FWIW, to arrive at my numbers, I used hourly sun data for a single site for a whole year and then optimized the amount of solar versus storage to minimize costs and prevent total depletion of the batteries while assuming a constant load for the whole year (and then iterating so that the amount of energy in the batteries at the beginning of the year is the same as that at the end). It turns out that even rainy days produce some solar power, which helps stretch the 48 hours of storage.
To make it work, it's cheaper to throw away about 2/3rds of your energy than to build 150 hours worth of storage. Some amount of "curtailment" is going to be optimal, and in my case, because solar is so cheap and my starting assumptions allowed no other energy sources, a LOT of curtailment is used. Literally most of the energy is unused.
If you ignore massive curtailment as a possibility, then yeah, haha, it's not going to work.