If you don't expose your plants to the sun, you're going to need a lot of energy to power their illumination. I'm not sure solar panels-> tuned led illumination

If you don't expose your plants to the sun, you're going to need a lot of energy to power their illumination. I'm not sure solar panels-> tuned led illumination is any more efficient than direct exposure, nor could it be any cheaper (and because solar irradiation on Earth is finite of course the amount of food you can grow is also finite). Then there's the profit per m^2 of agriculture isn't that high -- how much does a structure cost? I'd say at least $400/m^2. If your profit is <$10/m^2, you'd need decades to recoup costs (not a viable investment currently) -- that's only for the structure discounting all the other necessary additional infrastructure, lightning, irrigation, etc. Unless we switch to radically different/more efficient food growing than plants (say direct cell cultivation, or direct nutrient synthesis), then the current approach is probably the best (still with room for some automation and significant genetic improvements on crops, I'd say). I think direct nutrient synthesis probably has potential, but it will take a very long time to understand nutrition and synthesis of organic mollecules -- not an expert by any means, but I'd be surprised if it took less than several decades to develop the necessary tech/science. Even then the gains in efficiency will probably not be that large -- plants/life in general have been optimizing synthesis for billions of years, they're already quite efficient. What we would gain is minimizing to its limit parts of the plant we don't consume, and concentrate on the essentials/optimal nutrients.