If I was in change of a lunar colonisation plan, first thing I’d do is send up a rover with some limited ability to process lunar regolith — at a minimum into aluminium, preferably also PV. This rover would slowly crawl around the moon (doesn’t need to be the equator, but at least 5.14+6.68° from the poles to avoid being in darkness at the wrong time of year), leaving a fat conductor behind it. If the rover needs human interaction, it’s only a few seconds of light lag.
Once the processing has circumnavigated the moon, the conductor is already in a desiccated vacuum on non-conductive rock, so no extra insulation needed for that to be an HVDC line.
Then send some transformers and as much PV and industrial equipment as you want, because now the 14-day-long lunar night is easy to get through. The current going through the conductor gives you a free magnetosphere whose strength goes up the more PV is connected.
If you can’t build such a rover, you’re probably not ready to colonise the moon in the first place.
Humans can be surprisingly useful still, even if we are very costly to maintain. https://en.wikipedia.org/wiki/Moravec%27s_paradox
I'm hopeful about artificial magnetic fields - modern superconducting magnets could do wonders in orbit with solar power.
Or here's sending ions from Phobos to do it: https://arxiv.org/abs/2111.06887