5014

Considering Falcon 9 launches cost in the neighborhood of $60-70 million, that’s a pipe dream.

There is no customer for super heavy launch services. I’m not sure what SpaceX’s business plan there is. I guess the assumption is that we’ll set up a moon base or a deep space station that will require regular construction and supply runs. I’m very skeptical. The only viable way to explore space is with robots. The only only economically viable resources in space are communication, surveillance, and (gulp) military services. Potential energy and/or mineral resources in space are not currently possible to harvest at anything like industrial scale (or any scale really.)

5015

That’s the price, not the cost. As of a 2020 interview with Aviation Week, the marginal cost to launch a Falcon 9 with a reused booster is about $15 million, around $10 million of which is the expendable second stage, and refurbishing the booster only costs a few hundred thousand dollars, comparable to the fuel expense. It’s entirely conceivable that SpaceX could hit a $10 million/launch goal for reused Starships.

What they charge external customers is another question, but as you say, there isn’t much of a market for super-heavy launch services at $3000/kg. SpaceX clearly can’t be planning to make money on charging the same as the legacy options (including, at this point, Falcon 9), so I suspect they want to make it up on volume. $3000/kg is one thing, with a defined set of plausible, economic uses. $500/kg or $100/kg are entirely different things, and what exactly people will make of such cheap access to orbit is unknowable until it starts happening, in my book.

5016

Well, there’s NASA. But also, the reason there’s no (other) customer is because they’re so expensive. If they can get the price down, even to Falcon 9/Heavy levels, then you open up a load of potential use cases that just aren’t viable at the moment.

5017

Excellent point. SpaceX does have a good track record of creating economies of scale in a market that doesn’t look like it has much demand. But sure enough, when the price drops the demand materializes. It’s spooky, I suspect invisible hands are at play.

5018

Is there more demand for launch services with Falcon 9 in play? I don’t actually know.

5019

Yes. SpaceX built the capacity and the customers have come. Falcon9 launches have increased almost every year, with the exception of 2019, since the rocket was introduced. They are happening basically weekly so far this year.

5020

Sure, but this might be zero sum. They might be undercutting existing launch providers and taking their business. Not a bad business model for SpaceX, but not creating demand. But maybe they are.

5021

Last year was record setting in terms of total launches. Well technically I think they heyday of ICBM development in the 60s set the launch record, but last year took the cake for actual orbital launches with useful payloads as opposed to weapon-test launches. This year is on track to eclipse last year.

In fairness, SpaceX does seem to be creating some of the demand that they have created the launch supply for. Their own Starlink initiative requires a ton of launches for all those satellites.

5022

If the price/kg ends up substantially lower than Falcon 9 then I would suspect they’d be pushing ridesharing launches to orbit. Sure most customers won’t need anywhere near the full capacity but if you use it to launch multiple satellites at once then that’s a different story.

Since they seem interested in retiring Falcon 9 I’m not sure what the plan is for customers that don’t want to share. Seems like they lose price competitiveness pretty quickly then.

5023

I hadn’t heard this. That’s odd. They’ve had like 100 commercial launches with Falcon 9. Only 2 so far with Heavy (though they have a few more scheduled over the next couple of years: mostly military and exploration contracts with NASA and the Air Force.)

5024

It may never be reality and may just be more of Elon’s ambition talking where he wants/expects Starship to be so much better as to make F9 entirely obsolete.

5025

They’re interested in retiring Falcon 9 once (1) Starship is cheaper both per kg and per launch and (2) customers trust Starship. They’ve basically promised to keep the Falcon 9 going as long as customers prefer it. They’ll likely just offer cheaper prices to encourage most to switch to Starship, once this is possible.

I think Falcon 9 will stick around quite a while even if Starship succeeds quite well given things like NASA’s desire for traditional launch escape systems for human passengers during Earth ascent, not to mention the difficulty in scheduling, safety, and range issues for doing many super-heavy launches at any feasible launch site.

I bet they’ll retire Falcon Heavy a lot sooner since it’s economically worse for them and not human-rated.

5026

SpaceX isn’t exactly burning their boats, but they do seem to be going all-in for Starship.

