Re: Well Intel would have a chance
"The problem is that Intel doesn't understand what would be needed. They bring out x86 SoCs with 16 kilobytes of RAM."
I may have a *bit* more insight, as I've actually communicated "certain requirements" to some of the people at Intel, with respect to "what it would take to use XX processor on YY". I describe (basically) what the existing processor choice does and there are features on that CPU that are integrated with other hardware, and hard to duplicate without actually "being that". For example, a differential A:D device capable of having a gain adjustment and internally selectable reference voltage. If you want to scale it to read millivolts DIFFERENTIAL for example, you could with the existing CPU, and with 11 bits of precision. THESE are the kinds of things that IOT devices need.
what's interesting is that ATMel's 'xmega' and ARM 'SAM' series CPUs can do these things. they have the hardware support with "that kind of thing" in it. they are also VERY low power devices, with super-special power saving modes and wakeup timers built-in. If Intel wants IOT market share, it's going to have to address THAT.