I think the key message of all the debunkings is this: "it depends".
The device may work great in one condition but certainly not always like the vendors might tout.
Given the triple point chart for water, the solar power for that day, and your humidity and temperature, you can compute the max you'll condense from the atmosphere.
Right. So the "major" (in my mind) contribution here is exactly the work of "compute the max you'll condense from the atmosphere" over all relevant regions. In the linked nature article, the published work is modelling to help determine regions where atmospheric water generator is viable for different assumptions and thresholds.
The device may work great in one condition but certainly not always like the vendors might tout.
Given the triple point chart for water, the solar power for that day, and your humidity and temperature, you can compute the max you'll condense from the atmosphere.