The technology scales very well, it can be augmented by sunlight or run without it. Bear in mind that plants only need 17 hours of light a day, it's actually beneficial to run with darkness, so you only need to ensure that much light and light exposure. Plants can also be cycled from the inner areas towards the outer edges to ensure they get more actual sunlight rather than just artificial.
Also, research indicates that the light from a dual red-blue LED is actually as good or better than sunlight. LED's are cheap, and very efficient in terms of power consumption.
It's been cost effective for quite some time. Check out the site, they've already done the math, the systems have been tested extensively. Combine that high level technology with the natural concepts and organization from www.growingpower.org and it only gets better.
However, I would also like to point out that something the size and scale of a modern skyscraper would likely be a centralized system. It would be best if there was a harmony between any centralized systems (big skyscraper full of 3D farming systems), and smaller decentralized systems (community owned and operated greenhouse full of such systems, an omega garden in every home, etc), to smooth out the supply and demand as well as resource usage.
Just a quick point I want to make. Treating any such systems as mutually exclusive ("We can only build X or Y") isn't as efficient as treating them as blended and integrated. Not that I'm seeing anyone do so, but I just thought it would be prudant to point out that no one is suggesting a singular approach to any such problem solving. Especially relevant considering that my suggestion of 3D farming with systems such as Omegagarden only really apply to mostly the leafy green veggies, herbs, and some fruits such as tomatoes and a select few crops. It may not necessarily apply to grains or corn. Hence why a singular approach will not fix everything. It can still fix quite a few things though.