Datacenters in Space: Not a Sitcom, but Maybe It Should Be

Don’t tell anybody, but I’m into my seventh decade on this planet. I strive for equanimity — it reduces stress and it's good for my health — but sometimes I can’t quite get there. Oh, I try. I give it everything I’ve got to be equable and benignly detached.

My problem, if I can call it that, is that I’m an inveterate and voracious reader. Now you might think that shouldn’t be a problem, that reading induces contemplation and an introspective tranquillity. To which I respond: Well, it all depends what you’re reading. If you’re reading the news, the intake of which I should probably reduce — as with salt and sugar — then maintaining one’s composure is a grind. You have to work at it, and success isn't assured.

It’s not the anger that gets to me. No, it’s the mind-bending disbelief. I simply cannot believe half of what I’m reading. How, I ask myself, can satirists earn a living when reality has become more absurd than anything a surrealist or dadaist could possibly depict? You read ostensibly fact-based news reports and you say to yourself: “This can’t really be happening, can it?”

Maybe, I wonder, as I spin helplessly into a Hitchcockian vertigo tunnel, we really are living in a simulation. But whose simulation? And why am I relegated to the role of a bit player, practically an extra, in my own life?

I’m being facetious. I don't really believe that's true. But if we're living in a simulation, what I think is neither here nor there.

Bouncing Off the Ethical Ceiling

The fact is, I don’t believe we’re living in a simulation. Instead, I posit that we’re living in reality’s discount basement, a gimcrack knockoff of a better-appointed establishment. We took a bad step or a wrong turn on our collective journey of scientific progress and spiritual enlightenment. Admittedly, I concede, we have always achieved more of the former than the latter, but our contemporary lives, increasingly lived online, are increasingly demented.

Perhaps we’ve hit humanity’s ethical ceiling and we’re now bouncing downward in thrall to crass gravity. I grew up valuing the utility of logic and reason, and suddenly it’s revealed to me, in stark horror, that none of that matters. Very few things today make sense, and few of us try to reconcile all the diverging fragments and shards of our fractured world into something that imparts coherence and meaning.

Maybe humor, wry and knowing, is the only answer, or at least the only way to stay relatively sane. I’m giving that a try, too, progress report to follow.

But just look at the news. Every few days, I encounter an article that suggests AI will engender a vast leisure economy, in which displaced workers will have ample time to pursue hobbies, play games, and indulge in sybaritic decadence. The articles propose that this languid turn toward diversion and indulgence is a tremendous business opportunity for purveyors of gaming, entertainment, esoteric hobbies, and globe-girding travel.

In slipshod theory, I suppose that’s possible; in practice, though, how would it work? And work is the key word here because our economies are based on the premise that employees and workers exchange labor and productive time for market-based compensation from their employers. We’re still living in accordance with those economic precepts, and if employees and workers find themselves working and earning less, or not at all, how will they be able to afford copious not-free leisure in their free (but not-so-free) time?

There’s been some specious talk about a universal income. Talk is cheap. I don’t see the billionaire class reaching into their bottomless pockets to pay more in taxes, and that’s probably the only way any western country will have the wherewithal to sustainably support universal incomes.

When Free Time Isn’t Free

Fortunately, we’ll likely avoid the fate of leisure-based economies in which the only amusement and recreation that superannuated workers can afford will involve inexpensive hobbies, like chess and billiards. Then again, you’ll only be able to pursue those hobbies if you already have a chess set and access to a billiard table and pool cues.

Alas, I don’t think AI is going to claim nearly as many jobs as its most rabid proponents forecast. Yet all we see and hear are predictions about AI rendering vast swathes of the workforce superfluous.

I’m simply not buying what the AI propagandists are selling. AI is useful for closed-loop processes and self-contained workflows that involve minimal, if any, interpersonal communication and interaction. But think about it: Most jobs today in our post-industrial, information-based economies involve interpersonal communication and relatively nuanced verbal and non-verbal exchanges of information and meaning. AI should be able to enhance many of those roles, making interpersonal processes and workflows more efficient and productive, but the people will remain in place. Humanoid robots are not on the cusp of putting these people out to pasture, despite the beer-league soccer games Chinese robots seem to be enjoying.

Spacing Out on Datacenters

Meanwhile, Elon Musk continues to bloviate about establishing human settlements on Mars. Aside from the considerable near-term impracticalities of such an endeavor, we struggle, along with Musk, to identify a compelling business case for the venture. It’s worth noting, if only in passing and to shake our heads in incomprehension, that Musk’s principal motivation for his Mars quest is to ensure humanity’s survival. He is apparently deeply concerned about the threat of extinction from nuclear war, pandemics, or natural disasters. Naturally, his proposed solution is not to mitigate the risk or severity of such cataclysmic disasters, but to attempt a mass interstellar relocation to a planet that is inhospitable to human life.

Body-building enthusiast and tech billionaire Jeff Bezos, of Amazon renown, doesn’t seem quite as keen as Musk on the Mars idea, but he is enamored of the prospect of putting datacenters in space. He’s not alone, either. OpenAI’s Sam Altman and former Google CEO Eric Schmidt are also advocates of space-based AI datacenters. To be fair, Bezos thinks these beyond-cloud computing facilities are still 10 to 20 years away from being realized. He might be optimistic with that estimation, which brings to mind a comment a basketball scout once made about a rough-hewn draft prospect: “He’s two years away from being two years away.”

