I have a feeling the 10-10-10 hypothesis is a bit generalised as well. It's sounds more like rhetoric than research. The take away, I suppose is more of an appreciation of how mitigating the constraints can improve shipments.
Yes IoT if we think of implementing in present day things, it is very much possible using the devices and readily available protocol stack, but what the technology will demand is power efficient and reliable implementation now this will be an challenging opportunity for the chip giants.
IoT definitely is a bowl of ample opportunities for everyone involved in the electronics industry. Imagine everything in your home is on internet and can be monitored and controlled from any part of the world. It will be a boon for people who keeps traveling, for mothers who always need that extra time for themselves. It will revolutionize the humanity as whole.
"If we can get [power] consumption down by a factor of 10, compute performance up by a factor of 10, we can get cost factor down by a factor of 10, we can ship 10 times as many units. This will be key to differentiation and opportunity for this industry."
Well stated, and quite likely a prediction that will be borne out. Those "ifs" are pretty big though.