[Q&A+] Data Centers in Space. Genius or Just Dumb?

What I don't understand about putting data-centers in space is how can it be cheaper to put solar panels in space than to install solar panels on earth?

Yay! I got to dust off the old Lagrange bingo card.

Hi Fraser, I think I've missed something here. You keep saying things changed 3 months ago. What happened? Cheers, Edward

It’s 100% going to fail the commons. This is literally all of human history. Same shit different day. What’s also 100% from history is that elite heads will roll eventually - literally, and it’s already started. Maybe humans deserve what we get.

If you want to guarantee leverage over governments with the only recourse being to laser satellites to death when billionaires get too uppity, then yeah - it’s a great idea.. at least until the skynet becomes alive and you need to go to space to turn it off because it’s literally plugged into the sun… what could go wrong? 💭

Mass drivers on the Moon don't need to point up. You just need to clear the immediate horizon. Actually much better that way if you want to go into lunar orbit.

"Overly cynical"?!? I recognize both those words...

Congratulations on 400 Q&A episodes.

Pol

I feel like I’ve seen at least 300 of them. I Never get tired of the question show (or any of the others). This channel is my only Patreon subscription and it’s more than worth the money I pay.

Fraiser Cain replaced by AI? Hmmm. Well, he does only ever show his head in these videos, so there is a very Max Headroom vibe... Maybe Fraiser already is an AI talking head and we don't? He is part of the simulation. Bwahaha!

Scott Manley had his take on space data centers. As well ass Kyle Hill. And they are all like "biggest bullshit". Best one was comparing the datacenter to a thermos flask. You don't put your computers into a themos bottle, because you want to cool it. Thermos uses a vacuum, like space, so it does keep the inside hot. Servers DO NOT want to be hot.

You should feed all the question answer pairs into an AI, make CainBot. 🤣#irony

congrats on 400 ❔shows !

Finally , it looks like a Pump and Dump exercise designed to give the IPO a big markup. And Musk is not getting even a simple mission to the Moon. Then fabrication in an extreme environment when very high precision is required using what kind of not interruptable power.... Lift cost per kilogram are not feasible to cover

Are you going to make those chips in space? The machine that makes base equipment is HUGH. Clearly that is too big to send and it's very hard to make the chips on earth. On earth everything can be serviced but space is much more difficult.

SpaceX has been illustrating how costly it is to lift things in to high orbits. This is not the Falcons to LEO. Starship is years behind with a big portion of the cost paid for.

No bonus question this week? :-( There are conservatives with whom I normally agree who look at history and see how, whenever new technologies are introduced that allow more work to be done by fewer workers, people fear that increasing percentages of the population will be permanently unemployed. However, more new jobs are ultimately created than were lost because of the new technologies. What those technologies have actually done is increase our standard of living. However, the rise of AI seems fundamentally different. How many of the 3.5 million professional driving jobs that currently exist in the U.S. will exist in 10 years? In 20? AI is already writing legal briefs for lawyers. How long before radiologists are replaced by AI programs that do a better job of reading X-rays and other medical scans? Will our new economy really create more new jobs than are lost because of AI? I have serious doubts that will be the case. When we live in a world where most work is done by AI/robots, the companies that own those robots will be heavily taxed so that some form of basic living stipend can be given to people to allow them to buy things the robots make for us. Perhaps that's what living in a post-scarcity world is like. It's possible that people could have a higher standard of living receiving a BLS than the middle class enjoys today. What I worry about is what those people who do not need to work to live will do with their time? When you first learn about socialism, it seems to make so much sense. “From each according to his ability, to each according to his need.” What's wrong with that? It turns out quite a lot. Socialism doesn't take human nature into account. If you tell someone they won't be rewarded for working harder, most people will put in the minimum required effort. In First Contact, Patrick Stewart tells Alfre Woodward, "The economics of the future are somewhat different. You see, money doesn't exist in the 24th century… The acquisition of wealth is no longer the driving force in our lives. We work to better ourselves and the rest of humanity." We are still several centuries from the 24th Century, but I doubt human nature will have changed that much. I really worry about what not having to work will do to peoples self image. Yes, many people will choose to educate themselves and/or become an artist. Many people will strive to help their fellow men and women. There have always been self-selected people who don't need external motivation to be productive and/or help others. However, those people are unlikely to constitute a majority of the population. What will a society be like when a majority of the population never needs to work? I really don't want to find out.

TL/DR: Data centers in space are worth pursuing. At first glance, this seems insane, but… I’ve worked in the energy sector for decades, and about one of those ago, I did a quick look at the viability of solar power in geostationary orbit. No surprise, it made no sense – I couldn’t even get the cost of the satellites alone to the point where over a 25-year life, they could sell electricity cheap enough to pay for themselves, even at a zero percent rate of return. And this is the key, that’s zero launch costs (GEO is almost escape velocity, so it’s a lot more than synchronous orbit, regardless of whether it’s rocket straight to GEO or GTO with the satellite wasting mass to provide the circularization), zero investment in microwave beam technology and ground stations, zero cost of those ground stations and connecting transmission lines, zero lobbying cost to convince local governments to let you beam down microwaves to those ground stations. But with a data center collocated on the satellite, you will have zero costs for all those power transmission related costs and headaches. You have a cheaper orbit with much cheaper rockets. And satellites (that no longer have to be hardened to spend their lives above the Van Allen Belts) are no longer semi-bespoke, but are cranking out of a factory every four hours. So from that cost perspective, it makes much more sense that space solar ever will, and allows for much faster deployment of data centers and power stations (it can take years just to be allowed to hook up a new power source to the grid, and you probably need a dedicated substation built for those data centers). So, it likely does pencil out, but… It was easy to compute how much a megawatt-hour of electricity was worth, I don’t know what a chunk of AI logic is worth. But even if it doesn’t pencil out, all that infrastructure (maybe even that lunar mass driver) will be built and can become this bubble’s version of dark fiber. So there’s that. Go space data centers! And go for moon rather than Mars first (sorry Dr Zubrin, but your planned Ares V rocket morphed into the Senate Launch System).

Thanks for a very interesting Q/A. When I hear that trillions of dollars will be spent on AI, I hear trillions of dollars will be borrowed (via Govt bonds, maybe?) to plummet us into even more global debt, making us even more vulnerable to an existential financial crisis that no one can fix. We're almost there now. It wont be lack of jobs, it will be lack of food and water that sets the ICBMs flying. Sorry to be a doom sayer, but the coders of this world remind me of the piano players on the Titanic.

I haven't yet watched this video, but the email announcing it was right next to this article https://www.breakthroughjournal.org/p/data-centers-wont-be-in-space-anytime