My understanding is that the same thing AP have been trying to...

My understanding is that the same thing AP have been trying to do "integrate" everything into a small form-factor (MEMS + ASIC), is being attempted or has been achieved by USound using the same approach. In fact they are using the same nomenclature, to describe their product.

Not only that, but the scalability of their invention is also detailed in their patent:

"Due to the possibility of the separate production, of the respective modules, the asic - modules and / or the mems - modules also in different variants and then made to different sound transducer arrangements to be combined by, for example, various mems - module - variants with an asic - module - variant or a mems - module - variant with different asic - module - variants are combined. This makes possible the flexible design of an extensive family of products of different sound transducer arrangements in the case of a simultaneous utilization of the economies of. "

(Reference Mems - a loudspeaker arrangement with a sound generator and an acoustic amplifier, Patent No. DE102014112841A1).

USound's patent does read as though it is still a conventional speaker, but the integration would allow you to control the diaphragm in the same way the AP MEMS device is actuated, in a pseudo-binary manner.

I'm still questioning how AP and USound have overcome the dipolar response of both of their devices, it's a lot easier in theory than it is in practice.