Our industry has been working on something massively consequential for decades. Linking the physical world of places and things to the internet was always a huge idea. Add the power of AI applied to those things and the idea, and the opportunity, gets bigger still, especially if you believe that the full value of connecting physical things to the cloud has yet to be realized.
Yet despite shipping billions of RFID tags and wireless trackers, the world is still operating in the dark 99.9% of the time. Most things that are made aren’t connected to the cloud in any way that is autonomous. Scans are snapshots, not connections.
We will know we are succeeding when the links between digital and physical are counted in trillions, not billions, and, more importantly, when the cost of goods and the environmental impact of our supply chains are slashed. At that point we get to move beyond the basics of Maslow’s hierarchy to the joyful stuff: AI that goes beyond writing emails to giving us better tasting fruit, more effective medicines, and the kind of storytelling for everyday things that today we only get at a farmers’ market, a Louis Vuitton store or a top restaurant. Can we agree we are not there yet?
So why is IoT the awkward stepchild of AI, when what we are really working on is a version of physical AI that could transform the economy?
Billions, Finally
Let’s give credit where it’s due. The low-cost Bluetooth tag segment is at last experiencing the accelerated growth that has been a long time coming. Based on planned deployments from the likes of Walmart, Amazon, IFCO, the US Postal Service and McDonald’s, Bluetooth will cross the billion-tag boundary by the end of next year. Not bad for a technology originally designed to replace the wires between keyboards and PCs. Optimists like me expect annual production to eclipse the RFID industry’s 35 to 70 billion units a year before too long.
But before we pat ourselves on the back, this is a long way short of the addressable market that ABI Research and the Bluetooth SIG put at over a trillion units. The gap between a billion and a trillion is not a matter of cheaper tags. We have spent decades driving down cost, and it hasn’t happened yet. Something else is missing.
What Ambient Actually Means
“Ambient” in ambient IoT usually gets associated with harvesting energy from the light and radio waves around us. That matters, but the quality that makes ambient IoT transformative is ubiquity, and ubiquity is a function of three costs: tags, infrastructure and labor.
Labor is the one we underestimate. Scanning a tag with a handheld reader doesn’t seem like much effort. But the moment we ask a human to do it, we invite errors, training, compliance issues, and a multiplier that takes a few seconds and applies them to every product in every location, every time. A few seconds compound to millions of dollars when applied to every item in every store. IoT becomes ambient when we connect all the things around us using all the devices already around us: phones, access points, smart devices, doorbells, cars. At that point we are on the brink of an internet of trillions of things.
Why Apps Are Easy and the Physical World is Hard
We have all learned that the physical world is hard. I’d argue the reason is more specific than that: reading is still largely manual, and the readers that were supposed to be everywhere are not.
Compare that with ChatGPT, which scaled to 100 million users in a couple of months. Every one of those users already had a connected device that could run it, paid for, in their pocket. No ROI calculation, no meeting to decide whether to buy a reader. The network was already there. The lack of networks is what makes apps easy and the physical world hard.
That isn’t how IoT was supposed to go. When it was first conceived, RFID was going to be embedded in appliances and phones to link consumers, supply chains and manufacturers. When the device makers shook their heads at adding yet another radio to all the radios already in their products, we retreated from that vision and settled for a much more limited IoT in the middle of the supply chain. Today the radios exist all the way from the farm to the fork. We just aren’t using them.
End to End, at Last
Beyond low cost, the thing that closes the billions-to-trillions gap is something the RAIN RFID market has always lacked: end-to-end visibility, from the maker of a product to the person using it. Or, to use the words investors most covet, network effects, the point when the rock rolling downhill becomes unstoppable and everyone gets pulled along.
We have dreamt of this for a long time. Those “register your product now” cards we all ignore were a testament to it. The moment a CPG manufacturer can sense the consumption of its products in consumers’ homes in real time, and talk to them directly, supply chains change drastically: a big reduction in cost of goods, far better customer retention and acquisition, and product-as-a-service models that aren’t possible today.
Make it concrete. McCormick, the herbs and spices company, could bid to be the preferred provider to your family’s spice rack, so you never have duplicate pots of peppercorns, never run out, never have expired product, and always have a personalized palette of new flavors to make you the hero of the potluck. On the environmental side, a direct-to-consumer supply chain without the waste and redundancy would be one of the biggest wins in the fight against climate change we’ve ever seen, and one that grows profits and improves the experience of consumers at the same time.
The Networks Are Coming. That’s The Problem
The engine for all this is the rise of Bluetooth networks. Until now the only Bluetooth IoT networks of note have been Find My from Apple, Tile and the Android OEMs: closed to developers building enterprise visibility systems, and populated by devices costing an order of magnitude more than the sub-dollar tags driving the billion-unit numbers. That is changing. Networks that could drive explosive growth are now being built by companies like Nodle, Hubble Network (opening up the once-closed Tile network), Wiliot and InPlay.
Each has earned the right to be proud of what it’s building, and each is seeing real momentum. But if what they are building are standalone networks, we will have reconstructed the old fracture one level up: an IoT that can see from farm to fork, but only one carrier at a time. The cellular industry made exactly that mistake in the 1G and 2G era, and it took a common standard to fix it. That is the subject of my next piece.
Steve Statler is CEO and co-founder of ambientChat.AI, an AI context management company focused on connecting people with places and the makers of the products they use. He is author of Beacon Technologies: The Hitchhiker’s Guide to the Beacosystem (Apress/Springer), host of the Mr. Beacon podcast, and an advocate for open, interoperable IoT ecosystems. Prior to founding ambientChat.AI, Steve spent more than seven years at Wiliot as the company’s first Senior Vice President, leading Product Management, Business Development, Sales, and Marketing, and later serving as Chief Marketing Officer. He is a contributor to the LIVE IoT initiative, an open standard for secure Bluetooth IoT interoperability.