[jacquesm]

Ok, I read the paper. You got that completely wrong, the 53% is theoretical performance at 1 meter, the actual peak performance measured was 16% at an extremely small distance and two full orders of magnitude less (so 1%) at a distance of 100 mm.

The maximum output they measured was 37 uW, so 1000 of these would output ~37 mW, with an input power 1000's of times higher.

If anything this paper is a nice example of how theory and practice differ. It also highlights another big loss factor for ultrasound power transmission, the angle of incidence, the power falls of as the co-sine of the angle between the transmitter and the receiver (maximum at 0 degrees, minimum at 90 degrees).