This question let me wonder whether SpaceX' Starship has a launch escape system. It really seems to have none. The obvious question is why not, it sounds pretty negligent to me not to have some kind of emergency security system. The Starship is so big, why can't it have launch escape thrusters like the Dragon spacecraft? I remind that the death of shuttle Challenger's crew could have been avoided if the shuttles had such emergency system, since the crew survived the initial explosion. If Starship isn't gonna have a launch escape system, I'd say they didn't learn their lesson from the Challenger disaster (but that's obvious anyway since the shuttles didn't get such system anymore until their retirement).

  • 1
    $\begingroup$ @geoffc It would be too much to have so many parachutes, for up to 500 people per airliner. But sailplanes do have parachutes. The shuttles and the Starship could (have) introduce(d) some too. If the Challenger crew had parachutes they might have saved themselves by jumping out. $\endgroup$
    – Giovanni
    Commented Aug 19, 2020 at 15:16
  • 6
    $\begingroup$ Related: space.stackexchange.com/q/33837/6944 That answer quotes Musk " You know, parachutes don’t work too well and [you can’t have] some standard abort system, and just how do you abort 100 people it’s just not feasible, the key is to make the spaceship itself extremely safe and reliable, and have redundancy in the engines, high safety margins and have [it be] well tested. Much like a commercial airliner. Like they don’t give you parachutes." Basically the same rationale as shuttle, as you say. $\endgroup$ Commented Aug 19, 2020 at 15:20
  • 2
    $\begingroup$ @OrganicMarble Your answer/comment is good too. $\endgroup$
    – Giovanni
    Commented Aug 19, 2020 at 16:35
  • 3
    $\begingroup$ @Giovanni: you are saying this as if it were self-evident, but it is not. In fact, in some sense, Launch Escape Systems are completely anti-logical: you are afraid that rockets aren't safe enough, and your answer is to stick even more rockets extremely close to the human passengers. $\endgroup$ Commented Aug 21, 2020 at 17:57
  • 2
    $\begingroup$ @Giovanni: Again, a LES is a rocket. It is a rocket that cannot possibly allowed to fail, that must work 100% under any circumstance with no exception. But, if you can build a rocket that never fails … then you don't need a LES, and if you don't trust your rocket, then "strap even more rockets onto the rocket" is by far not an obvious answer at all. Starship weighs three million pounds. A Falcon Heavy would barely be able to lift it off the ground, an actual LES would need multiple Falcon Heavys worth of rockets. $\endgroup$ Commented Aug 21, 2020 at 18:53

3 Answers 3


If Superheavy fails during launch (or even fails to launch in an unsafe way) the Starship itself might well be able to just light its engines and fly a suitable suborbital trajectory to a safe landing spot (assuming it wasn't hit by too much shrapnel). By the time Starship normally separates, it is much too high and fast for anything like a launch escape, so the only contingency left is that Starship itself catches fire, or similar, on the pad, or early in the flight, despite the fact that its engines haven't lit up yet. That might reasonably be considered a low enough risk to live with.

They could also fly each new Starship unmanned to orbit and back, as a proving flight if they choose to, which should shake out most manufacturing defects.

  • $\begingroup$ Although it occurs to me that Starship lighting its engines to get away from Superheavy close to the pad might not be too healthy for the pad. I imagine a fast fire with a full Superheavy propellant load would be something to see -- from a distance! $\endgroup$ Commented Aug 19, 2020 at 16:24
  • $\begingroup$ Thank you. You may want to include Organic Marble's citation of Elon Musk from above into your answer. $\endgroup$
    – Giovanni
    Commented Aug 19, 2020 at 16:36
  • 3
    $\begingroup$ Well, big liquid-fuel rocket engines with turbopumps (like Raptor) have some startup time. So in case Falcon Superheavy rocket explodes on pad I doubt it'll be enough time for Starship to fly out. In flight phase - maybe. Are there any links this variant was ever considered by SpaceX? Afaik - Elon Musk told about Starship+Superheavy like a commercial airplane analog - it will have no escape system but should have low probabiliy of fatal failure. $\endgroup$
    – Heopps
    Commented Aug 20, 2020 at 8:00
  • $\begingroup$ I wonder if the Starship could really escape a Superheavy explosion. Look at the size of that thing. Compare with the Soviet N1 rocket, which when it blew up caused "one of the largest artificial non-nuclear explosions in human history." To evacuate the crew from the blast radius with maybe a split second forewarning the massive starship would need quite an acceleration. $\endgroup$ Commented May 2, 2021 at 10:22
  • 3
    $\begingroup$ Starship cannot lift its own weight off the pad $\endgroup$ Commented Apr 25, 2023 at 13:56

You are not the first to raise this question! It's worth reviewing Tim Dodd's accessible and detailed evaluation of escape systems: https://everydayastronaut.com/starship-abort/

He draws attention to a number of considerations:

  • escape systems are not a panacea.

