You can translate cd/m^2 + Area directly into lumen.

lumen = cd/m^2 x m^2.

That is, 1 lumen of light energy will illuminate a square metre of area with a brightness of one candela.

So your 1000 cd/m^2 source over 0.5 m^2 = 1000 x 0.5 = 500 lumen.

Modern available LEDs are achieving 200 l/W (just) (Cree XM-L2 top flux bin, lowest Vf), with LEDs with typical values of over 150 l/W being commercially available off the shelf (I have some). But, even allowing 100 l/W the 500 lumen source is equivalent to 5 Watts of input.

That's minimal compared to alternatives - it's well under a 100 Watt incandescent bulb.

You can get much more output from some of the new 60W equivalent LED bulbs with CRI's of over 90 and people starting to focus on providing CRIs of essentially 100. At the 5 Watt level you could achieve a very high CRI by simply mixing a range of low power LEDs of different wavelengths with your main "white" emitters.

Note that the lifetimes claimed as possible only approach those of good quality power LEDs (50,000 hours +) when expensive encapsulation is used. Quoting:

• "For our device to reach 20,000 hours, they still have to be encapsulated, but they are not so sensitive as OLEDs," said Carroll. "If you use expensive encapsulation, you will get 40,000-50,000 hours."

The article you cite is from late 2012. Given the typical time from lab to value for money buyable product I see nothing written there that suggests that the new devices will be available within a year with reasonable outputs and/or at reasonable prices. I'd be immensely pleased if they were.