They've made fluorescent tubes with daylight spectrum for a long time (I recall seeing them available, but expensive in the 90s). Those work with phosphors, so I don't see why an blue LED pumped phosphor couldn't work as well as one driven by a near-UV mercury vapour lamp.
To exactly match daylight you would need all the visible light, plus infrared and UV, and a total power of 1 KW/square meter. The good thing is that the lamp would also work as a space heater.
Thanks for the explanation. I've been learning a little about light sources; based on my limited knowledge, I wouldn't expect that the full spectrum (or spectral power distribution (SPD)) of daylight could be replicated by hacking together LEDs and phosphors, though perhaps the untrained eye would not be cognizant of the difference. Can they really pull off 'daylight'?