Data Center Debate, with Philip Johnston (CEO of Starcloud)

Scaling AI compute requires moving data centers to space because terrestrial energy and water permitting constraints create significant bottlenecks. Starcloud CEO Philip Johnston argues that as AI energy demand doubles annually, space-based infrastructure becomes a necessity rather than a luxury. Key economic drivers include reducing launch costs to approximately $500 per kilogram, optimizing watts per kilogram, and minimizing dollars per watt. Space-based data centers mitigate terrestrial energy shortages and provide a strategic advantage in global AI capabilities, particularly against international competitors. Despite technical hurdles such as radiation shielding and thermal management in a vacuum, Starcloud has successfully operated an NVIDIA H100 in orbit, demonstrating that specialized hardware can survive and function in space. Future scalability relies on lower launch costs and the potential for robotic maintenance to support long-term orbital infrastructure.