The hidden mission behind NASA’s Artemis expedition to the moon…and why Elon Musk is so suddenly invested in it

When NASA’s Artemis II lifts off from the launchpad at Cape Canaveral, Florida on Wednesday evening, it will be sold as a symbolic return to the space race for the United States.

Keen observers understand it to be something far more concrete.

America is sending astronauts back toward the moon again, in the first crewed test flight since Apollo 17’s landed on the lunar surface in December 1972.

NASA has said the Artemis campaign is about laying the groundwork for future Mars missions, harnessing space’s economic benefits and creating an enduring human presence beyond Earth.

It’s also—as is anything these days in the United States—about heading off China’s own lunar land grab: Washington wants to remain in front of Beijing’s own plans to put a man on the moon by  2030.

But Artemis appears to be the opening step in a much more consequential infrastructure race that also pits Washington against Beijing—a giant leap toward an extraterrestrial AI future. And the program has attracted an unlikely new advocate.

Elon Musk’s ambitions in outer space have shifted significantly in recent months.

Last year, he was maligning the Artemis mission as ‘extremely inefficient’ and ‘a jobs-maximizing program, not a results-maximizing’ one.

What a difference a year makes. Despite dismissing lunar efforts as a distraction from the human colony that he envisions building on Mars, Musk now calls Artemis a vital step in the development of spaceflight.

In February he posted to his millions of followers that getting to the moon was actually quite a good idea.

‘SpaceX has already shifted focus to building a self-growing city on the Moon, as we can potentially achieve that in less than 10 years, whereas Mars would take 20+ years,’ he wrote.

‘SpaceX will also strive to build a Mars city and begin doing so in about 5 to 7 years, but the overriding priority is securing the future of civilization and the Moon is faster,’ he added.

Around the same time, he made a public push to develop space-based AI data centers, arguing that the energy, land and cooling constraints choking terrestrial data centers – including his own in Memphis, which was accused of illegally operating gas turbines to power its production – make orbit increasingly attractive.

AI data centers are integral to the future of the technology. They’re vast buildings housing the computer chips vital to AI’s operations, and help train new, more proficient models, while also enabling ‘inference’ – the ability for AI systems to answer the questions and tackle the tasks users put to them. But they require a lot of energy and water to work, and are already straining power grids and national budgets.

NASA’s own language hints at that promise: they’ve talked about the systems needed to support sustained operations on and around the moon. 

Make the moon a place to work, service hardware, warehouse supplies and test power systems, and it quickly stops being just a scientific outpost and becomes a strategic foothold.

And if space-based data centers are to become a thing in the near future, then it’s important to plant your flag–literally–on the next frontier of where they’ll be built in order to compete against China, who has the same ability to launch its taikonauts (the Chinese version of astronauts) into space.

Like Musk, the Chinese also possess grand ambitions when it comes to space-based AI programs. Perhaps even more so. In January, China Aerospace Science and Technology Corporation—Beijing’s main space contractor—announced plans to directly challenge Musk and SpaceX and develop its own ‘gigawatt-class space digital-intelligence infrastructure.’ 

Their goal is to construct industrial-scale ‘space clouds’ by 2030 that—as Musk envisions—will utilize space-based solar power to fuel AI-generation, storage and transmission.

Thanks to his successful Falcon 9 reusable rocket program, Musk remains ahead of the Chinese—which have yet to develop a rocket capable of the repeat launches required to reliably and affordably offshore AI infrastructure into orbit. 

But with China making ‘space clouds’ a centerpoint of their latest ‘five-year plan,’ the Artemis launch sends a clear message that the US—and Musk—refuse to concede their leads to a rogue competitor regime like the CCP.

Of course, it’s unlikely either astronauts or taikonauts will be floating by space-borne data centers any time soon. Beyond the challenges of keeping solar radiation and interstellar debris at bay, the maintenance and launch demands make repairing hyperscale computers in orbit harder than swapping a server blade in Scranton.

We’re still, even by the most optimistic estimates, the best part of a decade away—if ever—from having space-based AI data centers operating, despite China’s shorter-term ambitions.

Yet the one thing Musk’s fans and opponents can agree on is that his life story is the history of turning ridiculous-sounding infrastructure claims into realities for everyone else.

Reusable rockets once seemed fanciful. Musk’s SpaceX now throws them up into air and catches them in a complicated chopsticks maneuver with alarming regularity. A satellite internet constellation on the scale of Starlink seemed like science fiction, until you could log onto it while flying in a plane cross-country. So even if space-based AI data centers arrive later than Musk claims, the logic behind them is not hard to grasp.

AI’s all-encompassing appetite for electricity, land and water is exploding. And space provides a seemingly limitless location for power centers with no environmentally-sensitive communities to piss off and a (theoretically) endless supply of power from the sun.

Of course, there are plenty of reasons for returning to the moon. Artemis long predates Musk’s current fixation on the moon. But if you did want to build a future in which vast amounts of reliable AI compute sit off Earth, you would need more than rockets flung up from Cape Canaveral every so often. You would need a reliable logistics chain. You would need fuel, power, maintenance capability, robotic systems, communications, rules for access and some kind of permanent operational base beyond low Earth orbit.

You’d need, in other words, exactly the sort of ecosystem Artemis is slowly, expensively and publicly trying to create. Artemis is meant to help build a lunar economy and support an expanding space economy, says NASA.

Just last week, NASA shelved plans for a revolving space station circulating around the moon, the Lunar Gateway, in favor of a $20 billion permanent base on the lunar surface.

‘It should not really surprise anyone that we are pausing Gateway in its current form and focusing on infrastructure that supports sustained operations on the lunar surface,’ said NASA head honcho Jared Isaacman.

Investors are already putting serious money into off-planet AI infrastructure.

Yes, Artemis is about America returning to the moon. Yes, it’s about heading off China’s increasingly aggressive advances. And yes, it’s about science. But it may also be the first serious public build-out for a much more sci-fi-style future where the race of AI supremacy goes extraterrestrial.