For the last 30 years PCs and the PC supply chain dominated the technology industry, and PCs and their components were used for all sorts of things that weren't actually personal computers - everything from ATMs to industrial control equipment. Today that's clearly ending - it's now the smartphone supply chain that dominates. This chart makes that point pretty well, I think.
One effect of this is that smartphones are both destroying and creating entire other product categories. On one hand, things like point & shoot cameras, voice recorders, clocks and music players and so on are very obviously being subsumed into phones. But on the other, smartphones are enabling new types of products. The smartphone boom is creating a flood of small, cheap, low power and yet very sophisticated components that would not have existed otherwise, or would have been much more expensive. The PC supply chain ultimately thought about components for $500-1000 boxes to go on your desk - the smartphone supply chain thinks about much smaller boxes that average $200 and go down to $30 or $40 and run on batteries. So you get smaller, cheaper, low power components, and you get all sorts of new types of sensors that a PC could never have used. These components are enabling everything from drones to wearables to connected home devices, 'internet of things', smart TVs and connected cars. And satellites.
When I first started going to tech trade shows, 15 years ago, people were doing 'demos' of '3G data services' using Compaq iPaqs with PCMCIA wifi cards clamped to the back, because outside Japan the actual phones still had four-line black-and-white screens. The vision was a long way ahead of the technology, and it took until the iPhone in 2007 for a product to come along that could deliver that vision. This is normally the way consumer tech works - people have a vision a long time before the technology is really there to implement it (remember the Newton, or the Nokia 7560).
Today it sometimes seems like things are the other way around. Want to make a connected door lock? Camera collar for your dog? Intelligent scale? Eye tracker? The electronic components all there, more or less off the shelf. The challenge is in the vision for what the product should be, what people would do with it and how you would take it to market. That is, hardware has lapped software, so to speak.
The immediate question is of course of specific use cases. Over the past decades we've slowly worked out more and more things you can do with an electric motor in the home, most of which would have seemed silly to our grandparents, and the same will happen with these new components. We'll work out things to do with sensors and intelligence that costs a dollar or a few cents. Our a16z investment Ringly is a fascinating angle on this - as the smartphone bundles up hardware categories we start creating new ones that unbundle it.
The challenges, I think, come from where the intelligence and the software will actually sit, and how much there needs to be. If much of the hardware tech is a commodity, then so is the hardware itself, unless you can find a way to create an addition layer of value with software and service and experience (and design, as with Ringly). It's the intelligence and the software that makes the difference between a commodity widget and something that has some real value for an entrepreneur (from outside or inside China). Not all device types really have network effects, and not all devices can be the hub of your connected home, at least not all at once. Some categories are going to be commodity widgets.
One way to get at this problem is to work out what the SKU might look like at retail, and who will sell it. Is it the Apple Store, Home Depot, Best Buy or your alarm company? Is it a point solution, or an add-on for an industry-standard platform, or an accessory for your Nest or iPhone, or Apple Television? Different channels will produce different results. And when we look at the thing itself, is this a thermostat/lock/garage door opener/pool heater/landing light with software, or software with a motor attached? Hence we can see entrants from all directions. There are software companies making hardware, and legacy hardware companies (locks companies, for example) adding software. There are channels (such as alarm companies) moving up the stack to hardware and then software, and software moving down the stack to hardware and channels. And there are strong players in each layer.
This is why I though the most interesting thing about Nest (besides the team) was to look at it not as a thermostat but a a route to market. Nest found a clear, easy-to-communicate use case with lots of scope to avoid commoditization through software, and also scope to be used as a hub - a trojan horse - to sell other things. Now your car will tell your Nest to turn on the heating when you're almost home. Conversely things like Apple's HomeKit and Quirky's Wink (Quirky is an a16z investment) aim for a flatter structure - your phone itself is the hub, and you can buy whatever point solution you want.
It seems to me that we'll see Venn diagrams - your door lock, garage door opener, thermostat and alarm might talk to each other, but not talk to your fridge. The fridge might talk to Instacart and nothing else, while the washing machine (and thermostat) might talk to the smart meter to pick the best power cycle. And your phone will talk to them over wifi and IP, and one app or many. That is, for some things the connecting protocol will be something closed and sophisticated, for some IP, and for some it'll be AC power.