It sounds as though their AI satellites are contingent on the Starlink V3 design, which in turn rely on Starship to launch them, as the sat bus is too large for Falcon 9.

Relevant Elon tweet: https://x.com/elonmusk/status/1998872465087541752

Starlink V3 will be 20kW and launched at scale around Q4 next year. No problem to scale that to >100kW if the satellite mass is shifted towards solar arrays and radiators for AI compute, instead of giant phased array antennas for Internet connectivity.

As you mention, it would use the same laser comms system as Starlink to connect to Starlink.

An AI satellite is easier, not harder, than the Starlink V3 design, which is a marvel of engineering created by an epic team of humans. I am so proud of the @SpaceX team.

If they think Starship is a long way from actual service, pursuing AI satellites is a strange move, as it seems these satellites would be dependent on Starship to launch them.

[–] Peppycito@sh.itjust.works 2 points 2 weeks ago

pursuing AI satellites is a strange move,

My theory is that it's total bullshit. Why put Ai in space? It solves the power problem with solar, but it makes the heat problem way worse. So why, exactly are Ai space satellites a good idea? What need does it solve, aside from using their infrastructure for something 'hip' and 'cool'? I think Elon is talking out his ass and setting bullshit timeliness to complete buzzword investment bait. 'Space ai' is solar roads.

Of course, I could be absolutely wrong and I'll have to wait till Q4 to see if we have space factories and data centers. Going to be an action packed year with all the starships making orbit and doing orbital refueling and going around the moon and all the other bullshit he's promised.