Comment by Robotbeat
15 hours ago
That already happened with the merger of X with XAI and then the merger of Xai and SpaceX.
But the reason is because SpaceX is trying to tool up for orbital datacenters. They're building a bunch of solar cell manufacturing plants and Starship launch pads.
Or at least they are selling the idea of orbital data centers since the technology for orbital data centers does not exist yet or in the short term.
It’s literally just a satellite with GPUs on it. SpaceX already has about 100MW of satellites in orbit. They’re making an upgraded version of these satellites, V3, designed for launch on Starship. Making a bigger versions of those with larger radiators (because of the duty cycle) and putting them in sun-synch is all that is required, if it can be done cheaply enough. And Starship, if it works, should be cheap enough as its propellant cost would be less than the fuel cost of using airliners to do transpacific air freight.
It didn't already happen. As you pointed out, people who funded the purchase of Twitter hold SpaceX shares, and this IPO is how they get their money back.
It did. After the merger with SpaceX, the Xai debt got renegotiated at a very low rate.
Orbital datacenters is just an excuse of a reason to merge the two companies when there’s nothing else tying them together. They won’t actually happen.
SpaceX will run out of things to profitably launch with Starship once they launch Starlink V3 (a few hundred launches per year), and orbital datacenters provides effectively unlimited launch demand. Like Starlink was for Falcon 9, orbital datacenters are a way to fully leverage the insane launch cost advantage and launch capacity that Starship brings to the table.
I see your point. All this launch capacity is coming online but, once Starlink is built out, who is going to buy it? Surely Starlink maintenance isn't big enough [edit: maybe it is, it's hard to get my head around the number of sats]
The datacenters in space thing just doesn't make sense to me. Datacenters get their value from scale which is why they're so freaking massive here on earth. I just don't see any datacenter in space being big enough (unless built over multiple lifetimes) to use let alone profitable/desireable. What am i missing?