AI Now Apparently Needs Space

Artificial intelligence is growing fast — so fast that some of the world’s biggest tech leaders are starting to wonder whether Earth itself can keep up.

As AI models get larger, smarter, and more power-hungry, the pressure on traditional data centers is intensifying. Power grids are strained. Water for cooling is scarce. And the costs keep climbing.

Elon Musk, unsurprisingly, has a bold solution: take AI off the planet entirely.

Elon’s Latest Big Idea: Servers, But Make Them Orbital

In January 2026, SpaceX filed an application with the U.S. Federal Communications Commission seeking approval to launch up to one million satellites into low-Earth orbit — not for internet access, but to function as “orbital data centers.”

According to the filing, these space-based systems would process AI workloads using solar power, bypassing what SpaceX claims are the growing limits of Earth-based computing infrastructure.

Musk’s argument is simple: AI demand is expanding faster than terrestrial power and cooling systems can handle, and space offers abundant energy and natural cooling.

SpaceX says the network could eventually deliver computing capacity to billions of users globally, framing it as a foundational step toward a future where humanity harnesses energy at a planetary — or even solar — scale.

From Starlink To Star Servers

The proposal would dramatically expand SpaceX’s footprint in orbit.

The company already operates nearly 10,000 Starlink satellites, a number that has drawn criticism from astronomers and space-safety experts.

Musk has repeatedly rejected claims that Starlink is crowding space, saying the orbit is vast and underutilized.

In a post on X, Musk said the satellites would be spaced far enough apart that it would be difficult to even see one from another.

Like Starlink, the proposed AI satellites would operate between 500 and 2,000 kilometers above Earth, powered by solar energy.

Space: Green Dream Or Costly Fantasy?

SpaceX argues that orbital data centers could be a greener alternative to massive land-based facilities, which consume enormous amounts of electricity and water for cooling.

But experts aren’t fully convinced.

Launching heavy computing hardware into orbit remains expensive, and maintaining, cooling, and protecting that hardware in space introduces its own engineering challenges.

There’s also the growing problem of space debris, which increases the risk of collisions that could damage equipment or send debris falling back to Earth.

Astronomers have already complained that Starlink’s radio signals interfere with telescope observations — a concern that would likely grow with an even larger satellite network.

AWS Says: Not So Fast

Not everyone in Big Tech is buying Musk’s vision — at least not yet.

Last week, Amazon Web Services CEO Matt Garman poured cold water on the idea, calling space-based data centers impractical and uneconomical with today’s technology.

Speaking at the Cisco AI Summit in San Francisco, Garman said the biggest bottleneck isn’t computing power — it’s transportation.

Server racks can weigh around 1,000 pounds, and the cost of launching that kind of payload into space remains massive. Garman also noted that humanity has yet to build permanent, large-scale infrastructure in orbit.

“There are not enough rockets to launch a million satellites yet,” he said, adding that space data centers are still “pretty far” from reality.

The Bigger Picture: AI Is Stress-Testing Earth

Despite the skepticism, the interest itself says a lot.

AI’s rapid expansion is pushing cloud providers, utilities, and governments to rethink how computing infrastructure is built and powered.

Traditional data centers are hitting real-world limits, sparking interest in unconventional ideas — from nuclear-powered facilities to underwater servers and now, orbital ones.

Even Alphabet, Google’s parent company, has early-stage plans to explore space-based data center concepts.

For now, though, the industry remains divided: Musk sees space as the inevitable next frontier for AI, while rivals like AWS say the economics simply don’t add up — yet.

AI’s growth problem is real. Whether the solution lies in smarter infrastructure on Earth or ambitious leaps into orbit is still an open question.

But one thing is clear: when AI starts outgrowing the planet, the tech industry’s imagination knows no gravity.

That's all for this week! If you found these updates useful, you'll like more from this newsletter. Get deeper dives, hot takes, and all the latest tech news delivered straight to your inbox.