It looks like so far we are still well off the pace, yet the market has not traded any lower.

The longer the beginning of the time span spends below the curve, the harder it will be to catch up.

To get to 500 before 2030, SpaceX need to on average increase the number of launches by about a factor of 2.28× each year starting from 2 launches in 2023.

Whilst only half the launches so far have been successful by this market's criteria, counting unsuccessful ones anyway as indicative of the trend, so far we have:

<year>: <launches> (<launches at 2.28× per year>)
2023: 2
2024: 4        (5)
2025: 2 so far (10)
2026: -        (24)
2027: -        (54)
2028: -        (123)
2029: -        (282)

So we need 8 more launches in 2025 (~1 per month) to achieve the needed average annual growth rate.

@chrisjbillington

1. Is the 282 figure for a single year something that is reasonable? That is a lot of launches and it feels like it might start to bump against both supply and demand barriers. How does it compare to established demand on existing rockets? Is it more about a forecasted demand based on the capabilities?

2. Let's just say they miss the targeted 10 in 2025 and instead end up with 6-8. Do you think this would have a genuine effect on the 2026 trajectory compared to getting the full 10?

@Eliza Falcon 9 and Falcon Heavy are doing the lion's share of all launches globally, and together they did 134 launches in 2024. Starship may be cheaper, and for missions going beyond Earth orbit (e.g. the moon, mars), there'll need to be multiple launches to do orbital refuelling. So missions like those could bump the numbers up. But I'd still think 282 in a year is quite unlikely unless they get the price down enough to open up access to a much larger market than that for which it currently makes sense to spend money on launches.

I guess 6-8 this year would seem to me like a bad sign compared to 10, yeah. Like, it would indicate an overall delay so that whatever trajectory they were on, they are now behind by some fraction of a year.

I don't think a constant growth rate model makes sense past a few years, such that even if they're "on track" I expect growth will slow when the launches get close to Falcon 9's cadence. But if they slow enough to not even keep up with the average growth rate needed, I think we should be able to pretty confidently expect they won't speed up enough to make up for lost time later.

@chrisjbillington @Eliza
This year likely finished at 5 rather than 10 in your table. It has been a bad year re flight 7 8 9 and ship 36.

How are the calcs looking now? I would suggest
<year>: <launches> <successful launches> (<launches at 2.2× per year>)
2023: 2 0
2024: 4 4
2025: 5 2 but initial rate considered to be lower of year/build time and year/pad turnaround time=min (365/30.5, 365/15) = 12
2026: - (26)

2027: - (58)

2028: - (128)

2029: - (282)
(4+2+26+58+128+282=500)

Ha, ended up with same 282 for 2029 by assuming a higher start rate (because 2025 success was so poor) and consequently needing slightly lower growth rate.

Growth rate spurts possible?
At present they have to build each ship as they haven't got reuse of ship. Ships taking a month+ after last launch to finish test and get ready but no ships nearly ready now so next launch likely 2026. Pad 1 also takes a couple of weeks to prepare. If they get rapid reuse of ships and boosters and pad 2 likely does not take anywhere near as long between launches then this build time constraint goes away as you add ships to the fleet that can fly again and again.

Another build rate growth spurt might come when Gigabay(s) are in operation. Probably reasonable to say only 8 stands in the Megabays at present. 2 Gigabays with 24 stands each could allow a much faster build rate once ships are coming off the end of those lines though that might not be until ~mid 2028 and maybe by then they might have enough to rotate and keep launching at the pad turnaround rate. So maybe it is extra pads that allow faster rate and they are planning 5 launch towers and maybe some more catch only towers.

At least 28 ships and boosters can be built in ~28 months starting in 2026 and a higher rate after mid 2028. 10 launches a year per ship seems like it could be possible. 282 launches from current 5 planned launch towers & mounts might be closer to being a limit in this timeframe but I don't think it is completely implausible.

However, then there is the demand side limits. How quickly customers transfer from Falcon 9 to Starship may depend on the price SpaceX charges. So far this year 91 Starlink launches 37 others in 9.5 months. At the same price the others might do 50 launches a year. While the mass of these might be within limit of 10 starship launches I doubt the volume and 10 orbits would be sufficient so there might be demand for 20 or 30 starship launches a year with similar price per launch. Maybe that increases a little with price cuts and more fitting on a single launch. But it is not approaching 300+ rate without SpaceX launches and it is hard to recover lots of development costs without both a lot more launches and a decent price per launch.

There are 2 possible large scale numbers that SpaceX might want to launch: Starlink and maybe Mars. They can probably do 20 times the falcon 9 bandwidth per starlink launch with Starship. Therefore it will almost certainly be cheaper per bandwidth launched by using Starship so it may be more profitable for SpaceX and more bandwidth should be launched. However how much more is hard to quantify. If 20 times more is justified then it is the same rate of launches about 120 per annum. However the way internet bandwidth grows rapidly it seems hard to rule out 50+ times more bandwidth being potentially usefully deployed especially with premium phone high connectivity and low latency being a possibility. You have to be pretty optimistic for this to be within reach so soon.

