Editing Openai/691a2d9f-c044-8012-bccf-942bfce6334d (section)

==== If you’re far from any star (thousands of AU away from all of them), the only illumination is: ====
* extremely faint combined starlight from the galaxy
* cosmic microwave background
* a bit of infrared from dust
* cosmic rays

The total energy density of that radiation field (starlight, dust IR, CMB, cosmic rays) is of order 1 eV per cubic centimetre IUCAA<ref>{{cite web|title=IUCAA|url=https://web.iucaa.in/~dipankar/ph217/ism.pdf|publisher=web.iucaa.in|access-date=2025-11-17}}</ref> which corresponds to a flux on the order of 10⁻⁵–10⁻⁴ W/m².

Compare:
* Solar constant at Earth: ~1,360 W/m²
* Deep interstellar radiation: ~0.00001 W/m² (ballpark)

So going from 1 AU to deep interstellar space, you lose something like 8–10 orders of magnitude in available flux. Even if you deploy, say, 1 km² of ultra-light solar sail / panel (a million square meters):
* At Earth: ~1.36 gigawatts of raw sunlight (if somehow used as collectors)
* In deep space: maybe 10–100 W total

That’s not life support for a city-ship, that’s a camping lantern.

You might harvest a bit from:
* starlight when you’re near the origin or destination star
* faint ISM interactions (magnetohydrodynamic sails, plasma, etc.)

…but for most of the cruise, the external environment is effectively energetically sterile for anything on the scale of a generational habitat.

So: ambient harvesting is basically a rounding error. You cannot run propulsion, ecosystems, or cryobanks on it.