It’s 2026, and four astronauts are preparing to be the first to return to the moon in more than 50 years. Should they pack the GPUs to get started on a lunar data centre now, or wait for prices to go down?
“The easy part is getting it up there, believe it or not.”
Jim Kalogiros,
Schneider Electric
It sounds like a joke, but the idea of building lunar data centres has been increasingly floated by companies as prominent as Google. Demand for AI has created a parallel demand for compute capacity, and the data centres behind that capacity consume a lot of power and a lot of water to stay cool. They’re also noisy eyesores that take up a lot of space. So, why not build those facilities in space—literally?
Jim Kalogiros, the VP of secure power Canada with Schneider Electric, sees the benefits. After all, the sun has unlimited energy to give and space is an inherently cold environment, he explained in a December interview with BetaKit. It’s a computer’s paradise, and the opportunity has prompted Jeff Bezos’ and Elon Musk’s race to build data centres in space. But Kalogiros isn’t sold on the feasibility of space-based data centres just yet.
“The easy part is getting it up there, believe it or not, which, in itself, is probably challenging,” Kalogiros told BetaKit. “If something goes wrong up in space, how are you going to service it? … getting equipment up there if it fails is not as easy as it is to drive down with a FedEx truck to your warehouse.”
RELATED: Kepler makes history with world-first launch of optical relay satellite network
Schneider is a multinational energy tech company that helps data centre providers implement energy-saving software and hardware. It has made data centres a key part of its business model; during the company’s earnings report last fall, it reported that sales in that market climbed by double digits. The company said at the time that it expected demand for data centres to keep growing this year—but that there will be challenges, like power constraints.
However, given challenges like solar storms (which can knock out electronics) and space debris, Kalogiros thinks it’s worth focusing on solving the energy, water, and cooling crisis on Earth before looking to the stars.
“We can really save a lot of energy just from optimizing what’s currently being built,” Kalogiros said, adding that investments in infrastructure and grid resiliency will help meet that goal.
“I think we’ve got a much easier solution on Earth than we do in space today,” he added.
A lot of this comes down to grid modernization. Kalogiros claims 20 to 30 percent of energy is being wasted by outdated gear that’s using more power than it needs to. He also foresees policy changes that spark investment in solar, wind, and potentially nuclear power before data centres in space are truly figured out.
Still, the race is on. Toronto-based Kepler Communications beat Musk and Bezos to the punch earlier this month when it launched its low-earth orbit satellite network that essentially functions as an orbital data centre.
Even if Kalogiros believes that humanity’s ability to solve its problems on Earth will be faster than solving the ins and outs of space data centres, he thinks innovating for space will help solve Earth’s constraints, prolonging the need to go to space in the first place.
“I suspect whoever figures it out is probably going to benefit drastically,” Kalogiros said.
Feature image courtesy Unsplash. Image by SpaceX.