Ok, here's the idea - it's a desalination plant like the windtrap from Dune.
You make a mega structure like the windtrap from Dune 2000. You then coat the outside of the shell with solar panels and the inside with thermoelectric strips that make it colder on the inside. So after a while there should be a cloud within this several story building.
Then what you do is you put a Tesla coil on the inside so that when it goes off the lightning makes it rain. It turns out that lightning inside clouds can make it rain! Which might work even on smaller clouds with less moisture content.
You'd probably have to make it quite large to have the necessary efficiency gains. And I don't know if this would work all the time as you might just end up sucking all the moisture out of the air around the windtrap unless there was significant amounts of wind.
Thoughts?
I also have to image this can be done with basic evaporation and condensation using the sun, without forcing it with solar panels and thermoelectric strips. Why not let the sun heat up the ocean and capture some area, then at night when it cools down, and condenses, it falls into a collector instead of falling as rain somewhere else. Of course, at scale, now we're removing potential rain from somewhere that used to get it... hurting one area to help another.
I guess most forms of "get some energy into water so that some of it evaporates, then condense it" will produce desalinated water, regardless of how simple or complicated they are. Or most forms of "make some ordinary air colder, so it can hold less water". But it's hard to form an intuition for how much water you get out based on how much energy you put in.