One interesting thing from that IEEE vid was the initial aim to transfer one watt, which even at the time would have been useless for phone charging, more so nowadays.

I've got calcs (using proper industry simulation data) that show with a @45KHz 155dB 1500Pa 2900W/sqm source, the likely typical efficiency is:

@1m = 1.8% giving 53W/sqm

@2m = 0.3% giving 8W/sqm

@3m = 0.05% giving 1.6 W/sqm

@4m = 0.01% giving 0.2 W/sqm available

The surface area of a typical iPhone thingo is about 0.01sqm

Assuming 100% efficiency on the receiver, a power density of 100W/sqm is required to charge at 1W

Even at 1m using a massive power hog transmitter and ideal receiver they would barely get 0.5W, quickly dropping to bugger all.

They claim a 4m radius

Absolute best case calcs are not much better at distance.

And the more power you pump in, the closer the saturation and nonlinearity happens.