When Starship reaches large-scale operations, space could become the most economical location for AI compute infrastructure. Full and complete reusability is the game changer here. Operating costs drop dramatically once you eliminate wasteful single-use expenditures. Think about it: with proper reusability architecture, propellant expenses per flight shrink to roughly a million dollars. That translates to just a million dollars in transportation costs for delivering 10 megawatts of continuous AI compute capacity. This cost structure fundamentally reshapes where compute infrastructure makes economic sense. Traditional data centers operating on Earth face mounting cooling, real estate, and power grid constraints. Meanwhile, the marginal cost of launching additional compute into space approaches something almost negligible once your launch vehicle amortizes across repeated missions. It's a compelling case for rethinking our computational future.