Musk needs a new vision for SpaceX and xAI. They landed on Moonbase Alpha.

“Join xAI if the idea of ​​more drivers on the Moon appeals to you,” CEO Elon Musk he announced yesterday following a reshuffle that saw a host of former officials exit the AI ​​lab.

This is an interesting recruitment strategy after the merger with Musk’s rocket company, SpaceX, and the company’s expected IPO. You might think that xAI workers should be interested in achieving AGI, using deep learning techniques to disrupt traditional software companies, or simply calling out bad words like “Macrohard.” But instead, Elon is going to the moon.

After explaining the objectives of generate AI data in orbitA major partnership between the two companies, Musk took the idea forward. “What if you want more than terawatts per year?” Musk asked. “To do that, you have to go to the moon… I’d like to see a big man on the moon shooting AI satellites into space.”

In Musk’s words, the next step in data orbiting the Earth is supercomputers in deep space. Also, Musk says that the best way to achieve this is to build a city on the moon to build space computers and launch them into the solar system using a large power train.

If all of this sounds a little familiar, Musk’s former executives know that there is a hint of where the conversation appears in the video of all hands meeting xAI shared with the public. A slide describing the moon’s base comes at the end of the presentation, where, during SpaceX pep talks, Musk often shares images of SpaceX’s rockets landing on Mars and waxes eloquent about the future of humanity.

Notably, the moon base comes shortly before SpaceX abandons its long-held goal of colonizing Mars. Now, with xAI in the company, Musk needs a new fictional metaphor for the future: For this reason, Kardashev Scalea hypothetical level of galactic development created by a famous Soviet astronomer in the 1960s. The idea is to increase the size of energy consumption – the first civilizations find out how to use all the energy on their planets, and then (hypothetically) they go to space and build infrastructure to capture solar energy.

With a moon base, Musk says the company could use “maybe even a little bit of solar power” to train and operate AI models. “It’s hard to imagine what this kind of intelligence will come up with,” he told his staff, “but it will be very interesting to see it happen.”

In the nine years since Musk revealed his plans to explore and control the Martian world, this vision has been a tool to help hire SpaceX: The story that started Musk’s interest in the Red Planet provided a long-term vision that united the company’s various development projects, and showed the company’s desire among other space contractors who were focused on additional government activities. The “Occupy Mars” t-shirts provided a visible sign of SpaceX’s ambitions.

That’s where the premise of the fictional moon comes in — part of Musk’s history of wrapping his company in a powerful story. There are one million people living on Mars, but now they are planning for a future where AI is the most exciting thing. The rise of the Martian mission was partially reflected in Musk’s Starship update for May 2025while the show ended with a now-deleted vision of Tesla Optimus robots traversing the Red Planet.

There was only one problem with SpaceX and Mars: No one wanted to pay them to go there. Plans announced in 2016 to repurpose the Dragon spacecraft as a Mars probe were shelved the following year as technical difficulties mounted. And since Musk uncovered The vehicle that will be the Starship in 2016, its power, which was originally designed for Mars, has been changed to focus on two other paying tasks: the installation of satellites of the Starlink comms network and the $4 billion contract to reach astronauts on the moon for NASA.

Unlike the development of many planets, it would make sense for SpaceX to buy expensive AI and social networks to build a data center in Earth orbit, especially if the predictions of increased demand and lower cost will be realized. Experts say it could happen in the 2030s.

Hypothetically, building satellites on the moon would require some of Musk’s dreams to come true first. Scientists and inventors are experimenting making chips and other precision components in space. But building tons of supercomputers on the moon means we’re living in an environment where it’s very cheap to get to space, the most important thing for those technologies, to get all the equipment to work on the moon, including everything needed to make it a “self-sustaining city.”

In a way, that’s the point: This is the, uh, stretch goal. If the meme-happy marketers buy into the controversy, they could turn shares of SpaceX into the next Tesla. Engineers, AI or aerospace, that Musk needs to achieve his goals can find the change. But the vision is one way to describe what xAI is about, except for LLM which is probably better known like a pervert. As one of the managers who are leaving the company he said on the way out, “all the AI ​​labs are building the same thing, and it’s boring.”

Building a super solar system on the moon is a lot of stuff (I’ll get emails that don’t use the word “crazy”), but it’s not the same thing, and it’s not boring.