“Join xAI if the idea of mass drivers on the Moon appeals to you.” That’s how Elon Musk is recruiting employees now, apparently. Not with promises of achieving AGI or disrupting the Technology industry, but with visions of a lunar city manufacturing AI satellites and yeeting them into deep space via a giant maglev train.
This recruitment pitch comes on the heels of xAI’s merger with SpaceX and a wave of executive departures that would make any HR department nervous. But Musk being Musk, he’s doubling down on the science fiction.
From Mars to Moon Factories
The shift is telling. For nine years, Mars colonization was SpaceX’s north star. “Occupy Mars” t-shirts became the uniform of choice for true believers. The Red Planet promised a narrative that united everything SpaceX was building, a story bigger than contracts and quarterly earnings.
But here’s the thing nobody wanted to say out loud: No one was actually paying SpaceX to go to Mars. The company quietly cancelled its Dragon Mars lander plans in 2017 when the technical challenges got too expensive. Starship, originally designed for Martian colonization, got repurposed for the more boring but profitable work of launching Starlink satellites and ferrying NASA astronauts to the moon for $4 billion.
The Martian dream was already dying when Musk’s May 2025 Starship update ended with a now-cancelled vision of Tesla Optimus robots stomping around Mars. Even the robots couldn’t save the pitch.
The Kardashev Scale Pivot
Enter the moon base. During an all-hands meeting that xAI shared publicly, Musk pulled out a new metaphor for humanity’s future: the Kardashev Scale. This 1960s Soviet concept measures civilizations by their energy consumption, from planetary energy mastery to eventually harnessing the power of entire stars.
Musk’s version goes like this: First, you build AI data centers in orbit around Earth. Then you go bigger. Way bigger. You build a self-sustaining city on the moon that mass-produces advanced computers and launches them into deep space using electromagnetic catapults. With this setup, xAI could capture “maybe even a few percent of the sun’s energy” to train AI models of unimaginable scale.
“It’s difficult to imagine what an intelligence of that scale would think about,” Musk told his staff. “But it’s going to be incredibly exciting to see it happen.”
The slide describing all this appeared at the end of his presentation deck, in the exact spot where he used to wax poetic about Mars colonies during SpaceX pep talks. The pattern is familiar for anyone who’s watched Musk operate.
The Business Case for Orbital Computing
Look, there might actually be some Business logic buried in here somewhere. If data center costs on Earth keep climbing and demand keeps exploding, orbital computing could theoretically make sense by the 2030s. Some experts think it’s possible.
But manufacturing satellites on the moon? That requires a completely different universe of assumptions. You need dramatically cheaper space access, which is the whole reason these technologies might work in the first place. You need to get raw materials to the moon. You need that “self-sustaining city” Musk keeps mentioning. You need scientists and startups experimenting with building precision chips in space to actually succeed at scale.
It’s not a roadmap. It’s a stretch goal wrapped in a recruitment poster.
The Narrative Economy
That’s actually the point though. Musk has always understood that the most valuable companies don’t just build products, they sell stories. Tesla wasn’t selling electric cars, it was selling a cleaner future. SpaceX wasn’t launching satellites, it was making humanity multi-planetary.
Now xAI needs its own mythology. Without it, the company is just another AI lab training another large language model that, as one departing executive noted, happens to be “perhaps best known as a pervert.” When all AI labs are building the exact same thing, boring becomes the enemy.
A solar system-scale supercomputer mass-produced on the moon is many things. Insane, probably. Impossible, maybe. But boring? Definitely not. And if that story is compelling enough to attract meme-happy retail investors and ambitious engineers who want to work on something different, then the moon base has already served its purpose.
The real question isn’t whether we’ll see mass drivers on the moon in our lifetime. It’s whether enough people will buy into the vision to make the less spectacular parts, like orbital data centers and cheaper space access, actually happen.
