[whatisthiseven]

The first premise is reasonable. The second premise is so outlandish as to require dozens of other assumptions, optimistic outcomes for launch, and pessimistic views on the cost of earthbound systems.

I care about the environment and I think we can keep earthbound systems, and also reduce their impact. Making assumptions about the feasibility of launch and the economic absurdity of orbital compute, but not affording the same assumptions for what could be done for earthbound systems, is confusing?

And no, orbital compute is absolutely not far lower than earthbound in co2 cost. Because it doesn't exist at any scale. All orbital compute is solely dedicated to switching where it is best served. If you were to spitball numbers, are we even willing to assume orbital matches earthbound in compute total, dollar cost, uptime, or any beneficial metric?

The only metric I see is just slinging silicon into space.

[GMoromisato]

I get skepticism that orbital datacenters will make sense economically. It would not surprise me if they never happen.

But if they made economic sense, then they would be much more environmentally friendly. On CO2 emissions, we're talking about a factor of 37. That's not a small amount. Even a factor of 10 would be a major issue multiplied by terawatts of power. Sure, maybe Earth-bound energy gets cleaner, but it's not going to get 10 times cleaner before 2036.

[jurgenburgen]

> Even a factor of 10 would be a major issue multiplied by terawatts of power. Sure, maybe Earth-bound energy gets cleaner, but it's not going to get 10 times cleaner before 2036.

Currently we’re not on that trajectory but if we don’t get there we might be too busy with civilizational collapse to need space data centers.

It would require doubling the rate of 2025 renewable deployments on average for the next 11 years so technically and economically it’s feasible.

[beAbU]

What aspect of an orbital data centre makes it more environmentally friendly than a terrestrial one?