Today was the first day I heard about space data centres. At this point I thought it wasn't possible but I was startled, shook even, by this claim. Surely there are some things that are too ridiculous claim even by AI boosters. But I keep being proven wrong.

I looked it up and there are many "research papers" about space data centres. But as far as I can tell there is only one startup trying to bring this pipe dream to reality. It's called Starcloud. They recently had their first successful test launch.

About the size of a small refrigerator, the 60kg satellite features the first Nvidia H100 GPU in space.

They have another test launch planned "by 2026" which they label "StarCloud-2".

The supposed end goal is this (according to Nvidia):

Starcloud plans to build a 5-gigawatt orbital data center with super-large solar and cooling panels approximately 4 kilometers in width and length.

I can't even begin to describe how ridiculous this is,

The linked post is extremely valuable because it is written by someone who has worked on space electronics themselves. But I hope the normal reaction when someone boosts "space data centres" for AI we don't really need is to laugh at them.