Tech • IA • Crypto
Why are Musk and Thiel fleeing the States?
7/10TL;DR
Tech billionaires are pursuing space- and ocean-based data centers to bypass Earth’s energy, regulatory, and environmental constraints.
KEY POINTS
A push beyond Earth’s limits
Rapid growth in artificial intelligence has intensified demand for computing power, straining terrestrial infrastructure. Data centers require vast electricity, water for cooling, and regulatory approvals that can take years. Increasing public resistance and environmental concerns are further complicating expansion on land.
Elon Musk’s orbital vision
Elon Musk is advancing a plan to deploy massive AI data centers in orbit, potentially at around 600 km above Earth. These systems could reach 1 gigawatt (GW) of compute capacity by 2027, with long-term ambitions scaling toward 1 terawatt (TW). Space offers continuous solar energy and eliminates many terrestrial constraints, but depends heavily on Starship, large-scale chip manufacturing, and enormous capital investment.
Energy and cooling advantages in space
In orbit, solar power is effectively constant, avoiding day-night cycles and weather disruptions. The vacuum of space also provides natural cooling conditions, potentially reducing the need for complex thermal management systems that dominate Earth-based facilities.
Peter Thiel-backed ocean alternative
In contrast, Peter Thiel has backed Pantalassa with $140 million, valuing the company near $1 billion. The firm is developing floating steel structures, roughly the length of a football field, positioned in the Pacific Ocean to host AI workloads.
Harnessing waves and seawater
Ocean-based data centers rely on wave energy as a renewable power source and use seawater for cooling, eliminating freshwater demand. Their offshore placement avoids land-use conflicts, local opposition, and much of the permitting complexity tied to national jurisdictions.
Operational limitations today
Current offshore systems are primarily used to run AI inference, not to train large models, which require far greater computational intensity. Data is transmitted via satellite, introducing latency and bandwidth constraints compared to terrestrial fiber networks.
A shared strategic conclusion
Despite different approaches, both initiatives reflect a common belief among tech leaders: scaling AI infrastructure on Earth is becoming increasingly impractical. The divergence lies in direction—space versus ocean—not in the underlying diagnosis.
Regulatory gray zones
Both outer space and international waters exist largely outside traditional national sovereignty. This creates opportunities to bypass regulation but also raises unresolved legal, security, and governance questions about who controls and oversees such infrastructure.
CONCLUSION
Space and ocean data centers represent competing visions for the future of AI infrastructure, united by a shift away from Earth-bound constraints and toward largely unregulated frontiers.