The idea of moving data centers into space faces a fundamental economic hurdle. While the potential benefits – abundant solar power, avoiding terrestrial NIMBY opposition, and solving future AI-driven capacity crunches – are compelling, replicating even a single large ground-based data center in orbit requires hundreds of satellites. This is a massive undertaking.

Historically, building anything in space has been astronomically expensive. The International Space Station, with less habitable space than an average home, cost over $150 billion. While launch costs have plummeted from $10,000 per kilogram to potentially under $1,000, and satellite hardware like Starlinks is more efficient, the sheer scale remains daunting.

SpaceX's ambitious plan for a million-satellite megaconstellation hinges on Starship becoming highly reliable and reusable, drastically cutting launch expenses. Yet, the cost of the satellites themselves and the monumental $20 billion investment SpaceX is making in its Terafab chip fabrication project to avoid Nvidia premiums add significant layers of financial uncertainty. The viability hinges on achieving unprecedented cost reductions across all three major economic factors: launch, hardware, and semiconductor production..