After a 24-second test of its engines, the New Glenn rocket is ready to fly | Hacker News Reader

After a 24-second test of its engines, the New Glenn rocket is ready to fly

arstechnica.com

113 pointsfenced_load143 comments

How much data do you guys reckon they gather during this 24-second window? Gigabytes? Terabytes? I imagine they have redundant sensors for every measurement, if possible. The LHC generates 1 PB/s, though it filters it down to roughly 200MB/s[0]. I wonder how this compares.

0: https://www.lhc-closer.es/taking_a_closer_look_at_lhc/0.lhc_...

I would imagine far less. The LHC captures many channels of sensors (for spatial resolution) and captures at very high frequency (because the time scales involved are fast and also because LHC is doing time-of-flight measurements).

I doubt that a rocket has anywhere near as many sensors (have you seen pictures of the LHC’s instruments? They’re basically all sensor), and I also expect that the timescales involved in rocketry are rather longer than in high energy physics.

Here’s a slide deck about ATLAS building an ASIC that reads something at 25 picosecond precision:

https://indico.cern.ch/event/799025/contributions/3486157/at...

Unless someone at Blue Origin is trying to localize a specific part of their flame by time of flight of light, I don’t see why time resolution even close to that would be at all useful. Perhaps they’re very fancy and want to tell which part of their rocket initiated an explosion by time of flight of sound, but that’s rather less demanding.

With the caveat, of course, that LHC events don’t explosively destroy the instrumentation. If you want useful telemetry in the last milliseconds before a rocket failure, you had better seriously harden your data logger or have very low latency transmission to a remote receiver :)

Combustion in the chamber happens at roughly the speed of sound in the F/O mixture. I don't know what the speed of sound in gCH4 is, but it's probably within an order of magnitude of air - air at 250 bar.

This is actually extremely important to model. Early F1 engines (Saturn V, not motorsport) were exploding and the engineers pretty much got lucky with the baffle design. Having a suite of sensors and then a computer model it would have saved lots of hardware and time - and really would have pretty much assured success. They were unsure if they'd succeed right up until they did.

Still way slower than what particle physics throws at you

(Wolphram Alpha gives me 743 m/s @ 250 bar and 1000C - could be wrong but probably the same OoM)

Yes, of course. I was clarifying, not disputing.

Speed of sound also depends on the temperature (i.e. speed of the gas molecules) which is high and variable. These kinds of coupled interactions make combustion instability really hard to model in CFD hence the importance of full-scale testing.

If you solve the formulas for calculating it, the speed of sound only depends on temperature. Mostly because temperature is dependant on pressure itself.

Temperature and molecular properties.. which guess what are also highly variable in combustion! Very messy stuff. Especially when you get into the actual nitty gritty of all the intermediate chemical species involved.

Not pure luck though, they tried 15 different baffle configurations. Would of course be easier to filter configuration in simulations.

There was enormous engineering effort put in with close coordination of physicists, but those 14 prior baffle configurations could have just as easily been 140.

For a relevant example, even just the oil formulation for Water Displacement on the exterior of the steel Atlas rockets took 40 iterations. Hence naming the product WD-40.

I wonder how many of the other baffle designs and water displacement formulations would still have been "good enough" - not optimal but usable?

At least for the baffles: None. That's why they kept going. Presumably there are even better possibly designs, but they were not discovered because one that works was found.

I don't know about the prior WD formulations.

You can fly to space with a 40ms control loop. It's a surprisingly low accuracy affair especially compared to particle physics. You're more likely capturing some actuator positions and discrete outputs and ensuring that everything sequences and responds to limits correctly, more on the level with an automobile CAN bus capture than it is anything else, and likely less than even 1 gigabyte.

There's likely more data stored in the video files from the cameras that observed the test than test data itself.

I'm just remembering that the images we got from Doves @ PlanetLab DWARFED the telemetry from the bus.

The system I worked on in the 2010s was pretty meager: 192kbits-per-second for the bus and about 54kbits-per-second (BITS, not BYTES) for the propulsion module. This is without video and doesn't count payload. It was also before we got the license for the beefy GHz S-Band, so we were sort of forced to make due.

But... you CAN get a lot of decent info from a low bandwidth link.

It depends on the number of PIDs and the speed of the data acquisition system. When I worked on the RS-25 engine project the engine had hundreds of PIDs, the low speed DAS was 250 samples per second, and the high speed DAS was 250,000 sps. The high speed DAS generated so much data that it was only started a few seconds before engine start so the recorders wouldn't fill up before the end of the tests (which were typically 500 seconds).

EDIT: something to add is that not ever PID was tied into the high speed DAS–only a couple-few dozen important PIDs.

Formula 1 car has over 300 sensors and generates 1.5Gb a lap.

So I would imagine this is generating hundreds of megabytes.

You are going to be limited by what you can transfer over radio.

I would wager the actually important data is far smaller. A single camera would probably dominate throughput

presumably the goal is to have the smoothest and most consistant combustion, and therefore less vibration, and known ,predictable resonant frequencys in the system. Which makes very small increments of time important.These rockets are consuming ? tons/sec of fuel, and then its a question of how "small" of an inconsistancy will break things. A simple but true engineering pricipal, is that any noise, is proof of ineficiency, and therefor a full examination of the audio frequency , coupled with video of the flame.Other data that might be nice would be to know exactly what is happening inside the turbo pumps,and fuel lines. The whole thing is so violent, that resonant phenominon not seen anywhere else, could be doing very weird things inside tanks and lines. And sensors for that, especialy video, could be challenging to engineer. Here is a wierd fact about a high pressure diesel fuel line: the pressure pulse (-+ 10000psi) causes a travelling expansion in the fuel line, that then causes a magnetic pulse, which can be detected, but you want to be quick about it, eh!

can confirm.

Not much, really. Most sensors are low speed: tens to hundreds of Hz. Then you have some high speed sensors, like strain gages, that are up in the 30+ kHz range.

I'd be surprised if it was more than a gigabyte. I work in spacecraft operations and the sampling and telemetry rates are not that high. At most they'll be monitoring a few hundred parameters at the order of 10Hz. That doesn't add up to a lot over 24 seconds. The biggest contribution to the data budget is most likely to be the video.