So in the grand scheme of things, to date, a mechanical abort system has only saved lives twice, may have prevented one tragedy and in one case caused a death. So out of the 320 orbital human flights to date, only three missions in total necessitated the use of an abort system, or less than 1% of crewed launches.

(And to paraphrase one commenter: if our concern is around the reliability and stability of rockets, it is ironic that our solution is "lots more rockets!". There are significant concerns about a) putting powerful, hair-trigger rockets and their fuel right next to the passengers, b) difficulty of testing, c) risks around carrying said rockets back through re-entry.)

  • the Challenger Shuttle disaster fatalities were caused by a series of failures at many levels.

perhaps the biggest problem with the Challenger disaster wasn’t a hardware problem, but a problem with program management and pressure to get that flight off the ground. It was known that they would be launching outside of the predetermined operating envelope of the SRBs and it was recommended to not launch that day.

(Of course, it doesn't matter what the cause of the failure is if you need to escape! But the big lessons lay elsewhere.)

  • trust is earned.

So that’s why I think it’s vital we see these things fly, fly often, and fly over and over. Only then will I think there’s a proven reliability record that would make it a safe enough option to not have an abort system.

A point that I think is rarely made, is that commercial airliners don't need escape systems because we can perform acceptance tests on them. Airliners are taken on test/commissioning flights before and during delivery to their customers, where their airworthiness is evaluated and certified.

You can't do this with disposable rockets: every flight is a maiden flight! And this doesn't exclude the Space Shuttle here, as the extensive refurbishment required after each flight would absolutely qualify for a test flight, if it were airliner, before passengers were allowed on board.

But with cheaply reusable and autonomous rockets, it's a completely different story. By specifically targeting low-maintenance and minimal refurbishment, SpaceX can do test flights and cargo flights until all the hardware and procedural issues are ironed out and the Starships can launch with an airline level of confidence. It is this capability that is unprecedented, and which is unsettling the established view of launch safety. Of course, they may fail in this goal, but that's why the whole grand saga has become so exciting lately!

  • 2
    $\begingroup$ @SusanW Yeah the Titanic is unsinkable, why put so many lifeboats on it? The huge cost, the delays and efforts, who needs that? And they might block the view. That's your and Jörgs argument. But in case of the BFR, an escape tower would really be not the best emergency system. Better to have such side thrusters like the Crew Dragon has. And parachutes per crew member wouldn't hurt perhaps, and might have saved the lives of the Challenger crew. $\endgroup$
    – Giovanni
    Commented Aug 22, 2020 at 18:05
  • 2
    $\begingroup$ @Giovanni: The Soyuz-U plus its 4 side boosters has less than a million pounds of thrust. Starship weighs 2 million pounds, so even two of them wouldn't be enough to have Starship just hover, let alone actually escape. The reason why the LES was needed on Soyuz T-10-1 was because we haven't figured out how to build such big rockets safely yet, and you are proposing to stick something at least three times the size onto Starship. The biggest rocket ever in the history of mankind was the Saturn V. Its first stage has 8 million pounds of thrust, which would give a thrust to weight ratio … $\endgroup$ Commented Aug 23, 2020 at 10:44
  • 2
    $\begingroup$ … of about 4:1, which sounds barely enough for a LES. So, you would need something as complex as a Saturn V first stage, or multiple Space Shuttle SRBs to even have a hope of getting Starship to safety. (And in all my calculations in all my comments so far, I have always assumed that the weight of the LES is 0, which is of course unrealistic, so an LES would in reality need even more thrust, because it not only has to lift Starship but also itself and its fuel.) Just es a point of comparison: Crew Dragon has 8 SuperDracos. Starship would need 125(!!!) just to hover, and probably up to 1000 … $\endgroup$ Commented Aug 23, 2020 at 10:47
  • 4
    $\begingroup$ … for an actual escape. The argument that I am making is not that a "Launch Escape System would block the view", the argument I am making is that the LES you are proposing is so insanely complex that, if you were capable of building it, you would also be capable of building a rocket that is safe enough to not need it. Again, the LES you are proposing has the same level of complexity as the most complex and most powerful rockets ever built in the history of mankind. $\endgroup$ Commented Aug 23, 2020 at 10:49
  • 2
    $\begingroup$ The idea is to have a standard platform that can take a large amount of crew and/or cargo. A separate capsule or compartment adds complexity and risk. $\endgroup$
    – Dagelf
    Commented Apr 20, 2023 at 22:24

Elon Musk has been mostly vague about launch escape on Starship, but he has indicated that he expects Starship to be reliable enough to not need it, other than a possible capability for the ship (second stage) to be able to pull away from the booster earlier than planned as a type of abort. The assumption apparently being that while the first stage might need escaping from, the second stage would not. Whether this rationale of making a space vehicle safe enough to not require launch escape in all phases of launch will be accepted by the regulatory agencies such as the FAA and DOT, and customers such as NASA and ESA, is unknown at this time. However even if Starship is ultimately not allowed to launch humans from Earth, it still has tremendous potential for many other tasks, including transporting or housing humans who are already in space, as well as revolutionizing the ability to launch large payloads at low cost and high frequency, which could transform space exploration over the next few decades.

