It seems like the play here would be to not create components but to create real designs. If you can get these levels of efficiency why wouldn't you go to market with just a component? The big profit will be in the end equipment...
I did ask Pete Magowan but he didn't want to give details.
I am thinking Amantys, like many startup companies, feel they can get more publicity by giving out key information in little chunks every two months or so while keeping a lot back.
But we could speculate that Patrick Palmer has come up with a load-balancing/feedback/feed forward mechanism/algorithm that allows heavy duty power devices to be operated in series without snubbers and maybe with a lot of other efficiency refinements. That algorithm could then be implemented in an ARM microcontroller (actually probably in something much simpler).
The Amantys go-to-market strategy would then be to produce power modules comprising IGBTs and programmed MCUs for inverter companies, while it talks about IP to power device makers with a view to in-package or monolithic integration.
Just a guess.
Very interesting but would be more exiting if they had elaborated how they accomplish this with ARM. This area could gain traction in the coming years since there is a huge gap between supply and demand and companies are stuggling to control the load distribution more efficiently.