The Cost and Network Limits of Space-Based AI Compute
arXiv:2607.14172
The paper assesses whether large AI data centers placed in low‑Earth orbit could be a cost‑effective alternative to ground‑based facilities, analyzing launch costs, power, cooling, radiation, and network performance, and concludes that while inference might be viable, training frontier‑scale models in space is unlikely to compete with terrestrial data centers.
Abstract
This paper evaluates whether large-scale AI data centers deployed in low-Earth orbit (LEO) could become a cost-effective alternative to terrestrial facilities. The analysis compares orbital and ground-based systems across launch cost, power generation, cooling, radiation exposure, and atmospheric reentry, as well as compute-network performance. A key distinction is the shift from terrestrial Clos networks to space-based mesh networks using laser inter-satellite links. Using bisection bandwidth, bisection intensity, and roofline-style models, we show that while LEO-based inference may be feasible, training frontier-scale LLMs in orbit is unlikely to be competitive with terrestrial data centers.