I wonder what the water-used-to-calories-produced ratio is for something like lettuce compared to something like corn or beans. I imagine it's a good number of orders or magnitude higher.
In other words, an increase in the price of fresh water does not necessarily significantly increase the price of feeding yourself, it significantly increases the price of water-intensive crops and likely increases the demand for less water intensive crops.
However, his estimates for the energy required were very conservative, and even if these numbers don't work out with today's water prices, they do make sense at a certain point that is likely lower than one encouraging violence. That effectively provide a cap where we start to go this route if necessary, and the vast supply means that we shouldn't go above that for very long.