The drumbeat of Internet of Things (IoT) marketing remains steady. Every company out there seems to be looking into the excitement. At this point, I am not sure who the intended recipient of the message could be? Are there customers out there looking to buy IoT? Is there some huge IoT order coming down the pipeline?

Of course not. The whole technology industry is still trying to figure out what IoT means.

In my big survey of IoT systems and follow-on pieces (here and here), I laid out my thesis which is that there is no such thing as “IoT”. There will not be a single IoT standard, nor will there be a single, common way to wire up all the machines in the world. Too much of the work involved in that is retro-fitting. If we were building a colony on Mars, we could probably agree on some common ways to connect everything. But here on Earth, we have to deal with 100 years of mechanization, and some very large vested interests, that are all going to want their own way of building IoT systems.

I do believe that we will eventually connect up all the machines, but I just think that the ‘Internet’ that eventually emerges will be much more fragmented than people realize (or admit).

Mention IoT and it conjures up images of wearables and smart watches and connected home systems. I certainly agree that will be a real category and possibly a very large one. However, as things stand now that industry will probably end up looking a lot like the rest of the consumer electronics industry. Which is to say it will be dominated by a small number of very large players – Google, Apple, Amazon, Alibaba, Xiaomi, etc. Under that scenario, most of the economics will accrue to those companies. Supplying components to those companies will not be very profitable despite the volumes. And all those companies will capture all the interesting bits of software and data themselves.

It is important to remember that consumer will be just one of the industry verticals for IoT. I counted 23 verticals that could be big IoT markets, and there are probably a few more that I have not counted.

Each of these verticals could be multi-billion opportunities. The important part will be learning to how to approach each one. The IoT opportunity will be fragmented, almost to the point of custom implementations. Today, all IoT systems are just that – custom-build. The real opportunity will be for the companies that find ways to build repeatable business practices. No one is going to find a single solution or go-to-market channel that will cover everything, but there will be tremendous value for companies that can find some common patterns in all of this.

I heard a great example recently, which I can use to clarify some of this. There is a company out there looking to tap into the Chicken Farming Market. According to the National Chicken Council there are 29,000 chicken farms in the US that ship $50 billion in chicken, or 9 billion chickens every year. So big numbers. This company has developed a chicken tracking system that can measure certain attributes of chicken behavior. They are apparently working on a wearable system that would include a few sensors. They have developed algorithms that predict certain important chicken behaviors based on a small sensor harness they want to install on each chicken. And remember, IoT is about actuators not just sensors. Under certain circumstances, chicken farmers need to euthanize chickens, and this company was considering ways to add that functionality to their system. No one is going to (profitably) take on Apple for control of the home IoT market, but I will gladly take a share of the chicken euthanizing market if I can charge $2 per chicken across nine billion chickens.

For component vendors, tapping into these verticals is going to prove incredibly challenging. Few semiconductor companies have sufficient software resources. And I doubt that even the largest have meaningful relationships with the companies who are going to be the gatekeepers for industrial IoT systems. The kinds of companies (and farms) that will be buying IoT systems in the future do not want to build on their own. They will not work with Taiwanese hardware ODMs or muck about with reference designs. They will turn to trusted IT partners, the companies who already supply their other technology needs. For agriculture this will likely mean companies like Trimble and John Deere. For nuclear power plants it will be Raytheon and General Dynamics. For auto plants it will be IBM and Accenture.

Let me ask a hypothetical question. Take the two biggest chip companies in the US – Intel and Qualcomm. What do you think their relationship with the two biggest systems integrators, IBM and Accenture, is? Which way do the dollars flow? I am willing to bet that even Intel pays more to Accenture than Accenture does to Intel. To the extent they have joint R&D projects, it is likely through third parties. So if even those massive companies are far-removed from the future IoT gatekeepers, how can small micro-controller and connectivity companies expect to reach them?

The real opportunity in IoT will be finding common patterns across all the industry sectors. Do not expect to find a lot of them. The work of tying up diverse systems is messy and rarely profitable.