Big Tech Dreams of Putting Data Centers in Space (www.wired.com)

Big tech and AI leaders — including OpenAI’s Sam Altman, Jeff Bezos, and Eric Schmidt — are seriously entertaining the idea of putting large data centers in space to dodge the mounting energy, water, and pollution costs of terrestrial hyperscalers. AI center power demand could rise as much as 165% by 2030; today over half of that power still comes from fossil fuels and U.S. facilities (about 5,400 and growing) may consume up to 12% of the nation’s electricity by 2028. Proposals range from near-term orbiting racks powered by space-based solar to far-out visions like Dyson-sphere–scale installations; startups (Starcloud, Axiom, Lonestar) and consortia such as the $500B Stargate project are already testing miniaturized satellites and lunar payloads. Technically, the idea is becoming less fanciful: launch costs have fallen to roughly $1,500/kg and lightweight, efficient solar arrays could theoretically deliver electricity near $0.10/kWh. But major hurdles remain — radiation exposure, repair/upgrade difficulty, obsolescence, latency and throughput trade-offs, and current per-kilowatt economics that still favor terrestrial sites. Early use cases are likely niche (space-data processing, national-security resiliency) rather than wholesale migration. A final incentive: limited space regulation could accelerate corporate moves before lawmakers catch up, raising urgent policy questions about jurisdiction, safety and environmental trade-offs.