Thermodynamics rules future orbital data centers

Economic feasibility and underlying motives

• Multiple comments stress that, even under optimistic assumptions about very cheap reusable launch vehicles, orbital data centers are still ~10x more expensive per GPU-year than terrestrial ones; a separate calculator suggests 2–3x in optimistic scenarios.
• Some see narrow niches where high costs might be justified (e.g., on-orbit collision avoidance), but doubt broad economic viability.
• Several argue orbital DCs mainly serve to justify high valuations for launch companies and create demand for high-volume, low-cost launches, rather than solve a real computing problem.
• Others note that, with very high inference margins and severe on-Earth siting constraints, even a 10x cost penalty could be attractive to some actors.

Thermal management and power in space

• Discussion focuses on the difficulty of dumping heat via radiation only; cooling is viewed as a central constraint.
• Ideas include running GPUs hotter to shrink radiator area, using heat pumps to raise radiator temperature, and exotic fluid/metal droplet radiators.
• Some question why radiators get more attention than large solar arrays, but most agree continuous high-quality solar power is a major attraction of orbit.

Radiation, reliability, and maintenance

• Several posts see radiation (cosmic rays, ionizing damage) as at least as concerning as heat, requiring heavy error correction, hardened parts, and possibly lower orbits.
• There’s skepticism about running state-of-the-art GPUs in such environments without large overhead.
• A major critique: data centers rely on constant hardware replacement. In orbit, repair is effectively uneconomic, making whole satellites disposable and further worsening lifecycle ROI and waste.

Alternatives: oceans, Arctic, and ships

• Many argue underwater, Arctic, or offshore data centers are far more practical: cold water/air for cooling, geothermal or hydro/wind/wave for power, and physical access for maintenance.
• Counterpoints: corrosion, severe permitting and infrastructure issues, sabotage risks, and power/fiber delivery in open ocean.

Environmental, regulatory, and social context

• Some see orbital DCs as a way to escape local NIMBY pushback and regulation; others think locals could simply be paid more on Earth instead.
• Concerns are raised about rocket emissions in the upper atmosphere if thousands of launches become routine, versus ongoing terrestrial DC emissions and siting in fragile regions like the Arctic.

Cultural framing

• The thread is laced with sci‑fi references (Skynet, matrioshka brains, Elysium, space manufacturing, asteroid mining), with many viewing orbital data centers as closer to speculative fiction than near-term infrastructure.