[beloch]

It's more interesting to ask, "Does AI need to follow the current model of evil megacorps building massive data centres that, collectively, guzzle more energy than most nations on Earth?"

Perhaps LLM's (or something better) will develop to be more efficient and quickly become something most people run on local hardware. Perhaps fad-obsessed management types will move onto the next big thing and AI will start being used more judiciously. Perhaps society will set sane regulatory limits that shape the direction AI is going in, from models that take jobs people want to models that, given the right hardware, can do the jobs few want.

Anthropic and OpenAI don't have to succeed for AI to succeed. If they turn out to be a bubble that bursts and torches a lot of investors, it might actually be a fundamentally good thing for everybody else.

[vessenes]

To be clear, nobody WANTS to have to go build all these datacenters. Well, maybe some pure-play hyperscalers do. But there's an immense amount of economic incentive to be able to do this more efficiently, capital and energy. And, what those hyperscalers want will not matter for a second if there isn't demand for the tokens output by those datacenters - they'll go instantly dark and have to seek new forms of valuable compute to offer.

If this current building spree ends in massive solar and other power generation being overbuilt and cutting energy costs, we've had a really good outcome.

[s08148692]

This is why there is so much interest in space based AI compute. It's not just SpaceX - Google, Anthropic and Nvidia have openly expressed interest

If you look at SpaceX plans and ambitions, they hope to deploy massive compute to orbit (multiple Terrawatts, hundreds of thousands of sats). If their ambitions even slightly materialise it would make ground based compute pale in comparison.

Whether or not they succeed in their plans is beside the point - the point is they know that terrestrial electric infra can't sustain the growth they need

[pjc50]

What is the multiplier in cost for a teraflop of compute in space vs on the ground? 100x? 1,000x?

> Whether or not they succeed in their plans is beside the point

No, I think that does matter eventually? Maybe for the IPO value?

[joebe89]

How would cooling work in space based computing? To a layman like me it seems like a significant hurdle to overcome.

[Pingk]

It really doesn't. You're purely relying on radiation fins to carry heat away, which are incredibly inefficient.

> The radiator surface area problem also scales uncomfortably. At 838 watts per square meter, rejecting 1 megawatt of waste heat requires roughly 1,200 square meters of radiator. Deploying that much surface area on a satellite is a structural engineering challenge that gets harder with every order of magnitude. The ISS solar arrays span about 2,500 square meters total.

So even a 2MW data centre in space requires a cooling array rivalling the international space station. Starcloud launched a single H100 in November and they were unable to run it 24/7 due to heat buildup.

Even with novel solutions to make heat transfer to the fins more efficient, like phase-change liquids, the limiting factor is that the vacuum of space is a tremendous insulator.

https://thecoolingreport.com/intel/starcloud-orbital-data-ce...

https://satnews.com/2026/03/17/the-physics-wall-orbiting-dat...

[dgellow]

Can we stop spreading the obvious bullshit that is space compute?

[jorisw]

Care to at least refer to some sources why?

[butlike]

It's literally a deflection mechanism to the fact they want to build data centers all over the land by proposing a fantastical better way that simply won't work.

[dgellow]

That has been covered to death since months, there is no good solution to cooling without an atmosphere. But even going into the technical details is a waste of time, think about how complicated maintaining such a system in space would be compared to having it on earth, the whole idea is a complete grift from beginning to end. There is literally no benefit to do it in space other than as a marketing tool.

Nobody who suggests the idea has ever presented a model that is even remotely close to reality

[brokenmachine]

What a surprise that something this implausible could come from the same guy who proposed trains that run in vacuum tubes.

I can't believe the people who actually listen to this guy can tie their own shoelaces. And somehow he became the richest person on earth. Humans are irredeemable.