Musk wants to launch AI data centers into space. The public hates them on Earth so naturally the solution is to strap them to rockets and fire them at the sky. This is what happens when a billionaire gets bored with terrestrial real estate.
The economic case is dog sh*t. You know what costs less than launching servers into orbit? Literally everything. A data center in Iowa requires electricity and cooling. A data center in space requires electricity, cooling, launch costs of $10,000 per kilogram, radiation shielding, orbital debris mitigation, and a prayer that nothing explodes. But sure, the NIMBYs in Palo Alto don't want to hear the hum of GPUs so let's spend forty billion dollars on the most expensive CloudFlare alternative ever conceived.
SpaceX is betting big on this. That's the phrase they use when a company commits resources to an idea that sounds like it came from a Reddit thread at 3am. Betting big. Like it's a calculated risk and not just ego with a launch manifest.
The public doesn't want data centers on Earth because they use too much power and water. Moving them to space solves neither problem. You still need to generate the power. You still need to cool the chips. Now you just do it in an environment where replacement parts cost more than a house and arrive via rocket. Efficiency gains through complexity. Brilliant.
Retail traders will see SpaceX and AI in the same sentence and immediately start calculating their Robinhood position. They'll ignore the part where this makes no financial sense. They'll skip the section about launch economics. They'll read orbital data centers and think this is the future. Then they'll wonder why their portfolio looks like it got hit with space debris.
The only thing getting launched here is money into a vacuum.
Photo by Michael Mwangi on Unsplash
Leave a Comment