'Putting the servers in orbit is a stupid idea': Could data centers in space help avoid an AI energy crisis? Experts are torn.
Summary
Google Research proposed 'Project Suncatcher,' exploring space-based AI infrastructure using solar-powered satellite constellations to address the rapidly growing energy consumption of data centers, which is projected to double by 2030.
However, experts are divided. Joe Morgan of Patmos dismisses the idea as impractical for the near term, citing issues with hardware obsolescence and the immense difficulty and cost of orbital repairs and maintenance, as well as unavoidable latency for most AI workloads. Conversely, Paul Kostek of IEEE notes that the concept stems from genuine terrestrial pressures, highlighting the advantages of 24/7 solar power and superior heat rejection into space over water-constrained Earth facilities.
The ultimate rationale might not be serving Earth users, but enabling space infrastructure itself, according to Christophe Bosquillon of the Moon Village Association. He suggests space data centers are crucial for a future lunar economy, handling sensor data and autonomous systems. While significant technical hurdles like radiation survival remain, the concept forces a consideration of whether computation should be constrained by Earth's limits when energy is abundant elsewhere, even if everything else is difficult.
(Source:Live Science)