Probably gigabytes. When I worked with other systems in the past we needed to upgrade our gigabit ethernet to handle the full-rate telemetry

Do you mean upgrade to gigabit or from gigabit to 2.5Gbps or 10Gbps?

1 gigabit wasn't enough for full rate telemetry

> Terabytes

Terabytes to petabytes. Much is noise. But you’re already making sensor-level keep/discard decisions due to the magnitude of the deluge.

Note that this includes cameras, of which modern telemetry includes many.

We're at 12 years of design and 5 years behind the originally announced first launch of 2020 now. Space is hard, rocketry is harder and I wish them luck, but Blue Origin has an awfully long way to go if they want to catch up to SpaceX: this was meant to be the Falcon 9 killer, but odds are Starship will be fully operational before New Glenn completes testing.

Starship will serve a different mission profile. People seem convinced Starship will just "solve" space, but if it works, there's still going to be demand for other mission profiles that it won't make sense to send on Starship.

But if your point wasn't to say that it will be obsoleted by Starship, and just to say instead it's slower development than Starship, yeah, that's true.

I suspect the head start in infrastructure spacex has is pretty valuable in developing new programs.

Space is hard. I hope Blue Origin succeeds.

I hope Blue Origin succeeds too, but I'm not sure which mission profiles wouldn't be served best by the Starship in comparison to other launchers. There could be some ultra-light ones, where the cost for the whole launch is less than refueling the Starship, but...

Obviously any mission that goes beyond LEO that does not aeorbrake at the destination, because that means you're expending an upper stage with aerodynamic surfaces for no reason.

Also anything that does a direct TLI or TMI, because you will have to stop for in-orbit refuelling.

The cost to get to LEO dominates in most, if not all, launches, and it's relatively trivial to add a booster stage which solves additional delta-v problems while being cost preferential to alternatives.

Maybe SpaceX will make those stages, if the market would interest them. Maybe fellow companies like Impulse Space would manage those requests. Starship still looks like a probable winner in the area.

Huh? If Starship can cheaply lift 150T to LEO, you can allocate some of that 150T to a propulsion system.

That means the customer has to buy an additional propulsion stage (plus time for integration testing and insurance in case it damages the launch vehicle, in case it becomes reusable). Also, it is not yet clear what the LEO payload capability of the vehicle is.

The cost of an expendable third stage is surely more than made up for the brute force cost savings of being able to launch 150T to LEO for the cost of fuel and a car wash.

I think it depends on the situation. It is not yet clear if 150T to LEO is real (and even if it is, you're likely sharing it with a bunch of other customers who might not want to fly with another rocket motor unless it has gone through quite a bit of vetting and paperwork; Shuttle banned the use of solid rocket motor 3rd stages after the Challenger accident) and it is not yet clear what the cost for a typical customer is going to be. Rocket engines designed to start in a vacuum are more difficult and expensive to develop and test than engines designed to operate closer to STP.

Starships model seems to hinge on being able to catch the booster every time and the second stage most of the time. I’m not sure if this is a Tesla Model 3, or a Cybertruck.

People had doubts about Falcon 9 landing on a barge every time...

I think they’re asking the question: “Is there really value in the product” or is it a gimmick like the cyber truck.

Which mission profiles does Starship not make sense for?

Starship is heavy lift / heavy expense. If you want to put 1 small-sat in orbit, you don't use Starship, you use a Rocket Lab Electron. It's MUCH cheaper. Or you wait until you find 600 other small sat users who want to launch into the same / similar orbit.

The talk I heard was that New Glenn was supposed to be BO's answer to Starship (or more likely Falcon Heavy) that could launch a bunch of Project Kuiper satellites into LEO so Amazon could compete with Starlink and feed Amazon's Ground Station as a Service offering.

If we are to believe the published numbers, New Glenn can lift 50 imperial tons to LEO, Starship Block III will hoist 200 tons and Falcon Heavy will lift 50-60 depending on how re-usable you want your launch to be.

I'm not a heavy lift sales-person, but I've been in the room when they discussed what they thought they could sell to govt / mil / commercial customers. So take this with a little bit of salt... Seems to me BO was targeting a slightly smaller launch vehicle than SpaceX was going with so they could decouple schedule with Amazon's Kuiper Project. You don't want to have that cool new rocket you developed dependent on a satellite constellation that gets delayed. So you have a rocket that would be easier to fill with a number of small to medium sized satellites to LEO or (fewer) to GEO.

And like other people on the thread have commented, it seems BO is a decade behind SpaceX, so... yeah... a big rocket that competes with Starship is pretty risky for BO.

And yes, I understand that BO is independent from Amazon, but from what I've seen this is just so they can execute on a schedule that isn't determined by AMZN's board of directors. They seem pretty closely related, just from talking with Kuiper, GSaaS and BO engineers.

I don't work for any of the above mentioned companies and have no insider information. YMMV. Just my guesses from watching some of the personalities involved for the last 30 years.

- "If you want to put 1 small-sat in orbit, you don't use Starship, you use a Rocket Lab Electron. It's MUCH cheaper."

I don't buy this. I think small startups like that can't get the economies of scale that would let them compete on price, for any payload. So long as they are targeting low-value niche markets like one-off smallsats, they won't have the revenue to support that.

At what Rocket Labs is currently charging, $7.5 million [0], it's within the realm of possibility you could even launch an entire reusable Starship with a one-cubesat payload for less than that. (The target figure Musk uses is $2 million/launch; take that with the appropriate bucket of salt).

How many tens of billions of R&D have gone into SpaceX, and how many launches are they able to amortize that cost over? How many decades have they invested in their manufacturing processes? Do their competitors' engines roll off factory assembly lines?

[0] https://en.wikipedia.org/wiki/Rocket_Lab_Electron

I think the $2m per Starship launch is "asperational." As in... they "aspire" to have launch costs that low. I do not believe Starship launches are currently as low as $2m (or $10m, which is the other figure I saw mentioned.) I believe their current launch cost (cost to launch the rocket, not the price they charge people to launch their payload) is considerably higher than $50m.

