5W of ultrasound focused on something the size of a phone (roughly .01 m^2) is at least 500 W/m^2. That's 147 dB, I think?
(10 * log(500 (W/m^2) / 10^(-12) (W/m^2))/log(10))
A giant 0.1 m^2 tablet charging at 12W (120 W/m^2) needs
only a 140 dB field. A watch, which might need 1W but with a cross section of more like .001 m^2 (for 1000 W/m^2), needs a 150 dB field.
If something intercepts the beam at 1m instead of the 3m the device is charging at, the cross section of the beam is about 1/3 by 1/3 the final cross section, which since it has to have the same energy (actually would need more to compensate for attenuation through air) would be nearly a 10dB increase?
Is 160dB of ultrasound (at 40kHz or 110 kHz) safe... for a few seconds? for a few minutes? a few seconds of exposure, daily? To your eardrum? To your eyes?
Higher energies needed if the phone isn't perfectly oriented. Worse case is small edge-on orientation to the transmitter.
Higher energies needed for less than 100% transducer efficiency, and I don't know what kind of engineering magic they've done for the transducer but what percentage of the energy could a thin skin over a device possibly convert? 80%? 50%?
The beam could be 170dB, or more.
I want to believe, but this is too sketchy without more information. Large companies have been conned out of millions by small teams peddling snake oil. I'd first want to see it demonstrated with nothing but the transmitter plugged into a socket (through a power meter), a phone at a known, low battery level, and nothing else with wires or metal in the room. Then I'd like to see a test of the transmitter aimed (from above) at a glass of water with a visible thermometer, to see what it does to water.