I was initially very skeptical about the viability of space-based data centers but after a couple hours reading papers, studies and summary technical assessments I realized there are a range of credible expert viewpoints from, "pretty unlikely" to "it could actually work". There at least appear to be plausible, though unproven, solutions to the most obvious drive-by objections I had off the top of my non-expert head.

Of course, there are still a lot of unknowns, any of which could prove fatal to the concept but I'm no longer comfortable just dismissing it as "obviously ridiculous."

Tesla has been out competed in Batteries, EV's and Robots so this is his new move. He did something similar with his solar panel company put it inside Tesla and then it has almost disappeared from the news. He puts the AI company inside of Spacex makes up a lot of unrealistic numbers to pump up price and captures most of the stock gains from Spacex IPO by diluting others people shares.

Don’t forget solar city. There is precedent for this, it is how Musk operates and it’s far more about protecting his investments and letting him use the company for his own enrichment than what makes sense for the company.

You forget that Musk has to make all the idiots who gave him capital for Twitter whole, somehow.