But... yes... if you can re-use your launch vehicle, then per-launch cost SHOULD go down. But at the current time, after 6 launches, only one booster has been snagged by the chop-sticks. The Starships themselves have NOT been re-used.

So if you're going to compare Starship with Electron, you should compare costs of Starship after it's fully re-usable with Electron after it's as re-usable as it's intended to be.

Rocket Lab claims they're at 7.5m per launch, but again, THAT price may come down as they use re-usable components on subsequent launches. So instead of comparing costs of Starship sometime in the future with Electron now, compare Starship sometime in the future with Electron sometime in the future.

F9 reusable is like $70M. $7.5M is already changes. Starship at $2M is not an innovation class event, it has to be some side effect of WW3.

Check out this great tweet on Elon's $2m/launch: https://x.com/Tim_X94/status/1864279927014740327

"As an auditor I review public company financial statements. These financial statements are prepared in Rocket Lab's case based on US GAAP. Adjacent to US GAAP figures companies also provide NON GAAP disclosures for informative purposes. Formally these are not audited, but auditors sanity check them anyway.

Since Adam and Pete are CEO and CFO of a public company, each time they give you a number in interviews / press releases etc., they will be based on US GAAP and if explicitly stated NON-GAAP basis.

Why is this relevant Tim?

Well if Adam and/or Pete give you a number, that number will includes ALL the (estimated) costs in accordance with US GAAP for what it takes to build something. In this case Adam/Pete on numerous ocassions told investors roughly the following:

We aim to have 50% gross margins on Neutron with a similar flight cadence as Electron. Electron has 50% gross margins at roughly 20 annual flights. Adam specifically also said they will probably arrive there even sooner with Neutron since its first stage + fairings reusable and Electron till this date is expendable.

With a $55 million sticker price and a 50% gross margin, the cost of sales under US GAAP will rougly in the $25-30 million ballpark based on a cadence of 20 flights. NOTE: If Neutron launches more then 20 times per year, cost of sales are going to even drop lower to the point where the build costs rougly resemble those of Falcon 9.

But for now let's keep it simple (20+ flights is the bull case scenario) and let's go with the 20 flights base scenario. In this case it costs between $25-$30 million to build/operate one Neutron rocket. Under US GAAP the main cost components to operate a rocket launch include:

Per flight: > Fuel costs for the rocket > Fee to the launch range > Seperation system

Per build: (allocated based on the number of annual launches) > Material costs for building the rocket > Salaries for the factory employee hours > Machine hour costs > Depreciation for factory buildings > Transportation costs

Per launch facility (allocated based on the number of annual launches) > Depreciation for launch infrastructure > Salaries for the launch staff > Security for the launch range etc.

SPB and Adam also said many times, that the per flight costs including material costs to build the rocket are only a FRACTION of the total cost of running a rocket program. The real costs are the fixed ones relating to factories, facilities, employees, depreciation etc (mainly the launch facility and build mentioned in the paragraph above) AKA the JOKE ADAM RUNS IS WE RAISE MONEY TO POUR CONCRETE. Why does he make that joke? Well more then 80% of the total rocket program cost are not related to the LAUNCH VEHICLE ITSELF

Now let's go over to our friends at SpaceX. Remember SpaceX is a private company and in the US there are zero reporting requirements for private companies nor obligations for CEO to quote US GAAP approved figures. Instead since Elon is not bound to any of these regulations, unlike at Tesla for all his SpaceX endavours he employs something what I will going forward refer to as Elon GAAP.

The most important rule of Elon GAAP is that there are NO accounting rules.

Let me illustrate that with this example:

Elon was quoted on multiple occassions about what Starship would cost to build. Basically he said it is the long term goal for a Starship launch is to cost $10 million. This $10 million figure is based on the following assumptions: > A Starship is fully reusable and will assume aircraft like operations. > It takes no refurb between flights between each vehicle, similar like an aircraft. In order for that to happen the heatshield tile issue will need to get solved, but ok let's go with his narrative.

As such if you assume the above Elon then says that the only costs for each flight that you will have mainly relates to fuel costs and if you build 100s-1000s of Starships each build will not cost a lot. Elon estimates this to be $10 million per flight.

Caveat are we really comparing a 20 flight Neutron cadence with a 1000 flight Starship cadence?

Yep we are which is totally insane it itself and non apples to apples comparison, but let's go with the leading narrative on X.

So what your finfluencers and SpaceX fanboys on X do is they compare the Starship $10 million number to Rocket Lab's $25-$30 million number and conclude Rocket Lab's Neutron is way to expensive to build and will run out of business long term.

Why is this not correct then Tim?

Well thats because we are comparing Adam's US GAAP cost to launch with Elon's Elon GAAP cost to launch. Its not an apples to apples comparison.

Why?

Because under Elon GAAP you only have to account for direct rocket material and fuel costs and don't have to account for these costs:

Per flight: > Fee to the launch range (assuming a Cape launch) > Seperation system

Per build: (allocated based on the number of annual launches) > Salaries for the factory employee hours > Machine hour costs > Depreciation for factory buildings > Transportation costs

Per launch facility (allocated based on the number of annual launches) > Depreciation for launch infrastructure > Salaries for the launch staff > Security for the launch range etc.

Why you don't have to account for these costs Tim?

Well you see, under Elon GAAP all the employees and suppliers work for free and buildings and launch infrastructure remains in perfect condition and never has to be replaced.

Well Tim Elon GAAP must be wonderful right?

Yes, it truely is an amazing place.

Especially since a large part of the costs of running a rocket program are not directly related to the rocket itself AKA the largest part of the costs are not included in Elon's 10 million number)

After Adam read this he is probably also going to apply Elon GAAP for the Rocket Lab financial statements. This means he just has to account for material and fuel costs for each Neutron launch and can leave everything else out. Under Elon GAAP a Neutron rocket launch will the same or less then $10 million (assuming a 20 Neutron cadence and 1000 Starship cadence), because Neutron is a significantly smaller vehicle then Starship and as such way cheaper to fill. This is partly offset by fact that the second stage is expendable for Neutron.

