Re: saturated market
Will such devices need the enormous processing capabilities which have been developed and deployed for smartphones (Modem, WiFi, Bluetooth, GPU(s), CPU(s), GPS, Audio, Video, NFC, Camera(s), etc...)?
Users interact with their smartphones on a very regular basis and are therefore very aware of benefits in speed, interactivity and the captivation of the applications (and by implication the processing resources they use - although they typically don't equate this). If there is a new killer app that runs like a dog on their existing smartphone, then they will be tempted to upgrade. Many just upgrade anyway due to the particular contract finance model that perpetuates the ongoing upgrade process.
What would an Internet of things device offer to be as captivating as a smartphone or tablet? The only devices that need heavy duty graphics/video processing in the mass market are: 1) The PC/Laptop, 2) The Smartphone, 3) The Tablet/Fondleslab, 4) The games console, 5) The SmartTV, 6) The Smart STB/PVR/Dongle, 7) In Car infomatics/entertainment
More likely is that Internet of things silicon devices will merely contain lower cost silicon chip devices which include Low power CPU, Wifi (or other digital modem), Sensor processing (could be image and/or Audio or other environmental sensors), device control (e.g lighting, motors, compressors, actuators, ..)
They will be unlikely to be in any way as captivating as a smartphone or tablet. Remember also, the smartphone is in effect the modern day equivalent of slapping down a set of executive car keys on the bar at the pub! ... i.e. a visible status symbol which drives growth (latest greatest model)...
So the industry is going to be looking for the next golden goose. Before the smartphone, many silicon vendors were focussed on picking up crumbs from Intels Table (i.e. the PC such as graphics processors, media processors, mixed signal devices ...). The smartphone has enabled stellar growth for Silicon IP vendors who had the right technology/biz model/engagement at the right time (ARM CPU, Imagination GPU... ) ... Just how much more graphics (and CPU cores) do we need in a smartphone that will be so compelling to drive the future massive customer upgrade cycles?
The CPU, by its very nature, is extremely generic and can be deployed in many markets and can leverage the tools and o/s utilised in the mass market smartphone environment. Less so for the GPU unless it is used more as a SIMD/MIMD style data processor for sensor processing. It will probably be much less efficient than a dedicated hardware (old school ASIC) approach, but time to market pressures and vast swathes of cheap silicon real estate will make it economically viable and offer fast time to market (so long as the softies don't screw it up).
So in this decade we will increasingly see IP vendors picking up crumbs from ARM's table and having to focus upon fragmented smaller emerging markets. Devices in the new emerging Internet of things will have a much longer in market utilisation time by the user than we have seen in the smartphone/tablet market. An internet fridge is not exactly a status symbol and not something that you are going to upgrade every 12-24 months!
Interesting times ahead.... Thinking caps on silicon jockeys!