Elon has talked of 1000 starships to Mars per window (or even per year?), but I see a few problems with counting on too many of these for the demand. First talking about 1000 is easy when you can say well I was talking about the rate in ~2050 or 2070 or .... Actually doing so prior to Jan 2030 is much harder. There aren't going to be many if any in the Nov Dec 2026 window. The Jan Feb 2029 window is in the period for this question but is the only one and I am not sure you are going to keep launching through the rest of 2029 as the next window isn't until Mar 2031. More importantly who is going to pay for and do all the development work for the cargos to be sent when they almost certainly haven't demonstrated they can land on Mars with Starship before the time the development work needs to be done? SpaceX may be able to afford and find suitable cargos for say 30 starships for 2029 window but hundreds seems more difficult for anyone to justify spending the amounts involved if landing on Mars hasn't be demonstrated. It is also worth noting that the starlink ships can be rapidly reused, but this is not the case for Mars Starship and the higher build rate allowed for by the Gigabays doesn't leave a lot of time to build many mars Starships. If you require 6 refuelling launches for each of those 30 mars ships then you are getting up to 180, a decent proportion of 282 launches a year but this all seems very optimistic as to whether it is possible.

The willingness to invest in the facilities to make it possible might be a sign of SpaceX's confidence in how rapidly they expect demand to grow. However in the before 2030 timeframe of this question, I suggest the only way you get there is with a lot of optimism for Starlink internet service and rapid expansion of services to phones.

@ChristopherRandles 26 feels extremely optimistic for 2026! Even just missing by a few means more and more catchup.

@Eliza Yes it is very optimistic. Requires assuming they get reuse pretty quickly and other things to go well. They may struggle to build and complete 12 ships and ~3 boosters in 2026. The first few ships will only launch once. If they catch one in May then they are able to learn lessons, design upgrades to parts that need it, and if they can all be retro fitted to ships in the production line then maybe you can imagine adding a ship a month to the fleet starting in August so you get 2 launches in Aug, 3 in Sept, 4 in Oct, 5 in Nov and 6 in Dec so 20 in last 5 months of 2026 and even with that, 6 before Aug is also a struggle and all that is needed in order to reach 26. Almost everything if not everything has to go right for this. BTW I am also assuming no refuelling campaign for a mars ship at the end of 2026 because that assumes the ships built would have to be tankers and there would only be test launch(es) with these in Sep and Oct, so the ships built would have to be pez dispensers and able to be reused within a month. A launch every 5 days from a single pad might be asking too much but a second and even third pad may be coming online by the end of 2026. Testing facilities might also need to be expanded to achieve that rate depending how much testing before reuse.

58 in 2027 seems relatively easy in comparison to that timeline as above is already past that launch rate. So maybe there is some room to underachieve on 26 in 2026 and still have a viable path. However, I suggest you would still need to be an optimist to believe in 500 before 2030.

Basically the question should have a low probability, but just how low?

@ChristopherRandles when is your forecast date for the next launch?

If they launch a tank of space fuel into space, how long can it...sit there...before being used?

@Eliza March 2026 maybe? That seems possible but even that might be a bit optimistic hence a struggle to do 6 before Aug.

Boil off is a problem. That is, if cryogenic liquid oxygen and methane fuel warms up it evaporates and significantly increases in volume. The pressure becomes too much for the tank unless you vent the excess pressure. There are various ways to deal with this including: lots of insulation, sun shade to reduce insolation, leave ship/depot/tank well away from moon and earth re reflected and reradiated heat (this is why Artemis plans on using NRHO rather than lunar orbit), and adding refrigeration to condense vapour back into liquid.

So how long it can sit there depends on how much effort you put into doing these things. You probably have depots orbiting. You likely leave off all the flaps header tanks and re-entry heat shield so it cannot return to Earth and add lots of above to these "depots". Tankers likely have the re-entry equipment but don't need much if any boil off reduction as the launches plan to quickly put the fuel into a depot in a very low earth orbit. The depot carries the fuel higher if that is needed. Mars ships will likely need some or all of the boil off reduction equipment that the depots have to do a 6ish month trip to Mars and still have some fuel on arrival in order to land.

What is not at all intuitive is that it takes far more fuel to go to the moon (and back) than to go one way to some of the outer planets that are much much further away. A full ship load is probably fine for going to Mars but probably inadequate for going to moon, landing on moon, launching from Moon and returning to low Earth orbit/land so you likely need some refuelling in lunar NRHO, or a highly elliptical Earth orbit or maybe a pusher/depot or extra tank on nose of the lunar starship or ...