So the conclusion of this accounting rant is:

No matter the GAAP (US GAAP or Elon GAAP) Neutron will almost always be cheaper to operate then Starship on a dedicated ride basis. You just have to do an apples to apples comparison. Sure Starship will be the king of price per kg, but Neutron is not in the price per kg business, but in the DEDICATED RIDES BUSINESS. For other this will become more obvious in the years to come, when Neutron will ramp cadence well above 20 flights per year.

Why?

Because Starship is a significantly larger vehicle and rocket program costs don't scale LINEAR, they grow EXPENONENTIALLY with the size of the vehicle. How else can the entire Starship program cost $10 billion (Payload estimate) versus Neutron $300 million?

Also Starship is optimized to go to Mars, Neutron is optimized to require minimal infrastructure (no launch tower and other optimizations due to vehicle size. Remember infrastructure is the largest cost of a rocket program. SPB is very smart and exactly knows what he is doing. So don't let yourself get fooled by finfluencers and fanboys.

Next time you see someone quoting Elon GAAP for rockets, you can refer this post :)

Full disclosure: I love Elon, as a Tesla shareholder, I just don't like Elon GAAP "

I am going out on a limb and say that Project Kuiper satellites themselves are nowhere near ready. Amazon is supposed to launch 50% of the constellation of ~6000 into orbit by July 2026 or risk losing frequencies. So you would think there would be some urgency.

Amazon purchased pretty much all remaining Atlas 5 launches from ULA. This is a proven rocket and ready to fly. Why aren't the satellites being launched? The only thing I can think of is that they are not ready yet.

I am unfortunately skeptical. As you say theyve got two test units in orbit and every day the frequency allocations tick closer. Whats the launch schedule look like for 3000 successful units inside 18 months? What about the in house build process. I havent heard anything that reflects the exponential process growth & delivery theyll need.

That first real launch also starts the recurring costs ticking on the limited orbit lifetime of that hardware. I havent thought through the numbers in a while, but something like a billion or three per year to maintain the constellation? Swag (an optimistic?) $2m satellite and $2m launch cost and 7 year lifetime, thats an yearly average of $1.7b in “maintenance” to keep the very minimal constellation up there. Easily double if their

Im also pretty skeptical of their business. AFAICT its a bunch of ex telecoms and space/defense contractors. So theyre going to try and soak US DoD for connectivity with a more uh, “reliable”, company and a consumer side of “space comcast.” Im pretty skeptical on consumer space “broadband” due to the density problems. And I use comcast as a perjorative for their business & network interop coming from CDN & ISP land.

Lastly on a positive note, I dont expect the same employee resignation bloodbath that AMZN at large is going through at the moment. Kuiper (afaik) has been pretty top down “you shall come in to the office” the whole time, so any rto mandate is unlikely to change the existing experience.

>If we are to believe the published numbers, New Glenn can lift 50 imperial tons to LEO

Well 45 tons, but this is in a reusable configuration.

>Starship Block III will hoist 200 tons

This is definitely not a reusable configuration. Maybe 150 tons if they are lucky.

>Falcon Heavy will lift 50-60 depending on how re-usable you want your launch to be.

60 tons for Falcon Heavy means zero reuse, that is a fully expendable launch. Falcon heavy also hasn't carried payloads heavier than 18 tons so far. So this number is something you can whip out to pretend that New Glenn sucks and yet completely miss the mark.

I was going off data from the wikipedia, so yeah, it may not be the most current or most accurate. It's in the ballpark, though.

And most of what I care about is small-sats, so Falcon Heavy's are more than enough lift for me. But yeah, other people may have have a completely different set of requirements and see the Heavy in a different light.

<https://news.ycombinator.com/item?id=9792046>

I think the cheapest smallsat launches now available are being pushed out of the Japanese airlock on the ISS. It's very popular.

Really limits your options in terms of height/inclination but it's been popular enough that they're almost at capacity.

Aye, NanoRacks is pretty cool. I often bump into some of their people at conferences. They're cool and competent as well.

> If you want to put 1 small-sat in orbit, you don't use Starship

Why not? Unless you need a custom orbit that nobody else is interested in then Starship will be by far the cheapest way to put a small satellite in orbit, as part of a ride-share mission.

Who do I talk to at SpaceX RIGHT NOW to launch my small sat?

Sales? Accounts receivable?

Give me a phone number. Electron and NanoRacks have phone numbers (email, actually.) I can contact them and get info on dates and costs. SpaceX will give me info on Falcon launches but not Starship launches (and they pick the orbit.)

rideshare@spacex.com

I just googled SpaceX rideshare, it was the very first result.

Rideshare doesn't give quotes for starship.

No shit. Does Rocket Lab give quotes for Neutron?

Smaller launch to a specific orbit. As an analogy, freight trains are one of the most efficient, cheapest ways to move freight. But moving my laundry to the laundromat is better served by another vehicle.

This changes if SpaceX can get full reusability working, which they have been making very steady progress towards.

Yeah, that's the main proposition Starship has for the market. Going to analogy with laundromat, you need to include the cost of single-use laundry cart and remember that your laundromat is on another end of the same asphalted road which the truck nearby can easily drive on.

Not sure at all if costs - and hassles, of course - of buying a cart are less than some change for the costs of the fuel - the truck is autonomous, of course.

10t to LEO in an unusual orbit.

Being smaller than Starship while still huge isn’t necessarily a disadvantage.

Starship should be able to put 10t into LEO significantly cheaper than New Glenn can. Why would anybody pay more?

I believe this is factually inaccurate. What's your source? Though admittedly, we have to depend on BO's predictions for per-kg launch costs, and those may be a lot higher for the first few missions and gradually ease off.

Google's AI tells me the current costs for a Starship launch is somewhere between $100m and $2b. Wikipedia says a Falcon 9 costs about $50m and can lift about 20t to LEO. I see a blurb that says Musk says Starship launches will get down to $10m each. But... that seems like an "asperational statement." He also said Full Self Driving Mode would be available in 2018, 2019, 2022 and 2025. Not trying to take away from the absolutely cool stuff his companies have done, but it seems like it will be a while before it costs $10m to launch a Starship.

This link from 2 years ago estimates a New Glenn launch costing $68m. I have no idea how accurate that number is. But if we're going to use Musk's "asperational" cost estimate for Starship launches in the distant future, we should let BO use an "asperational" figure as well.

https://www.cnbc.com/amp/2022/04/05/amazon-signs-rocket-deal...

