Typical coefficients of performance for air-source heat pumps, like air conditioners or dehumidifiers, are around 2; that means for every 2 joules of heat that it removes from the cold reservoir, it consumes a joule of electrical energy (dumping it as heat into the hot reservoir along with the heat it's pumping). The enthalpy of vaporization of water, at almost 41kJ/mol, is huge compared to the specific heat of air or even water itself, so that's a good approximation to how much heat you need to extract to condense the water. Water's molar mass is about 18 g/mol (two hydrogens of 1.0001 or so and an oxygen of 15.9994) so that works out to about 2.3kJ/g. At 10¢/kWh, this means your water costs 6.3¢/ℓ or US$78000/acrefoot.

This is a lot cheaper than bottled water, and a lot more expensive than what farmers can afford to pay for irrigation water, and a lot more expensive than reverse osmosis desalination, which I think is around $8000/acrefoot. (Corrections welcome!)

If we figure that a person needs 8 ℓ / day for cooking, drinking, and bathing (more or less what we use at Burning Man; if you're somewhere less hot and druggy then you might need less) which is about 210 watts. If a photovoltaic panel produces a 24-hour average of 13 W/m², which is what a friend of mine in England is getting, you need about 16 square meters of photovoltaic panels to power the device, which costs about US$2000 at this point. That's well within the bounds of feasibility, although it's a big enough asset to be a temptation to thieves.

So when you say, "I feel like I'm missing something here," I agree: you were missing quantitative understanding of the subject, and as a result your quantitative conclusions were completely wrong, even though they were based on correct qualitative understanding.