This is very important, and really we’ve not seen orders of magnitude cost reduction in many things.
Except for computing power.

The first Microprocessor the Intel 4004 turned 50 last November, and there were a number of celebrations/articles.

This is one of better comparisons I’ve seen between then and now.

CPU Model Intel 4004 (1971)* Intel Core i9-12900K (2021)*
Date Publicly
Announced November 15, 1971 October 27, 2021*
Price
(2021 Dollars) $401.41 $589.00
Price
(1971 Dollars) $60.00* $87.82
Max Clock Speed 0.00074 GHz(740 kHz) 5.20 GHz
(5,200,000 kHz)
Word Size 4-bit 64-bit
Cores 1 16
Threads 1 24
Memory Limit 0.000004 GB(4 KB) 128 GB
(134,217,728 KB)
Power Usage 1 W 125-241 W
Process Size 10,000 nm(10 µm) 10 nm
(0.010 µm)
Die Size 12 mm² (4 mm × 3 mm) 215.25 mm² (20.5 mm × 10.5 mm)*
Transistor Count 2,250 ~21,700,000,000

The transistor count 10 million times increase is most impressive, but what is more important is the cost reduction. You can get a 16-bit microcontroller (including on board ROM and RAM) for $.40 and an 8-bit Microcontroller for $.03 now. (pre shortage), that is 10,000 time cost reduction.

I can say talking to engineers who were there at the beginning, they had had a few dozen ideas of what people would use a microprocessor. A decade later, a few years before I joined the company, Intel embarked on a campaign (operation Crush) to find new uses for 8086/8088 microprocessor, and take business from Motorola. They manage to win about, 3000 new designs across several hundred different applications, including the IBM PC. The application that won the award for most original, was a device to help with the treatment of erectile dysfunction (ED). Basically, it used an 8088 microprocessor to measure how hard a guy’s boner was. Needless to say in a 98% male salesforce this was hysterical. funny.

I don’t know what the equivalent of the ED device in space will be, but a $100/KG I just know there will be one.

5027

In case anyone misses it, the key thing about SpaceX stopping Crew Dragon production:

“We are finishing our final (capsule), but we still are manufacturing components, because we’ll be refurbishing,” SpaceX President Gwynne Shotwell told Reuters, confirming the plan to end Crew Dragon manufacturing.

She added that SpaceX would retain the capability to build more capsules if a need arises in the future, but contended that “fleet management is key.”

I expect that they’ll similarly stop production on the Falcon 9 first stage while keeping some options open in the next few years but keep producing second stages and keep flying them for many years since they (1st stages) are so reusable.

5028

I mean, there’s only so many crewed flights in a year, so it’s not like they need all that many if they can get a good turnaround time.

5029

I guess you’re right, because this is confounding to me. With a microcontroller you can automate anything. They’re so ubiquitous I have to work to come up with control solutions that don’t use a microcontroller (because they tend to hate neutron bombardment.)

5030

The days when analog EE was king were very different. By the 90s is was getting hard to find new folks who were analog specialists. Everyone was doing digital for pretty much everything. At least in my neck of the woods.

5031

The Hubble Announcement: Oldest known star spotted by Hubble. It’ll most likely be put on a list for JWST to look at when it’s ready.

5032

Pft, I’ve seen older.

5033

Well if you are an involved in nuclear engineering, the costs of sensor and microcontrollers is very small compared to the overall cost of the project, and that was even true in the 1980s.
I was the product manage for embedded microprocessor, in the mid 1980s it was $99- headed down to $20 in few years, at the time the cheapest microcontroller were about $2. So there were plenty of applications in things, like cars, planes, industrial robots. At the consumer, level very little under $100 other than calculators was available.

I remember thinking it sure we be nice to have a digital thermometer because I always struggled to read,the damn things and breaking a mercury filled thermometer was dangerous. But thermometers were $5 at drugstore, so how could spend you $2 for electronics for $5 or even $10 product,. Yet today, it is really hard to find a non-digital electronics thermoter. (I just went through 5 pages on Amazon and found 4)