Those Starship costs you quote assume an expended booster and the New Glenn costs assume a recovered booster.

We don't have to use Musk's cost figures for Starship. Starship has been built in the open and can be relatively accurately cost estimated by experts.

Ok, so your argument for Starship not being the best choice for all mission profiles is based solely on cost. If Starship was cheaper, then you'd agree that it serves the mission profiles?

An estimate of 2 billion per launch is laughable, and suggests you are not arguing in good faith. 100m is more accurate for a fully disposable launch, and SpaceX has demonstrated great progress on reusability of the booster, which will cut costs considerably.

100m is vaporware pricing. It's the $30k Tesla that drives itself. Or any of the numbers on the hyperloop. It's a made up number for the press.

I'm begging the internet to please be critical and do some basic analysis and not just believe everything they hear from that guy!

Are you arguing in good faith? What are you based on to say $100mil is vaporware? Most commercial flights (i.e. non-Starlink) are priced starting at $70 millions - [0], increased from $60m previously. That’s not for the press, although government customers such as NASA often pay much more. SpaceX is very dominant at this point that they’d be foolish to charge under cost.

And then there are Starlink launches. They made money on it on 2024, according to Shotwell, so launch cost must be way lower than external price.

[0] - https://www.spacex.com/media/Capabilities&Services.pdf

SpaceX is mass-producing the engines, mass-producing the heat shield tiles, the fuel is cheap, the structure is stainless steel (cheap), payload capacity is huge, and full rapid re-usability seems well under way. There is no way Starship will not be the cheapest launch platform (per kilogram) if and when it is operational. It will definitely have a launch price tag for under $100m.

But it's not just that guy.

> Starship rocket to less than $10 million. However, Starship is still very much a development program, and Payload estimates it currently costs around $90 million for SpaceX to build a fully stacked Starship rocket. The vast majority of this cost goes toward the rocket's 39 Raptor engines and labor expenses.

So it's going to be somewhere over $100 for a fully disposable launch. What happens when they start reusing the booster? What happens when they have optimised production further?

Are you sure that your anti-musk bias isn't clouding your judgement?

https://arstechnica.com/space/2024/01/rocket-report-a-new-es...

$100m might be correct though. Elon Musk himself has said that just the hardware for each Starship test flight has cost $100 million. This doesn't count all the hardware that hasn't flown so in practice each test flight was even more expensive than this, but there is no reason to argue that Starship will cost more than $200 million for an expendable launch and maybe a quarter of that assuming reuse via booster catch.

2 billion at the high end isn’t actually unreasonable when compared to falcon 9’s costs which are sitting around 100m/launch right now largely due to inflation.

There’s a have a fairly linear relationship between rocket payload and size, and for large structures going big tends to increase cost per pound so ~10x the size resulting in ~20x the cost is just mildly pessimistic.

If and only if they the thing is both rapidly reusable and individual starships are actually used for hundreds of launches do those highly optimistic numbers become vaguely possible. Even just a 0.2% failure rate would represent a massive increase over their optimistic estimates.

> when compared to falcon 9’s costs which are sitting around 100m/launch right now largely due to inflation.

SpaceX's financial situation argues very differently. They have raised relatively little money for a company that is spending multiple billions on two very expensive development programs (Starship and Starlink).

If Falcon cost $100M per launch the 134 launches this year would have bankrupted the company. The $1.7B they raised in spring 2022 was their last major capital injection, and have been self funded since.

If Falcon cost substantially more than $20M to launch SpaceX would need to be getting external money from somewhere. They aren't. Their revenue is well understood and is around $10B per year, and salary costs fot 13,000 people are going to consume most of that. What NASA and the Space Force pay is public knowledge, what they charge for a private launch is known, and the number of Starlink subscribers has been revealed.

SpaceX has several million Starlink customers providing around 6.6 Billion dollars of revenue in 2024. It not clear if it’s profitable yet, but it’s been stated to kick off 100’s of millions in positive cash flow.

As to the salaries of its employees, that’s a major component of launch costs. You can’t point to it and say launch costs must be cheap because they are paying all these people when a large fraction of them are directly or indirectly working on launches.

They are spending ~2 billion per year on Spaceship, but what they charge per launch varies widely. 5 crewed falcon 9 flight cost the government ~260 million each, and the 2 ISS missions where 145 million each. https://payloadspace.com/predicting-spacexs-2024-revenue/

2 billion is ridiculous, and I can only imagine that number was a misunderstanding SpaceX/Musk saying that they were spending 2 billion in a full year of R&D on Starship.

https://spacenews.com/spacex-investment-in-starship-approach...

That doesn’t justify why it’s ridiculous, it’s just a coincidence.

I doubt SpaceX’s internal costs are ~100m/falcon 9 launch, but companies need a markup to be profitable. 100m - 2B is a huge range covering everything from giving up on reusability and paying back R&D over a small number of flights to significant success resulting in a 90% reduction in costs per kg to LEO.

Also, having spent 5B on R&D and doing 5 test flights up to this point works out to 1 billion per flight. That’s not the actual marginal cost per flight, but when people say how expensive each shuttle flight was that’s the number they use. Nothing guarantees they continue to do Starship launches, they could fail it’s among the potential outcomes.

It's ridiculous because the much ridiculed SLS has a launch cost of 2 billion dollars. If you think SpaceX is throwing billions of dollars into developing a vehicle that costs thisuch to launch, you clearly haven't been following SpaceX at all.

You know there is going to be more that 5 flights, and you know people in this thread are not amortizing total R&D into flight costs. People are talking about 68 million per flight for New Glenn, which no doubt has has many hundreds of millions on R&D spend, and hasn't flown one time.

> You know there is going to be more than 5 flights

No, I don’t actually know the future. I can make predictions, but we could have a thermonuclear war tomorrow etc.

Yes. An estimate for 2 billion is laughable. I did not say the cost WAS $2 billion, I said the Google AI gave me a range from $100M to $2B. Maybe the total cost of the program for the 1st launch was 2 billion? But you're going to amortize that cost over (hopefully) several launches.