One of Musk’s more detailed statements about launch escape was made in 2016 at the International Astronautical Congress conference (transcript timestamp 26:29):

the spacecraft itself is capable of aborting from the booster ... Launch abort on the spaceship itself is kind of pointless, if you’re on Mars you’re taking off or you’re not taking off. You know, parachutes don’t work too well and [you can’t have] some standard abort system, and just how do you abort 100 people it’s just not feasible, the key is to make the spaceship itself extremely safe and reliable, and have redundancy in the engines, high safety margins and have [it be] well tested. Much like a commercial airliner. Like they don’t give you parachutes...the spaceship could separate from the booster and fly away from the booster if there’s a problem at the booster level.

Musk recently reaffirmed this philosophy in an answer to a question about launch escape at the February 10, 2022 Starship update (timestamp 1:04:32) at Starbase, where he said:

Starship will not have an independent abort system. But I think something that would make sense is to have the thrust to weight of the ship be enough that it could take off from the booster even if the booster has a failure at the pad level, so you get the thrust to weight of the ship at sea level above one, then even if something goes wrong with the booster the ship can essentially fly away from the booster. And so that’s something that I think would be important for carrying people, and also for high value cargo to have the ship have thrust to weight greater than one even at sea level, that would be like the nine engine version, and then even if you lost one engine I think you should still be able to do an abort. So I think for crewed missions we would essentially maybe detank the ship to some degree so that would have kind of a launch abort capability with the ship even if you lost an engine. That would be my recommendation.

Musk gave no indication of how second stage engine pre-chill would be affected by this approach, whether for example pre-chill can be designed out of the system, or if second stage engine pre-chill would begin on the pad as soon as fuel starts being loaded (or prior to astronaut boarding), and continuing through the count and launch. Or would there simply be an acceptance that launch escape is not instantaneous but would take several seconds to initiate. These are details that presumably could be worked out.

However Musk’s overall philosophy of making Starship safe enough to not need launch abort capability during all phases of launch sounds strikingly similar to what was accepted in the original design of the Space Shuttle, as you alluded to, which had phases of launch, most notably the first two minutes, where any major anomalies during those phases would be inherently unsurvivable. As well as virtually no chance of survival (pre-Challenger) of an abort that resulted in a water landing. Another similarity with Starship is that Shuttle astronauts rode to orbit inside of what was essentially a rocket stage. NASA’s rationale at the time, similar to Musk’s, was that the Shuttle would be tested thoroughly and have multiple redundancies, eliminating the need for launch escape capability during all phases of launch. To the point that early Shuttle astronauts rode to orbit wearing only coveralls and oxygen helmets which provided no protection from depressurization. This half a century ago opinion about the Shuttle eventually changed once the vehicle came under increased scrutiny after the 1986 and 2003 accidents. Musk seems to be asserting that space technology has now advanced to the point that the original Shuttle philosophy of the 1970’s can be reinstituted. Whether the regulatory and space agencies will agree with this assessment is nearly impossible to predict at this time.

However Musk seems confident that it will be allowed, based on his well known public statements made over the past several years about building Starships capable of launching 100 people to Mars. As well as his announcement in 2018 of a project to send several artists around the Moon in Starship as early as 2023, as well as the (currently unscheduled) Polaris Mission III. And also including his vision of eventually sending 100,000 people to Mars during a roughly 30 day period every 26 months on 1,000 Starships, a number which doesn't seem to include the additional Starship fuel tanker and cargo launches that would be needed to support this. The viability of this futuristic vision of launching tens of thousands of people to colonize Mars is likely decades away and using technology yet to be developed, and thus is beyond the scope of the question or this answer. I am only attempting to answer whether a launch escape system will ever be needed on Starship for launching people from Earth.

What regulators would likely be influenced by is that currently the safest way to launch people to space is in a capsule, and this will not likely change in the foreseeable future. SpaceX would have to prove that Starship can be made as safe as a capsule before it would be allowed to launch people to orbit in it. During the Humans to Mars virtual conference on August 31, 2020 Musk said that he expects to do hundreds of uncrewed flights before launching people on Starship. Again the question is whether this will be enough to satisfy regulators. Falcon 9 has successfully landed 165 first stage boosters in a row without mishap. If hypothetically SpaceX decided to start selling seats to tourists who want to ride inside the Falcon 9 first stage, the first thing regulators would likely ask about is launch escape capability. If the answer is, “there is none but we’ve landed the first stage 165 times in a row without any problems”, that answer might not be considered adequate.