Did this fuel question have some relevance to the likely launch cadence? Were you wondering about mars ships and fuel stockpiling launches in 2029 after Feb for the 2031 window? Not sure, maybe if the boil off minimisation equipment works well you could fill up lots of depots well in advance Perhaps requires more depots that way but this might not be a big issue. I think you would tend to want to launch the mars ships nearer the time in order to give time to do the development work on the cargos. Lots of roll out solar arrays could be launched early but you might want mix the cargos across the ships so if you lose a ship or two you are not out of any one of the (essential) things you are sending. Maybe attempting to land low necessity items first on the higher risk landings until you gain more experience can alleviate this concern somewhat.

Hope this is useful.

Considering Falcon only flies 60 a year and Starship still has some years of development I'm going to say no.

@Riley12 Falcon has flown a wee bit more than 60 times this year...

@DanHomerick I have no clue where I got the number 60 lol

Do refueling flights count?

@Martlet definitely, they would meet the criteria the same as any other flight.

I updated the description to match the proposed new criteria and unpinned the comment. I'd have left just the tally pinned if I could, but it doesn't make sense to keep it in the description + that comment is super long (and now largely redundant).

I'm borrowing from the original creator's criteria:

I think we have to fit the same criteria to the 1st launch that we do to the 500th. I'm considering the first IFT "Not Successful", in that it didn't insert Starship into a near-orbital trajectory. Even though personally I think it was a successful data gathering flight. The guiding principle for success is something like "Did this flight prove to customers their payload will go where they want it to go?".

and will say that the guiding principle will remain "Did this flight prove to customers their payload will go where they want it to go?". If this means at some point flights are being made that reach 99km by design, the 100km rule could be revised. I trust this will not be the case...

in defence of the 100km rule, from the original description:

I'm considering the first IFT "Not Successful", in that it didn't insert Starship into a near-orbital trajectory.

Thanks @chrisjbillington and @ViniciusMoreira for your considerable effort in making this market better!

If there are any standing objections, let me know - I am not an expert

I read through some (not all) of the comments. I think it is reasonable to adopt the criteria outlined in the pinned comment -- assuming it really does include the special case suborbital flights that meet

The guiding principle for success is something like "Did this flight prove to customers their payload will go where they want it to go?".

from the original criteria

I'll give an opportunity to respond if you like, I'd rather everyone agreed so we can move on, but if anything is insufficient lmk before I change it.

I don't know who is counting though, 500 launches seems horrible to keep track of and if there is no good source that tracks this criteria we may have trouble in 5 years if it's close to that number..

I figured surely somebody is counting, but I had no idea how right that was: https://en.wikipedia.org/wiki/List_of_Falcon_9_and_Falcon_Heavy_launches

Not that wikipedia is a good resolution criteria, but it is thoroughly cited

All seems reasonable to me, I'd support adding the clarification to the description.

I would like to propose a more objective and straightforward method for resolving this market to minimize disputes. Using an independent organization, not affiliated with Manifold or SpaceX, such as the FAA, can provide clarity and fairness.

Each time a Starship Super Heavy + Starship launch occurs, it adds 1 to the count of total launches. If the FAA initiates a mishap investigation for any of these launches, that flight will be considered a failure and will be subtracted from the total count.

If, before January 1st, 2030, the count exceeds 500 successful launches, the market resolves to YES. If the count does not reach this number by that date, the market resolves to NO.

I have previously commented on why I believe Chris Billington's clarification method is not ideal. Therefore, I am proposing this alternative method, which I believe is more objective and fair for all parties involved. This approach ensures that success is determined by an impartial and independent body.

Here is the FAA definition of what would be defined as a failure.

The new FAA regulations describe nine events (see below) that would constitute a mishap (14 CFR 401.7). The occurrence of any of these events, singly or in any combination, during the scope of FAA-authorized commercial space activities constitutes a mishap and must be reported to the FAA (14 CFR 450.173(c)).

• Serious injury or fatality

• Malfunction of a safety-critical system

• Failure of a safety organization, safety operations or safety procedures

• High risk of causing a serious or fatal injury to any space flight participant, crew, government astronaut, or member of the public

• Substantial damage to property not associated with the activity

• Unplanned substantial damage to property associated with the activity

• Unplanned permanent loss of the vehicle

• Impact of hazardous debris outside of defined areas

• Failure to complete a launch or reentry as planned

@ViniciusMoreira

> Unplanned permanent loss of the vehicle

This one is specifically allowed to happen in some situations by the prior criteria, so there would need to be a strong consensus to override that one IMO?

It seems like the original intent of the question is really heavily focused on "payload goes where it is supposed to go", to a degree that makes some of the reasons the FAA might consider a launch a failure a little bit less relevant. I wonder if that one would be different enough to constitute a separate question entirely.