I think you misunderstand my argument. Let me restate it.

Someone, sometime said the Starship launch was $2b. The Google AI picked that up and included it in its answer. Someone, sometime said it was around $100m. The Google AI picked that up and included it in its answer. There is a lot of range between 100m and 2b, which implies there's a lot of data getting thrown around and we don't have good numbers.

If observing that we don't have good numbers is arguing in bad faith... I don't know what to tell you.

Musk at some point said $10m for a Starship launch. I think I found a reference for that in a CNBC interview... I'll look it up later. But my point is... It is unlikely that Starship launches are $10m RIGHT NOW. But sure... maybe they will be in the future. I take Elon with a grain of salt because of his comments regarding Full Self-Driving Mode and Robo-Taxi deployment dates.

I said we should not compare New Glenn estimated launch costs RIGHT NOW with Elon's asperational price target of $10m. We should compare Starship's cost per kg to LEO RIGHT NOW with New Glenn's estimated cost per kg to LEO RIGHT NOW. Or we could compare them at a particular point in the project history. We could compare per-kg costs at first launch or estimated per-kg costs at the 10th launch.

Both companies are saying they want to do a lot of launches, so we'll eventually have MUCH better data.

I'm suggesting we compare apples to apples and oranges to oranges and not apples to oranges.

At the current moment, all Starship launches have been fully disposable (though yes, one booster was caught by the chopsticks so it's probably more accurate to say the whole system is about 1/12th re-usable.) At this point in the program, you have to pay for each vehicle that lands or crashes in the water. I agree with you when you say "100m is more accurate for a fully disposable launch." Starship is currently more disposable than it is reusable.

When SpaceX re-uses the boosters and the Starships, then it will not be fully disposable and the price per launch will go down. We are not at that point at the moment. You can tell this because a number of boosters and starships have fallen into the ocean, some crashing, some coming to a controlled stop just over the ocean and then falling over.

But the important part here is that the equipment that wasn't caught by the chopsticks doesn't get to be re-used. So if you want to do another launch, you have to build new equipment. That new equipment will cost money.

So if the current, mostly non-reusable Starship launches cost $100m a pop, that's after several launches. Even though we have someone estimating the first couple of New Glenn launches cost $68m, let's wait until it has 6 launches and THEN compare costs.

> He also said Full Self Driving Mode would be available in 2018, 2019, 2022 and 2025.

He did. He also said it'd be available in 2016, 2017, 2020, 2021, 2023 and 2024.

So yeah...

Dudes. You don't have to down-vote me 'cause I'm just asking for where you got your numbers. I'm not saying you're a horrible person and that SpaceX sucks. I'm saying I have numbers that don't match yours. Let's compare numbers / sources and see what the most likely values would be near-term and long-term.

Non-orbital. But it can do that as well. Just sacrifice the rocket and use the landing fuel for more delta-v.

I’m not sure which profiles don’t make sense unless they can’t figure out payload doors. Stage 2 reuse means you can put up 50 tons or so of stage 3 in LEO for approximately free in current payload pricing terms, it’s a crazy number. (Advertised capacity of 100-150 tons is probably pointless and/or volume-limited anyway for a third stage.)

Not all profiles end up in orbit. At this point in time, reusuability sacrifices delta-v for fuel. If you want to launch scientific missions you want to escape Earth orbit. If we're going to land on Mars, we're going to need a lot of stuff sent there first.

But it doesn’t matter if you can launch 10x the mass for the same price to LEO and light your stage 3 to get anywhere you’d like to be. Stage 3 will have enough delta v and starship’s doesn’t matter as long as it gets you to the parking orbit.

Is Starship really not following the exact step of N-1 && VentureStar?

Computers and materials came a long way since the N1, as Starship’s successful tests can attest to.

But they're still losing roughly as many engines as Russians did in N1, so that sounds like a dubious claim.

What about the latter? Are they really not tracing the footsteps of the X-33 program?

What? Flight five didn’t lose any engines. N1 kept blowing up.

X-33 never got to a test flight, let alone a successful one.

N-1 only flew 4 times. Flight 5 of N-1 was supposed to use upgraded NK-33 engines.

X-33 never got to test flight because the engineer minded NASA director kept pushing an unrealistic technical goal.

Besides, SS still hasn't gotten its payload system working.

Are those that dissimilar to how SS program is going?

Yes, they are. Starship has proven multiple novel engineering approaches and technologies. Examples: 1. catching the booster in the air to save on landing gear weight and complexity 2. firing 33 engines simultaneously in a tight cluster 3. the only large scale production and use of CH4-fuelled engines (all other CH4 engines have only been fired a handful of times at most)

N-1 failed because of a lack of time/funding, not because the design was inherently flawed.

But that wasn't certain until Starship proved it technologically possible.

> but if it works, there's still going to be demand for other mission profiles that it won't make sense to send on Starship

I'm rooting for any and all US launch providers to succeed, but I don't think this is true. Starship at full reusability will be better than any other launcher for every single mission profile imaginable.

There's simply no way one vehicle will be the best option in every case.

If you have a small satellite you need placed somewhere unique, firing up a huge launch vehicle makes no sense.

It does if it's still cheaper than any other rocket.

I don't see how that makes any sense. Starship is 100t dry; simply the fuel costs of a launch will necessarily be higher than disposable alternatives.

Fuel is cheap: it's $900,000 per Starship launch, according to Musk [0]. No disposable rocket comes within even a factor-of-10 of that.

[0] https://techcrunch.com/2019/11/06/elon-musk-says-spacexs-sta...

I was perhaps exaggerating a little with the fuel cost comparison, but we've had rockets like Astra R3 and SS-520 fly for under $5M per launch. Rocket Lab's Electron is ~$7.5M. That's all within a factor of 10.

fuel is cheap

most expensive part are the engines

Your launch price may somehow be cheaper; although, that's incredibly unlikely. Anyways as your insurance underwriter I'm going to jack up the rate on you so high for that launch it will no longer be cheaper.

Starship’s launch price is highly likely to be cheaper if New Glenn’s second stage isn’t reusable.

