AWS Chief Garman: Orbital AI Data Centers Are 10-20 Years Away

AWS CEO Matt Garman says data centers in space are "highly unlikely to be economically feasible or practical in the next decade or two," directly pushing back on a vision championed by Elon Musk and Jeff Bezos.

Speaking on the Sources Podcast released Tuesday, Garman laid out the math that makes orbital facilities impractical today. The core problem is launch capacity.

"We as a world don't do one launch a day today," Garman said. By contrast, "We're landing in a data center today one GPU every 14 seconds."

That comparison frames the gap bluntly for anyone planning AI infrastructure. The world's rocket fleets can't move hardware at anything close to the pace terrestrial data centers already absorb it.

What did Garman say actually blocks orbital data centers?

Garman, who leads Amazon Web Services, listed three hard constraints beyond launch frequency:

• Rocket launches remain "relatively infrequent and expensive."
• Heat management is a serious engineering problem. "Dissipating heat in space is actually quite hard," he told podcast host Alex Heath. "It's just a really hard environment. There's radiation everywhere, so you have to build things."
• Building servers in space would cost "way more" than on the ground.

The pitch sounds attractive on paper, Garman conceded — orbital facilities could tap free energy from the sun. But he doesn't buy the economics yet.

Still, he stopped short of calling it impossible. "These are all technically solvable problems," Garman said. "It's just going to take a while."

Where do Musk and Bezos stand?

Garman's skepticism puts him at odds with two of the most powerful figures in tech — including his boss's rival.

Elon Musk, CEO of SpaceX and Tesla, has promoted space-based data centers as an answer to AI's soaring energy demands. Earlier this year, SpaceX said it aims to build a "constellation of a million satellites that operate as orbital data centers" and had begun hiring engineers toward that goal. In January, Musk said that within three years "at the latest" space could be the cheapest place to put data centers.

Amazon and Blue Origin founder Jeff Bezos is more measured but still bullish. In a CNBC interview in May, Bezos called space data centers "very realistic," while cautioning on timelines.

"Some of the timelines we hear are very short," Bezos said. "People would talk about two or three years. That's probably a little ambitious."

So the dispute isn't really whether orbital data centers happen — it's whether the horizon is three years, ten, or twenty.

What about oceans and deserts?

Garman extended his skepticism to underwater facilities. "Oceans, there's some of those same problems," he said. "I find oceans to be also probably impractical."

Deserts won his vote among the three frontier options. "I think deserts of those three that you mentioned are by far the most likely," Garman told Heath. "You have relatively good environmental areas. You actually have heat sinks. You have solar areas that you could do it."

He added one caveat that matters to local officials: building data centers far from communities means towns lose out on the increased tax base and high-paying jobs these facilities bring.

How much water do AWS data centers actually use?

Garman also addressed growing concern over data centers' water consumption. Amazon's largest facilities, he said, use less water than 40 US households — roughly what a small residential street consumes.

For cloud customers and infrastructure planners, the practical takeaway is simple: budget for land-based capacity, likely in sunny, dry regions, for the foreseeable future. The orbital option stays on the whiteboard for another decade at least — by the estimate of the person running the world's largest cloud business.