Space data centers could solve problem of 165% surge in AI power hunger
SMRTR summary
Former Google head Eric Schmidt acquired a launch startup to build data centers in space, and now researchers from NTU Singapore are detailing exactly how that ambitious vision could work.
The team proposes sending data centers to low Earth orbit, where the extreme cold provides free cooling and solar panels generate round-the-clock power. This matters because AI-driven energy demand could rise by 165 percent by 2030, according to Goldman Sachs.
"Space offers a true sustainable environment for computing," explains study lead Professor Wen Yonggang. "By harnessing the sun's energy and the cold vacuum of space, orbital data centers could transform global computing."
The researchers outline two approaches: orbital edge data centers that process satellite imagery in space, reducing data transmission to Earth by a hundredfold, and orbital cloud data centers using satellite constellations equipped with servers and radiative coolers.
The remarkable claim? All of this is possible using existing launch and satellite technologies, potentially achieving net-zero carbon emissions for computing infrastructure.
SMRTR provides this summary for quick context. The original article belongs to Interesting Engineering.
Read the original article