It's an upper-bound, sure, but I don't think it's relevant because I think that if we have a moon base, it's going to necessarily decrease by orders of magnitude. At that point, we're dealing too much uncertainty; we don't know how to mine water on the moon, and we don't know how much the transportation cost will decrease.
It's going to "necessarily decrease" in inflation adjusted terms: Sure. It's going to necessarily decrease by orderS of magnitude? No, I think that's too strong. Our most wild technological extrapolations only cover 2 or 3 orders, tops, and there's no guarantee they can or will be built. I do not see prices dropping to 2 cents per ton any time soon.
On the flip side, yeah, actually we can make a pretty good guess at how to mine this water. We actually are pretty good at mining. Run in through a machine that literally boils it out. Done. Throw the leftover rock in a big pile. At 2% concentration that's more than enough to get you started, it just needs energy, which is also fairly easily to get in this context. A real miner might do something else but that would certainly work.
My claim is that if transporting a liter of water to the moon does not decrease by orders of magnitude, then we won't have colonies on the moon. No colonies on the moon mean no reason to mine the water.
In other words, I think that the only circumstances in which we would want to mine water on the moon, transporting water would be orders of magnitude less expensive than $20,000/liter.
The "unknown" bits I was referring to are the infrastructure required to be on the dark side of the moon.
