Interestingly, when frozen planets get terraformed, they become vastly better. eg: a 100% hazard from 175% because it removes the usual "Extreme Cold" but also a "Poor Light" condition. This is better than an originally-water world (which was also at 175%) which had other negatives like storms etc. Clearly we should shade the water world so it freezes and then warm it back up so there's no storms.
Also, Organics shows up (standard quality) which is handy since food (water world) and organics deposits are a bit rarer than say volatiles and the two ores.
Do all water worlds spawn with organics in them?
In this case, I think upkeep cost would work better than initial cost as a balancing mechanism. At 3 cycles, 10m credits, and 1 structure slot, terraforming is something people won't bother considering until they're just looking for an excuse to spend money. I have nothing against money sinks, but 10m would seem out of place, and I think such high initial costs would make terraforming non-viable for practical purposes, especially with such a long delay before seeing any benefit.
Agree, compared to vanilla structures, the time on these at 1080 days is much longer so that a "full up-front" seems less of a match to the "it's a long process" feel of it.
Take 10,000,000. Say 1 mil is your upfront (like the battle station -> star fortress) then your upkeep would be 9mil over 36 months or 250,000 a month. Now, might have to adjust though, since you have 1. hazard, 2. upkeep bonus from faction-supplies and 3. cores.
Then again a lower hazard world would be easier to make stellar shades in the orbit of... maybe? And I can see how using in-faction supplies and cores would work. Also, borrowing from the Waystation, perhaps making the structure demand various goods, with a steep upkeep penalty for shortages would also work. Like a lot of metals, transplutonics, (building) crew, heavy machinery and ship hulls (building in space) etc
Perhaps not make the goods requirements change with size (would be odd) but high so that you need likely player production or very high accessibility (though if the AI factions suddenly can't send enough, it will hurt) which again makes sense.