Edit: You can never know in advance, but I would guess that if the chip is good in a random small scale device, it is even better if you pack a few hundred in a dedicated device with proportionally designed power supply, fans the rest of the stuff that is needed. Is it possible to construct a chip that is better than current mining, and at the same time can not be used large scale?
Probably irrelevant. The users won't care nor notice the electricity used. No one here in the Philippines studies their electric bill to see how much their smartphone charging was.
NO ONE. NO ONE. NO ONE.
Capice?!
Now, my phone gets pretty warm when it's charging so the new mining chip can't draw that much more power. But then again, a lot of times my phone is plugged in and already fully charged, and so the mining chip might as well be running. So let's say that 21 Inc. can boost the yearly energy consumption of the typical smart phone from $0.50 to $3.00 before people start to complain about "hot phones." $2.50 goes to mining bitcoins
Thankfully there are some rational engineers on this board.