A capsule is small and sturdy and less susceptible to destructive aerodynamic forces than larger vehicles. Robust and sturdy heat shields that are less susceptible to damage are more feasible on a capsule. Capsules also provide a proven safe method of landing. Especially important is that a capsule has the ability to handle a much larger number of anomalies that might occur during all phases of launch and landing, including aborts that result in a water landing.

And while not a given, a capsule has a reasonable chance of surviving a first stage or even second stage booster explosion that occurs prior to launch escape activation. Two data points for this are the survival of the Challenger crew cabin after the breakup, and the apparent survival of the Cargo Dragon CRS-7 capsule after the second stage exploded. There is evidence that the CRS-7 capsule would have survived if it wasn't for the unfortunate, and in my opinion inexcusable failure of SpaceX to include a routine in the Cargo Dragon software to deploy the parachutes in a launch anomaly. This lack of foresight resulted in the loss of a large amount of expensive customer cargo for many countries and agencies, including an international docking adapter that was to be delivered to ISS.

Starship is a second stage which will have the capability to land propulsively. And this is where they propose to place the passengers, inside the second stage. Any failure of the second stage's propulsion or control capability during launch or landing that removes the ability of the stage to make a controlled propulsive landing would be unsurvivable. And any breakup or explosion of the second stage would be inherently unsurvivable. Capsules of course are not totally without risk, there is the possibility of a fuel tank explosion for example. However in space that would be true for all vehicles. Once in space it's a level playing field, where capsules have little to no safety advantage over other vehicles. But the risk of fatality caused by a capsule failure during launch and landing is less than it would be riding inside of a rocket stage that produces 1,500 tons of thrust.

Musk made a comparison to airliners not having parachutes, but the FAA may consider this an apples and oranges comparison with little in common. Considering that in the past ten years there have been only three people killed in U.S. airline incidents, out of about 80 million flights. Even if a 1 in 10,000 accident rate could be achieved in spaceflight, it would still be over 2,000 times more dangerous than flying in an airliner.

Adding an escape pod or something similar to Starship is unlikely because of the added weight, cost, and complexity. In theory a capsule with LES could be launched on top of Starship, but that in fact would confirm exactly what I am referring to. Any of these approaches would greatly reduce the number of people that Starship can launch to orbit.

But even if Starship does not receive approval to launch humans from Earth, it can still potentially carry out the dearMoon (sic) mission, send people to Mars, etc., by acting as a space ferry, with astronauts arriving in capsules at a space depot, or perhaps docking directly with an already fueled Starship. In fact this method is exactly what is planned for the Artemis III Moon landing, which NASA still maintains is planned to occur in 2025.

Musk’s belief that space travel should become as safe and reliable as airplane travel is certainly a goal that no one would disagree with. What is uncertain however is whether this goal is achievable, and if it is, whether the timeline to accomplish this will ultimately be measured in years, or in decades. And as applies to this question, whether Starship without launch escape capability could even come close to meeting this goal.

  • $\begingroup$ I wouldn’t say that Starship will not carry humans. Currently it is planned to land the next astronauts on the moon (Artemis 3). It is also planned to carry tourists around the moon (DearMoon). In addition, there are alternative methods of getting to space in the foreseeable future. For example the company spin launch wants to use a giant catapult to send payloads across the Karman line. Space tethers also have been under discussion and could see use in the near future $\endgroup$ Commented Nov 15, 2023 at 15:59
  • $\begingroup$ @TheRocketfan - I am specifically talking only about launch from Earth and landing on Earth. I mentioned ferrying people to the Moon as well as the planned Artemis landing. The dearMoon project was announced in 2018 with an initial planned launch in 2023. There is also a planned (using the term loosly) Polaris Mission III. I put little credence in announcements at staged events, conferences or interviews. However Musk is extremely reliable in tweets about current activities regarding Starship, Falcon, and Dragon.This is the Musk that I listen to, not the other one. $\endgroup$ Commented Nov 15, 2023 at 17:15
  • $\begingroup$ Personally, if I were SpaceX, I would launch Starship uncrewed at first, and launch the crew separately on the very reliable Falcon 9 (using Crew Dragon, which does have an escape system). Once Starship's reliability is established, only then would I start using it to launch crew. Why? If a Starship launch kills, say, a dozen astronauts over two or three launch failures, how long will it be before public pressure stops further flights, as happened with Shuttle? $\endgroup$ Commented Nov 20, 2023 at 4:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.