They are fabulously expensive for the water output you get. The San Diego one is (or was) the largest in the Western hemisphere, cost about a billion dollars to build, requires 40MW to run, and only covers about 7% of the potable water needs for the San Diego area.
I think some context can be helpful here. As a nation, we're happy -- hell, our elites (yes, we on HN) practically begged -- to spend $6.4 trillion dollars over 20 years to have fun foreign adventures in the sand, playing guardian of freedom and feeling real nice about it [1].
$1bn for 7% of the water needs for a major (and growing) US metro area sounds like a bargain to me. For a measly $14bn we'd cover the whole city. Defund the Pentagon, turn it into a WeWork, and let's do fun, neat stuff with our civilization. If we're gonna throw piles of money at things with low ROI, why not throw it at fusion and desalination instead of drone striking children at Pakistani weddings? 100% sincere here, not being cheeky.
Sure, maybe -- I'm no scientician -- but $6.4tn can buy mega megawatts. For that much money we can put everybody in Southern California on treadmills hooked up light bulbs.
Do you have a figure for price per gallon? It might be interesting to do a cost benefit analysis between desalination and the water pipeline from Canada mentioned elsewhere in the comments.
The best I have is from [1], which estimates 50 million gallons a day at an ongoing operating cost of around $50 million per year. You'd have to pick a timeframe to also amortize the cost of the plant itself and add that in.