I would totally buy into that. One of the major advantages of the ESP32 over the ESP8266 is that there are two microcontroller cores; one to deal with wifi events and one to deal with user code simultaneously. This basically means there's no (or much less) chance of the MCU crashing or dropping signal due to user code tying keeping it too busy to deal with wifi events.
The major problem with the ESP32 right now is not lack of hardware in the channel, it's library development... so if Erik made an Oak Pro or something, that's where a lot of work would end up going if he wants to push some units. Another thing would be to sell it cheaper than the competitors. The Sparkfun ESP32 Thing is about twenty bucks, and the generic Ebay units are about seven. He would have to be able to turn a profit somewhere in the middle.