In an article published today by Reuters, Bezos opined:

"One of the things that's going to happen in the next – it's hard to know exactly when, it's 10 plus years, and I bet it's not more than 20 years – we're going to start building these giant gigawatt data centres in space," Bezos said during a fireside chat with Ferrari (RACE.MI) and Stellantis (STLAM.MI) Chairman John Elkann at the Italian Tech Week in Turin.
The concept of space-based data centres is gaining traction among large tech companies, as the energy needs to maintain such operations on Earth are growing sharply.
"These giant training clusters, those will be better built in space, because we have solar power there, 24/7. There are no clouds and no rain, no weather," Bezos said. "We will be able to beat the cost of terrestrial data centres in space in the next couple of decades."
Bezos said the shift to orbital infrastructure is part of a broader trend of using space to improve life on Earth.
"It already has happened with weather satellites. It has already happened with communication satellites. The next step is going to be data centres and then other kinds of manufacturing," he said.

Bezos makes it sound inevitable, doesn’t he? It’s all of a piece with a new twist on the space-time continuum. Then again, the spoilsports at Reuters can’t help concluding the article by mentioning that a few not-so-minor details need to be resolved before the tech industry can graduate to space invaders. The nettlesome challenges include “cumbersome maintenance, limited scope for upgrades, and high costs of launching rockets as well as the risk of failed rocket launches.” I’m sure we could add to that list if we had the time and the inclination, but other debunking takes precedence.

That some of the richest men in techdom are reaching into space to find abundant real estate for their datacenters tells us more about the limits to AI growth than it does about the vaulting ingenuity of tech’s intelligentsia.

AI and the New Nimbyism

Why do you think they’re casting a covetous eye toward space? That answer is obvious, as a surfeit of articles from newspapers (yes, a few of those relics can still be found in the wild) attest. AI datacenters are being built nearly everywhere that tech giants can find wide-open spaces and adequate energy supplies. The preceding sentence, however, is easier to type than it is to enact.

When agrarian and rustic inhabitants learn that their new neighbor will be a loudly humming, energy-guzzling, water-depleting, city-size datacenter, many of them are not in the mood for a celebratory block party. You can appreciate their unease. Put yourself in their clogs. AI datacenters have spawned a digital nimbyism. Yes, the tech giants and others want to build these gargantuan datacenters to serve an allegedly booming market that remains mostly aspirational, but do you really want to live next door to one?

There’s also that problem that the earth is only so big, and that if AI does attain the mind-boggling success forecast by its Pollyannas, we will eventually run out of open spaces in which to put the requisite datacenters. We’re not there yet — datacenters have not sprouted everywhere — and I don’t think we’ll get there in my lifetime (and probably not in yours, either). Nonetheless, you can see how the AI market optimists perceive a land-supply problem.

That’s to say nothing of the strain that such datacenters would place on energy and water supplies, and the ensuing harm they might inflict on the environment. (I realize that “woke” environmentalism is unfashionable in the Trumpian epoch, but we can only go on an environmentally destructive bender for so long before we do serious damage to ourselves.)

Where the Space has No Name

In an article published by Grist a few weeks ago, the datacenter-capacity challenge was framed as follows:

The companies frantically building and leasing data centers are well aware that they’re straining grids, driving emissions, and guzzling water. The electricity demand of AI data centers in particular could increase as much as 165 percent by 2030. Over half of the energy powering these sprawling facilities comes from fossil fuels, threatening to reverse progress toward addressing the climate crisis.
Some of the biggest names in artificial intelligence say they have a solution: Just stick these colossal computer clusters in space. OpenAI CEO Sam Altman told manosphere podcaster Theo Vonn that he considers a massive expansion of data centers inevitable. “I do guess a lot of the world gets covered in data centers over time,” he said. (This is not, in fact, inevitable, but the result of unfathomably wealthy companies choosing to invest unfathomably large sums of money. Altman has speculated that he would quite literally put trillions into it, and OpenAI is part of the consortium behind the $500 billion Stargate project.)
Altman is aware, however, that some people might not like this. “I’ve spoken with environmentalists,” he said. Then, he offered a suggestion. “Maybe we put [ data centers] in space,” he said. “I wish I had, like, more concrete answers for you, but like, we’re stumbling through this.”

There’s a lot of public stumbling on display these days, Sam, so join the blundering chorus line. Perhaps a choreographer can give it a better look.

Space, as was declaimed in the intros to old Star Trek episodes, is the last frontier. It’s also a libertarian’s wet dream, an empty space without regulations or pushback from pitchfork-wielding citizenry.

The Grist article frames the frontier opportunity in its concluding paragraphs:

For now, it’s much more expensive to put a data center in space than it is to put one in, say, Virginia’s Data Center Valley, where power demand could double in the next decade if left unregulated. And as long as staying on Earth remains cheaper, profit-motivated companies will favor terrestrial data-center expansion.
Still, there is one factor that might encourage OpenAI and others to look toward the heavens: There isn’t much regulation up there. Building data centers on Earth requires obtaining municipal permits, and companies can be stymied by local governments whose residents worry that data center development might siphon their water, raise their electricity bills, or overheat their planet. In space, there aren’t any neighbors to complain, said Michelle Hanlon, a political scientist and lawyer who leads the Center for Air and Space law at the University of Mississippi. “If you are a U.S. company seeking to put data centers in space, then the sooner the better, before Congress is like, ‘Oh, we need to regulate that.’”

It’s absurd, even crazy, when you think about. How did we get here? Some of us are willing to pave over the earth and clutter space, at great expense and even greater cost, for a market that might never reach the exalted heights projected for it. So much of it doesn’t make sense, but sensible objections fall on deaf ears amid the frenzied cacophony of a madhouse.