Elon Musk issued a unique recruitment call yesterday, stating, “Join xAI if the idea of mass drivers on the Moon appeals to you.” This followed a restructuring that led to several former executives leaving the AI lab. The appeal comes after the company’s merger with Musk’s rocket manufacturer, SpaceX, and ahead of the combined entity’s expected IPO.

While one might assume xAI would focus on pursuing AGI, applying deep learning to challenge established software firms, or even indulging in playful branding like “Macrohard,” Musk is setting his sights on lunar ambitions. After detailing plans to construct AI data centers in orbit—a key synergy between the two companies—Musk expanded the vision further. He posed the question, “What if you want to go beyond a mere terawatt per year,” and answered, “To do that, you have to go to the moon…I really want to see a mass driver on the moon that is shooting AI satellites into deep space.”

According to Musk, the next step beyond Earth-orbiting data centers is deploying even more powerful computers into deep space. He proposes that the optimal method is to establish a lunar city to manufacture these space computers and launch them across the solar system using a large maglev system. For those familiar with Musk’s style, this grand vision fits a pattern evident in a publicly shared all-hands meeting video from xAI.

The presentation slide introducing the Moon base appears toward the end, a segment where Musk has historically shared aspirational concepts, such as SpaceX rockets landing on Mars and visions of a multi-planetary society. Notably, this lunar focus emerges just as SpaceX has publicly stepped back from its long-standing goal of colonizing Mars. Now, with xAI integrated, Musk is advancing a new narrative for the future, framed around the Kardashev Scale.

This theoretical framework, developed by Soviet astronomer Nikolai Kardashev in the 1960s, classifies civilizations based on their energy consumption. Musk described advancing along this scale, beginning with harnessing planetary resources and eventually capturing solar energy in space. With a Moon base, he suggested the company could utilize “maybe even a few percent of the sun’s energy” to train and operate AI models. Reflecting on the potential, he told staff, “It’s difficult to imagine what an intelligence of that scale would think about, but it’s going to be incredibly exciting to see it happen.”

Over the past nine years, Musk’s vision for Mars exploration and colonization has served as a powerful recruitment and unifying tool for SpaceX. The founding narrative around the Red Planet provided a long-term direction that aligned the company’s projects and distinguished its ambitions from more conventional government contractors. “Occupy Mars” merchandise became a symbol of these aspirations.

The proposed Moon base continues this tradition of embedding companies within a compelling story—shifting from a million-person Mars colony to a future centered on AI. Signs of this evolving focus were visible in Musk’s May 2025 Starship update, which concluded with a since-cancelled concept of Tesla Optimus robots traversing Mars.

However, a significant challenge with SpaceX’s Mars plans has been funding. A 2016 proposal to adapt the Dragon spacecraft for Mars landing was abandoned the following year due to prohibitive costs and technical hurdles. Similarly, the Starship vehicle, originally designed for Mars colonization, has been refocused on more immediate revenue sources: deploying satellites for the Starlink network and fulfilling NASA’s $4 billion contract for lunar astronaut landings.

In contrast to multi-planetary colonization, there is a practical logic behind SpaceX acquiring a resource-intensive AI and social media platform to build orbital data centers, especially if predictions of escalating demand and terrestrial costs prove accurate. Experts indicate such orbital infrastructure could become feasible in the 2030s.

Producing satellites on the Moon, however, depends on several other ambitious advancements materializing first. While researchers and startups are testing the fabrication of chips and precision components in space, mass-producing advanced computers on the lunar surface would require dramatically reduced space access costs. This includes transporting all necessary raw materials to the Moon, along with everything needed for a “self-sustaining city.”

Ultimately, this vision serves as an ambitious stretch goal. If retail investors embrace the narrative, it could potentially elevate SpaceX’s market valuation similarly to Tesla. For the engineers—whether in AI or aerospace—whom Musk relies on to realize these objectives, the shift in focus may be disorienting.

Yet, this grand narrative offers a distinctive explanation of xAI’s purpose beyond developing a large language model known for unconventional behavior. As one departing executive remarked on exit, “all AI labs are building the exact same thing, and it’s boring.” Mass-producing a solar system-scale supercomputer on the Moon may be described in many ways—some might call it “insane”—but it is certainly not the same as everyone else, and it is far from boring.