It's also a bit of a straw-man to assume people will only fly on the cheapest lift option. Virgin Orbit lasted as long as it did only because there was one (maybe two) customers who were willing to pay a premium for being guinea pigs for a new technology. Their reasoning was something like "if this pans out, it'll be very cool long term" so they were willing to put their small-sats on a largely untested platform.

The same could be said about some of the entrenched players in earth observation. They're willing to pay a bit of a premium for a reasonable amount of time to ensure there's not a monopoly player (which definitely looks like it will be SpaceX.)

How much of a premium they're willing to pay and for how long seems like anyone's guess.

The rocket equation doesn't participate in capitalism -- moving mass to orbit takes delta-v and that takes fuel.

If you are going to imagine a hypothetical future where starship has made technological leaps forward sufficient to be the cheapest possible option despite being significantly heavier and larger aerodynamically, you have to imagine someone else could also improve their rocket. A smaller rocket requires less fuel to fly.

The US is not the only place flying rockets, and spacex has a lead, but if the industry takes off, there will be other contenders. Once rockets start getting more similar as they all start contending with physics, a smaller rocket will necessarily be cheaper.

This. If, and it’s a big if, starship can be 100% reusable and less than 24 hours turn around, then it will be the space truck NASA envisioned with the shuttle

Even if it's not 100% reusable it's still likely to be cheaper than New Glenn. New Glenn uses extremely expensive methods of construction. Starship is assembled like a glorified water tank.

I'm not sure I see it that way, I wonder sometimes how much of the expensive construction can be systematized into something less expensive, and how much of Starship's "cheap" construction will remain as they work harder and harder to get to their reuse and turnaround goals. My intuition is that Starship will get more expensive to make over time and New Glenn less expensive but beyond that I can't say if they would meet in the middle or cross.

That said, platform construction costs only dominate when you can't re-use the platform. Anything you can re-use gets amortized over each re-use. That is what had made Falcon 9 so cost effective. Mostly because they get nearly 10 flights per booster.

> That is what had made Falcon 9 so cost effective. Mostly because they get nearly 10 flights per booster.

You’re way out of date. Multiple boosters have flown over 20, a couple are at 23.

https://en.m.wikipedia.org/wiki/List_of_Falcon_9_first-stage...

I guess 10 flights per booster is an average over, say, the whole time since the first booster successfully landed. In the list by this link there are ~70 boosters, and I don't think Falcons flew ~700 missions yet, so 10 flights per booster looks on average approximately correct, even though some - even quite a few - boosters flew significantly more that 10 times.

And recently, many payloads had their booster fly exactly once because the payload needed some extra 'oomph' to get into orbit so the booster was going to fast and to high to re-enter. When you are pricing for an average of 10x booster re-use you need boosters to go much more than 10x because one booster going 1x is going to bring your average down quickly.

I'm looking at overall re-use numbers. Yes, boosters are now being certified for more launches which will, over time, bring both the average and the median number of launches per booster up.

In this conversation, I think the relevant point is that as the 'resusability' of the F9 booster has gone up, so has the cost to make a single booster. That's because they've added things and changed how they make them in order to boost re-usability. I expect the same evolution in Starship/Booster which will increase the unit cost in order to make them more reusable which will lower overall cost of launches because you can amortize those costs across multiple flights.

> ... Falcon 9 ... nearly 10 flights per booster

Just curious: what year is your data from?

There's no way they're getting rapid turnaround with the damage it's has sustained on the last two tests.

I’m always amazed by statements like this. Starship is currently a test program and is rapidly evolving. They’re improving it as they learn.

What you said is akin to looking at the Wright brothers plane and saying “no way that thing crosses the Atlantic.”

It will improve rapidly until it does.

I agree w/ Chuck. It's a BIG IF. I'm hoping for them. I really hope BO turns out to be a good competitor for SpaceX so they're both highly motivated. I've seen the sausage being made at a couple of new space companies (and at least one old space company.) Even more than chip vendors and PC clone manufacturers, Heavy Lift providers are their own worst enemies. If SpaceX fails, it will be because of something they decided to do, not something that was foisted upon them. Ditto for Rocket Lab and Blue Origin. But the cool flip side of that is they're free to do extremely cool/innovative things. The guys who are privately held aren't beholden to a board that wants increasingly dependable revenue and profits (looking at you, ULA.) And this does not seem to be the industry that can tolerate that. I'm bullish on SpaceX, BO and Rocket Lab (assuming Rocket Lab doesn't run out of runway.)

>odds are Starship will be fully operational before New Glenn completes testing.

What does this even mean? Fully operational? Starship has three versions and they are still testing the first one which isn't supposed to reach orbit in the next flight nor is it supposed to carry any payload, not even a mass simulator. When you ask people why the booster hasn't been reflown, you get this confusing answer that the booster is "already obsolete" even though they have planned to launch three more "obsolete" boosters after the first successful catch.

Everyone is bragging how fast SpaceX is, but they are starting to drag their feet. It's like those people who build a demo that looks like the product is almost finished, but it turns out those were the easy and visible 80% that you can show off, now you're left with the hard and time consuming 20% and you're going to run into delays like everyone else.

And then there is the fact that New Glenn is going to launch on 5th of January and attempt landing on the first flight. Barring an explosion on the way to orbit, New Glenn will be flying at least half a dozen missions carrying payloads throughout 2025 including a moon landing of Blue Moon MK1.

Your comment comes across as pessimistically predicting the failure of the first launch or being ignorant that it will launch in four days.

"the easy and visible 80% that you can show off"

There was nothing easy about the Raptor engine, for one. It is absolutely the best rocket engine in the world by far, and the only methane-based engine that ever reached space.

AFAIK the only "real" problem that SpaceX is now having with Starship is the heat shield vs. rapid reusability. It is an important problem, but it also means that many other complicated problems (such as precise exercise of the belly flop) are fully solved.

> the only methane-based engine that ever reached space.

TQ-12 that is used by Zhuque-2 reached orbit first. Of course Raptor is much better, as TQ-12 is a simple gas generator engine

> New Glenn will be flying at least half a dozen missions carrying payloads throughout 2025

12 missions is their upper limit of capacity, so it’s “at most 12”.

https://x.com/jeff_foust/status/1835610703174255074?mx=2

Do you have some references of New Glenn being positioned as a Falcon 9 killer by any industry professionals, commentators or press with credibility (at the risk of no true Scotsman)?

