My question is: should it actually cost 3Y? Sure, if you take an off-the-shelf design and triple it, it will cost around three times as much.
But there should be at least some efficiencies that can be gained. For example, energy conversion is not cheap. Rather than building solar inverters that produce 480VAC and using 480VAC pumps or pump drives, what if the pumps ran on DC?
Or, for more fanciness, what if the pump drives were themselves the MPPTs? Have the solar panels arranged so they produce a sufficient voltage for the motor drive bus even at minimum output, and make the drivers smart enough to run the right number of pumps at the right power and to do so while tracking the maximum power point? That’s a lot of power electronics that could be avoided.
Also, if a plant is willing to store a day’s worth of seawater and brine, then the intake and discharge systems only need to be sized for maximum daily average flow, not peak sunshine flow.