In a recent filing with the U.S. Federal Communications Commission, Elon Musk outlined an Orbital Data Center System consisting of up to one million satellites, implying scaling on the order of ~100 kW of electrical power and compute per satellite at approximately one metric ton of spacecraft mass. This paper provides a first-principles engineering assessment of whether such a capability is physically feasible in low-Earth orbit using solar power alone. The analysis evaluates orbital duty cycle, solar power generation, battery sizing, thermal rejection, radiator geometry, deployable structure mass, and constellation-level operation. An optimal operating regime at a 70–80% sunlight duty cycle is identified, in which high-power computation is limited to sunlit orbit and batteries are sized only for eclipse survival. Under this regime, average delivered compute power of ~70–80 kW per satellite is achievable without prohibitive battery mass. Quantitative mass-closure and thermal sensitivity analysis establishes a hard feasibility boundary near ~1,000 kg total spacecraft mass, with heat rejection - not compute silicon - emerging as the dominant constraint. Required solar array areas (≈400–600 m²) and radiator areas (≈100–250 m²) are shown to be practical with current or near-term technology and consistent with existing spaceflight heritage, provided radiator operating temperatures are elevated and areal densities remain aggressive but plausible. While individual satellites must duty-cycle compute, the paper shows that compute continuity is achieved at the constellation level: at any instant, hundreds of thousands of satellites remain sunlit and capable of peak computation, enabling continuous aggregate service through distributed scheduling and checkpointed workloads. The analysis concludes that the FCC-filed Orbital Data Center concept is technically credible when interpreted as a peak-power, duty-cycled architecture, and that continuous 100 kW operation per satellite is infeasible at this mass without nuclear power or substantially larger spacecraft. The decisive constraints are thermal geometry, mass efficiency, and system architecture - not the availability of compute hardware.
Sanjay Kumar (Sun,) studied this question.