This was certainly thrown around with Tesla, but it's not something I've personally come across with SpaceX.

Regardless, if semiconductors have taught us anything over the last decade, strong competition is essential for a healthy market. It's hard to imagine a true (haha) fan of space exploration that isn't cheering on Rocket Lab and Blue Origin (as you are), even if they're destined to forever be runners up. Even if you believe Musk will operate SpaceX entirely selflessly, he won't be in control of it forever.

Boeing will probably buy one of the also rams to improve their offerings

The moment SpaceX is sold to Boeing is the moment Musk doesn't get to go to Mars. And while he may never be able to have the martian colony he seems to covet, he still seems to believe he can do it, so there's zero motivation to sell.

Also... Boeing? They've been having some issues lately.

I was thinking that Blue Origin was a more likely target.

Can an also ran buy another also ran? Does that even make sense?

It’s got more than twice the payload capacity of Falcon 9 to LEO and a 7m payload fairing vs 9m for Starship.

The real question is if there’s going to be enough demand to justify these systems. With enough reusability it might make sense to fly these things 2/3 empty, but that’s only going to be so profitable.

New Glenn's main customer will be Kuiper, Amazon's answer to Starlink.

But also whatever Amazon's equivalent of NanoRacks is 'cause Amazon also wants to feed their GSaaS (Ground-Station as a Service) play. I think long-term Kuiper is going to be BO's main customer, but selling lift capacity to a NanoRacks-alike could reduce the impact of a Kuiper schedule slip.

Starlink already scaled to 1M US households using Falcon 9 launches and this is twice as large. I think it’s reasonable to question just how many launches a competitor would support.

> Blue Origin has an awfully long way to go if they want to catch up to SpaceX

Maybe we should just give the whole space industry to SpaceX because obviously nobody can touch them. /s

They seem perfectly good at doing that all on their own

Blue Origin is a funny company. I worked for TerraBella/SkySat/Google Satellite, Planet Labs and Kubos, each as a consultant/contractor/whatever. I struck up a conversation w/ Rob Meyerson @ a conference and told him what I was working on (mostly comms / ground station work, but with a tiny bit of GNC test infrastructure.) I mean, I'm not God's gift to rocket scientists, but... I was pretty heavily recruited by them but was working at Amazon and didn't want to leave, then spent some time nursing a family member through cancer treatments. Finally I'm ready to talk to them and I start with back-channel chats to make sure they still have open reqs. They do so I informally talk about specific things I've done and what I might do at BO. Finally I put my resume together and submit it to the hiring manager I had been talking about, emphasizing the bits he said he was interested in. 15 minutes later I'm rejected for the position they had been recruiting me for.

I mean... I'm not a SUPER rocket engineer, but I'm pretty solid for the things I did work on. I've slogged through design meetings where I had to analyze protocol specs and make sure we agree'd on details, wrote code, wrote A LOT of tests. I mean, I'm solid. The only thing I can think is I don't have a Ph.D., but that DEFINITELY wasn't a requirement when Bob was running the shop.

Folk have told me it's evolved into something much more like Amazon where each team optimizes it's tiny bit and teams communicate only via APIs and the only opportunity you get to optimize complete functional or value chains is when something breaks.

Just seems a bit weird they went from "we're 10 guys in a hangar" to "we've re-implemented Amazon's small-team/local optimization religion" in less than 10 years and with less than 1/30-th of the number of engineers.

I wish them the best. I think SpaceX really needs some decent competition to focus their collective minds. But... they've gone weird.

I'm probably too senior and too "weird" to them to get hired there, but I absolutely encourage young engineers interested in an intense experience to check out their jobs page.

There's several funny stories of their hiring oddities on HN going back at least 10 years... For myself, I only talked to them about a back-end SWE role, not avionics, but they were the only company to ever ask about my college GPA. (Not good.)

You worked for Kubos? How is it? We were using KubOS the linux OS for our cubesats when they went bankrupt (?) and got acquired by someone else

I really enjoyed working with the KubOS team (though I was more on the Major Tom side.) They were a great bunch -- everyone involved on the project has my highest recommendation.

I was definitely worried when I heard KubOS (the company) was struggling. I don't know too much about Xplore, but I think I saw Tyler was working there now, so there's SOME continuity at least.

KubOS (the codebase) doesn't look like it's received any love in the last couple of years, which is sort of sad. Some really bright people poured a lot of time and effort into it. Just looking at the contributors list: Ryan, Catherine, Kyle and Tim... all very bright people and a pleasure to work with.

Was this before or after Limp joined?

It was about two or three years ago for me. Not sure when Limp joined (actually... not sure who Limp is.) Oh wait... google is your friend... David Limp... new CEO... Looks like he joined in December 2023. So this was before that.

I overlapped with David at Amazon and had a few interactions with him and some of his direct reports. Don't know if that's good or bad; pretty sure I didn't embarrass myself there.

I'm a big SpaceX fan, nonetheless I am looking forward to some competition for SpaceX. If New Glenn succeeds, it might become a workhorse like Falcon 9/Heavy. And don't forget New Armstrong is the next generation, should be in the works.

I'm sure that as the new pseudo-president he will avoid doing anything that might sabotage SpaceX competitors.

It's sad that this has to come into it, but given both of the companies CEOs "kissing the ring" so to speak, is there really any competition here, or just politics now?

Good luck to them. SpaceX is great, but healthy competition is always important at driving progress. I'd like to take a trip to space one day, and even if SpaceX nails rapid reusability, the best chance of them being motivated to pass on launch savings is a competitor hot on their heals.

I would also love to see Blue Origin spend more time on building space habitation. If Starship does bring heavy lift costs right down, I want to see all the interesting things that people start putting into space

I wish them all the best. I really want to see humans land on the moon again. We'll bring humanity together again. One big step for mankind.

I hope they succeed!

I wonder if Jeff Bezos also hoses-down Blue Origin workers with ice-cold water to passive-aggressively attack demands of livable wages and sensible working conditions for Amazon warehouse workers.

Reply on news.ycombinator.com