Headlines

Sundar Pichai and Elon Musk, Amazon’s computing head Matt Garman does not agree with the data centre idea you call ‘future’; he sees it as impractical as …


Sundar Pichai and Elon Musk, Amazon's computing head Matt Garman does not agree with the data centre idea you call 'future'; he sees it as impractical as …
Sundar Pichai and Elon Musk, Amazon’s computing head Matt Garman does not agree with the data centre idea you call ‘future’; he sees it as impractical as …

​Sundar Pichai and Elon Musk are pursuing space data centres. Yet, Amazon Web Services CEO Matt Garman sees the idea as impractical for another decade or two. He has questioned the economics of putting computing infrastructure into orbit, even as Pichai’s Google and Musk’s SpaceX pursue plans to run artificial intelligence hardware beyond Earth.​“I think that it is highly unlikely to be economically feasible or practical in the next decade or two,” Garman said during the Sources podcast with Alex Heath. His concerns include expensive launches, the weight of computing equipment, cooling and the difficulty of maintaining servers in space. The remarks highlight a disagreement over how quickly access to solar energy in orbit could translate into commercially viable infrastructure for AI.​

Why Matt Garman is questioning space data centres

​Garman pointed to launch costs and frequency as obstacles to moving computing infrastructure into orbit. “Number one, launches are still relatively infrequent and expensive,” he said.​He compared that pace with AWS’s deployment of computing equipment: “We’re landing in a data center today, one GPU every 14 seconds.” GPUs, or graphics processing units, are chips used to run AI workloads.​The weight of the equipment and the scale of the power supply would create further challenges. “People talk about gigawatt data centers and megawatts of racks. Most solar panels and things [similar systems deliver] kilowatts, not gigawatts. So you need these gigantic things that are up there.”​Garman also questioned the assumption that the cold environment would make cooling straightforward. “Space also doesn’t have air. Dissipating heat in space is actually quite hard.”​Radiation exposure and the difficulty of servicing equipment would add to costs. However, he acknowledged that these obstacles could eventually be overcome: “I’m not saying any one of these things that I could highlight is impossible.”​

What Sundar Pichai’s Google plans for space computing

​Google’s Project Suncatcher is exploring networks of solar-powered satellites carrying Tensor Processing Units, the company’s specialised AI chips. Announced in November 2025, the programme initially outlined plans with satellite company Planet to launch two prototype satellites by early 2027. Google’s announcement​In a September 2026 update, Google described an initial hardware test and a separate two-satellite experiment planned for 2027. Future satellite clusters would exchange data through laser links, allowing their processors to work together on larger AI workloads.​Google also identifies cooling and radiation as challenges. Its engineers are testing heat pipes and radiators to remove heat, alongside experiments assessing how its chips withstand radiation. The programme remains a research effort into whether computing in orbit can scale. Google’s project update​

What Elon Musk’s SpaceX is proposing

​SpaceX is proposing a larger satellite network. In January 2026, the company applied to the US Federal Communications Commission for permission to launch and operate up to one million satellites as an orbital data centre system. The application sets out a proposed network. FCC public notice​Garman favoured deserts when asked about alternative locations for data centres. “I think deserts of those three that you mentioned are by far the most likely,” he said, citing access to solar energy and cooling options.



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *