should be important for Starship program, given SpaceX commitments to help NASA Artemis. If we won't have orbital Starship then - we can come back to this question.
SpaceX have already proven that the iterative approach works with Falcon 9, literally the most successful rocket program ever. SpaceX have also proven that this specific Super Heavy/Starship rocket design isn’t a dead end. Criticising them for failing to succeed in the future is a valid but uninteresting opinion.
Just looking at it should tell you a lot about why:
https://www.metal-am.com/wp-content/uploads/sites/4/2024/08/...
It’s cheaper and faster to make in volume. It doesn’t require nearly as much shielding, because it’s less fragile, which saves a lot of weight. The engine itself is lighter. And on top of that, it develops more thrust, at higher fuel efficiency.
The net result is cheaper and lifts significantly more mass to space, which significantly drops the cost per kg to orbit.
It already worked, they’re making it much better, and getting it ready for a level of mass production that we’ve never seen anything close to in the space industry, even from SpaceX. They are much more ambitious than I think people who haven’t been watching them closely understand. The US grid is 1.4 TW of generation, they’re aiming to put up 1 TW of AI compute every year. Maybe they’ll stop well short of that, but their stated goal is insanely ambitious.
For reference, SLS has been in development for 5 times as long, and cost 15-20 times as much, as Starship, and they still haven’t landed people on the Moon, which has been one of the stated goals since the Constellation program in 2005.
I don’t see how the number of failures matters if the end result still happens faster and cheaper than anything else.
It worked out in the end, but I can’t imagine being so confident in a new system, no matter how much money and brainpower has been spent to make it safe.
That's undeniably true. Nonetheless "Better than the shuttle, which sucked" isn't the design goal.
The question is not even just "is it better to blow up 12 Starships?", which would probably still be true. It's "Why isn't Starship working yet?" and the implied "Maybe Starship sucks too?!".
'5 times as long' is dubious too. SpaceX claims to have been working on the design since 2012 vs 2011 for SLS. Ultimately though the start date of a complex program is not well defined, as early conceptual design stages can take years without leaving the drawing board. Government needs to put a start date on such efforts for legal/budget reasons, but a private company does not.
Also relevant - SpaceX has been given a lot of tech and expertise from NASA at a tiny fraction of the cost and time it would have required them to develop it themselves. Therefore, the costs of NASA programs like space shuttle actually includes some of the development costs of SpaceX.
Both programs pale in comparison to Saturn V, which was faster, cheaper, and more technically demanding at the time.
Most obvious improvement was having no re-entry heating problems, secondmost was deploying with zero issues and with a faster pace. It appears they decided to pause the "horizontal" movement of the pez dispenser before a final push away, probably to avoid vibration causing those "bonks" on the payload door, like we had once before.
Flight 12 was a relative success. Some engines failed to light but that's an unintended good test. Rockets are typically designed such that they can have a certain number of engines fail and still achieve their mission.
At this point, the entire SpaceX project is a bet on telecommunications services, specifically direct-to-satellite handheld Internet. That's the only market that will recoup the program costs.
We don't have exact figures for the current true cost of a Falcon 9 launch factoring in reuse but many think it's somewhere betweenm $10 and $20 million. Well, SpaceX has spent 100 F9 launches on Starship so far and that's how you have to look at it. Say F9 is $20M and Starship once it starts launching Starlink is $10M that's 150-300+ launches just to break even.
You might be tempted to say there are other missions for Starship but there really aren't. Satellites aren't that bug, as evidences by there being ~1 Falcon Heavy launch per year (usually for the military and/or to geostationary orbit AFAICT). You can't economically put multiple payloads in one Starship because they all have different orbital parameters.
F9 is rated for human spaceflight. It's a long road for Starship to be certified for human spaceflight. SpaceX hasn't even begun to test in-orbit refuelling yet. Gases are weird in microgravity.
F9 is the cash cow funding all this and that too might go away if Blue Origin or one of the other wannabes ever gets a reusable launch platform to commercial operation.
There are big launches like interplanetary missions but those are few and far between.
It would be fascinating if what ends up dooming SpaceX is actually Twitter.
There's also a military angle here. I'll leave it as an exercise for the reader to look into Musk's history with Michael D. Griffin from the Reagan SDI/'Star Wars' program.
Obviously a few hundred kg of payload in orbit are not equivalent to the same payload delivered directly to a target.
Then the deputy director of the program met a young man named Elon Musk, and the rest is history.
I’m imagining a launcher in a spacecraft that kicks out a bunch of payloads, one at a time, out the back, into orbits with perigee on or before the ground. (An LLM calculates the needed delta-V at under 200m/s, which is likely quite manageable with a small mass driver-style launcher or a very small rocket.) The payloads will lose a bunch of energy to the atmosphere, but all the remaining energy is kinetic energy delivered directly on target, assuming that you can inexpensively aim the thing at a target. Look up “Rods From God” on Wikipedia — you don’t even necessarily need any explosives.
So the question becomes: how economically can one build the guidance systems, avionics packages, and whatever heat shielding is needed to survive reentry?
(Cold War-era ICBMs with MIRV payloads are sort of in this category, but they treated launch vehicle as disposable, which means that the launch would be far more expensive but the reentry system could likely be a bit simpler as the payloads could be launched from a launch vehicle on a non-recoverable orbit. And it appears that Russia has attacked Ukraine with a MIRV-equipped missile with non-nuclear payloads, so there is precedent.)
I seriously doubt that. Just for example, mining a single asteroid has the potential to flood the market for any number of metals. I don't pretend to know how expensive it would be to achieve that in practice; my point is that there are quite a few different ways to recoup program costs at some handwavey point in the future.
An asteroid is much, much further than that but more important than distance is the delta-V required for change its orbit to reach an Earth orbit. So you not only need to get there, which, as discussed, requires in-orbit refuelling with Starship (or any vehicle), but you have to carry all the fuel you need for the orbital burn to bring it back. The rocket equation just kills this immediately.
You really hope you have to get incredibly lucky that an metallic asteroid is on a near-intercept course with Earth that is just shy or going into orbit. The odds for that are, well, astronomical.
Assuming they deliver the same payload, sure, but that’s very much not the plan.
[1]: https://www.bloomberg.com/news/articles/2025-06-17/musk-s-xa...
No. If it is just $15B I can think of dozens different usecases ranging from military applications(fast transportation, it is the cheapest ICBM) to asteroid deflection to moon mining to science applications to space datacenter.
Are you seriously thinking $15B is big? Artemis by comparison has spent $93B and has cost of $4B per launch.
[0]: https://www.reuters.com/investigates/special-report/spacex-m...
[1]: https://techcrunch.com/2026/05/18/osha-probing-worker-death-...
Less recent that I thought: https://spacenews.com/explosion-mojave-air-and-space-port-ki...
Over 100 died building the hoover dam. Over 5000 building the Panama canal, and over 30,000 if you count the failed attempt.
30,000 people trip and die in their homes per year.
70,000 Americans die per year from medication error or accidental overdose.
Starship's design is so far beyond where the rest of the world is that even if it takes another 3 years of iteration to perfect it will still be by far the best rocket in the world for many years afterward, to the point where it would hardly make sense to choose to launch on any other existing or currently in development rocket for any reason other than political ones.
If a design with a bunch of modifications works, then it's a good design. Thinking you need to clean sheet redesign everything is how you get Second System Syndrome.