Is there a chance they launch some of these vehicles within this timeframe in a manner that somehow does not involve FAA jurisdiction?

---

I don't want it to look like I am stepping on Chris's toes here either. I did respond to your other comment since Chris's original post does say he'll allow other moderators to intervene if there are controversies. Maybe we need to get an additional moderator involved, then.

Regardless, it's much better to figure this stuff out now, than in 3 years or 9 months or whatever. It's good that we are trying to figure out if the current count is 1 or 2 or whatever, and more objective criteria. Maybe they'll easily do 600 in which case it was all a waste of time, or maybe they won't even get to 400. But if it's 'close', then this stuff matters.

@ViniciusMoreira I appreciate the desire to make criteria more verifiable, and if the market were being made from scratch, perhaps counting non-FAA-mishap launches would be a good criterion to base a market around. However, FAA mishaps and successful payload deliveries (or demos of such capability) are simply different things, and this market is about the latter. You can have unsuccessful payload delivery and yet no FAA mishap, and you can have successful payload deliveries (or plausible demos thereof) despite FAA mishaps.

As for your point below about the 100km criterion, it's a space rocket and 100km is the most common definition of space. Any launch that could plausibly have a customer or be a test of something a customer might want will definitely be reaching at least 100km. That very much includes point-to-point flights on Earth, which I deliberately wanted to include with these criteria.

This is not a picky criterion. In recent flights, at the time Starship had reached 100km of altitude, it had only travelled about 130km downrange. After stage separation in the most recent flight, Starship ignited its engines at 72km of altitude, with about 0.75 km/s of vertical speed. Ignoring wind resistance, without igniting its engines, Starship would have continued upwards an additional ~30km over the next ~80 seconds, reaching an altitude just above 100km, before falling back to Earth. So a launch that only reaches 100km is one where Starship doesn't even ignite its engines.

The only launches that are excluded by this criterion are low-altitude tests that are not a trajectory any customer would want, or close to what any customer would want. For example, there were many Starship low-altitude test flights before Starship first flew with SuperHeavy. The 100km criterion is simply to exclude out such low-altitude test flights, which I don't think the original market intended to include, since it defined success as delivery of payload to "intended orbit or landing zone for interplanetary bodies". Previous low-altitude flights are excluded by the original criteria already by virtue of not involving SuperHeavy, but I think it's unlikely the creator would have wanted to include similar flights that do.

I'd be happy to replace the altitude criterion with something else, but you need something to exclude test flights that aren't a trajectory any customer would want. And "no mishap" isn't a good criterion, since SpaceX could do some low-altitude tower catch tests or whatever and have it go off without a hitch, and yet it would not be a demonstration of anything a customer would plausibly want - which is about taking a payload somewhere.

For what it's worth, I don't think we're going to see any successful Starship–SuperHeavy flights reaching less than 100km of altitude ever. If we remove the criterion I think it won't make any difference, so I don't think it's the case that it's a significant change. But on the off-chance there is some low-altitude test flight, it's good to have a way of excluding it without us having to argue about whether a customer would have wanted it. I think this very conservative altitude criterion is a good threshold - it would be a very weird customer who charters a space rocket to go a couple of hundred km without leaving the atmosphere when they could have taken a plane.

@chrisjbillington

it's fine not to use my definition of success, but I still believe your 100km criterion is arbitrary and was not in the initial question. However, if you include Earth-to-Earth the flights as shown in this video https://www.youtube.com/watch?v=zqE-ultsWt0 , regardless of the altitude reached or whether they have passengers or not, that works for me. If not, I want moderators, I took a look at this market.

Starship | Earth to Earth

Starship will be capable of taking people from any city to any other city on Earth in under one hour.

@Eliza
As I mentioned in a comment to Chris Billington, it's fine with me not to use my definition if the types of flights shown in this video are included: https://www.youtube.com/watch?v=zqE-ultsWt0

Regarding the resolution of this market, I believe it will be very unlikely to be 'close.' It's either going to be 'embarrassing' for the YES side with fewer than 100 flights, or it will be the opposite with over 700+ flights. It all depends on when the slope of the curve starts to go exponential.

Starship | Earth to Earth

Starship will be capable of taking people from any city to any other city on Earth in under one hour.

Here will go
https://manifold.markets/ViniciusMoreira/will-the-market-starship-reach-500

@ViniciusMoreira those flights will be included, fear not. They will be greater than 100km in altitude, I promise you. There is no plausible customer-relevant Starship flight that doesn't involve it reaching at least 100km.

The 100km criterion was chosen in order to include such flights (and other technically suborbital flights) whilst excluding low-altitide demos such as the trajectories of SN8,9,10,11, and 15.