The new numbers match what I said on YC the last time this came up.[1] Although uBeam's web site talks about megahertz ultrasound, as I pointed out, that won't work - air will absorb all the energy in less than a meter. They have to operate in the 40KHz - 100KHz range, where everybody else does air ultrasound and where the attenuation is moderate. Sure enough, they're now saying 45KHz to 75KHz.

As I pointed out last time, if you put together a 1 meter square array of off the shelf American Piezo ultrasonic transducers, about 4000 of them, you could get to their claimed 155dB level. But it would take about 2KW to drive the thing, comparable to a small clothes dryer. Their new figure is 3KW, a medium-sized clothes dryer.

They claim to be able to deliver 1.5W (not KW) to a device, for an efficiency of 0.05%. The rest of the 3KW will heat the air. The energy has to go someplace.

As I pointed out last time, this thing can potentially transmit some power, but it's an insanely inefficient way to do it. It can also double as a space heater and steerable ultrasonic welder.

I'll bet we'll see a demo of this at a trade show, with about a 2-3 meter range, a transmitting array about a foot square, and some mobile device with a special case as the target. The demo will work. The transmitter will draw about 300W, and deliver a few watts at the target end. It will have enough smarts to turn off if you put an obstacle in the beam. You could build that right now from off the shelf parts. As a deployed technology, it's unlikely to be useful.

Remember that guy who demoed power transmission over a few meters by ganging up enough WiFi transmitters to get some real power? Whatever happened to him?

[1] https://news.ycombinator.com/item?id=10421628