We can put a quick upper bound on this, as it is being sold to the press. Everyone here probably have wifi analyzer or something similar on their phone:
Right now, I have -55dBm from my closest (8m) access point, and -70 dBm from next closest. -70dBm is 30x weaker in linear power than -55dBm, so we can ignore that, and the neigbours.
0dBm is one milliwatt coming from the antenna in the phone. -30dBm is one microwatt. -50dBm is 10 nanowatt. -55 is 3 nanowatt.
As a sanity check we can take 100mW EIRP and divide by the area of the sphere, we get an upper bound of 12nW/cm2 with no walls.
Ahem. And that assumes 100% tx duty cycle on the access point. Reality is 5%.
for the led demo, they'd need in the order of 100uW after losses, 200uW into the antenna, for a 10x10cm antenna we need 20mW/m^2, we are 0.7m from a continuous 100mW transmitter.
(I am not going to try to convert to sheep in vacuum * norris )