The rapidly falling cost of rocket launches is unlocking a radical new possibility: making space a viable and even economical location for datacenters. Google is seizing on this trend, announcing research that suggests by the mid-2030s, running an AI datacentre in orbit could be comparable in cost to running one on Earth.
This economic calculation is at the heart of Project Suncatcher. While the $3 trillion projected spending on terrestrial datacenters balloons, driven by the AI boom, Google sees a future where the economics flip in favor of space. The falling launch prices are making the initial deployment of orbital hardware increasingly affordable.
Once in orbit, the operational costs could be significantly lower. The primary advantage is energy. Space-based solar panels can generate up-to-eight-times more power than on-ground arrays, providing a virtually free and unlimited energy source. This eliminates the massive electricity bills that plague terrestrial datacenters, which are a primary driver of operational costs.
Furthermore, the orbital model removes the need for expensive land and the vast quantities of water used for cooling, both of which are major cost factors on Earth. The data would be beamed back via optical links, bypassing some traditional infrastructure costs. Startup Starcloud, also entering this field, claims the energy is “unlimited” and “low-cost.”
This financial bet is not without risk. Google acknowledges that “significant engineering challenges” in thermal management and reliability could drive up costs. Moreover, this calculation must also factor in the environmental “cost” of launch emissions and the potential backlash from astronomers, which could lead to regulatory hurdles.