You could probably distribute the light among a larger number of plants (think plants stacked on a shelf, or on a tilted surface), so you grow more instead of faster.

IIRC there were some studies saying that higher CO2 in the atmosphere leads to better growth but lower nutritional value so that's one possible side-effect.

I think you have a point. There's no free lunch in evolution based systems, at least that's my perception.

No evolution needs to happen in a controlled environment.

Plants have evolved to adapt to a certain environment, and if we change those parameters to promote a certain behaviour (e.g. higher yield) then most likely it comes at a cost (e.g. is the produce less nutritious?)

I'm curious to know if there's a limit to selecting plants cultivars. What should we optimize for? Vitamins, minerals, proteins, glucids, lipids, fibers or other macromolecules? Should we allow some sacrifice in taste, appearance, yield, resistance, conservation?

Why does the change have to be negative for humans? The products could just as well become more nutritious.

Evolution does not optimize to global maxima, it prefers local maximas. The path it has taken so far is also pretty convoluted.

It's pretty clear that sunlight isn't the absolute best spectrum for growing plants, so we know plants haven't evolved optimally to make use of the sun's spectrum as-is. I think a local max is the right idea. this tech is shortcutting that evolution and providing a better spectrum without using electricity, like LEDs would.