Over a decade ago, I got into the blockchain business because I wanted to fix the Internet of Things.

A decade later, both businesses are thriving and both still have big problems with their business models. It may well be that, like the original internet, we can never really escape bad business models once they take root. I am, for one, skeptical that we’ll ever be free from the our-service-is-free-because-you-are-the-product model of social media, for example.

Still, I have some hope that, as the Internet of Things (IoT) is still relatively nascent, we might be able to use blockchain – specifically in this case blockchain-enabled Decentralized Physical Infrastructure (DePIn) to fix it. At heart, the problem with the Internet of Things is the business model: companies need a constant stream of revenue to maintain their products. Consumers, very understandably, don’t think it’s reasonable to pay a subscription to maintain, say, the software on their door knobs or fridge. The result is a great deal that usually comes with a nasty hangover: products that are free from subscription fees that one day get discontinued because the company selling them wants to stop maintaining the product.

Blockchain offers an alternative, combining open-source technology with decentralized systems, allowing us to build Internet of Things networks that manage themselves and can operate more sustainably.

At the heart of the problem is the mismatch between the life of products and the life of the product-line being sold by the business. We might ditch our smartphones and PCs every three to five years, but generally speaking, lightbulbs, doorknobs, refrigerators and other home devices are expected to stay in place for longer. If you need cloud infrastructure to manage these devices, you have a recurring cost that goes on for a decade or two after you may have sold the product. Combine that with software maintenance costs and it’s easy to see how you can consume all your margin over time.

The result is that, with depressing regularity, companies decide to “switch off” online services for products they used to sell. The result often turns a device you integrated into your life into a brick. Alternatively, the vendor offering the service that was “included” in the purchase prices starts charging.

A few years ago, I was suddenly hit with a $90 annual fee to keep my smart door locks running. I suppose that’s better than obsoleting them, but I was so incensed I went out and bought new locks and installed them as replacements. It probably cost me about eight years of service to replace the locks, but my decision was driven by spite, not rational analysis.

Despite some very frustrating experiences, the IoT industry has made some very good progress in the last few years. Devices that integrate with the HomeKit standard and those using the new Matter controls and Thread radios are built from the ground up to run without internet connections. This means their basic functions do not require cloud infrastructure and the burden of maintenance doesn’t fall on a single enterprise.

However, if we want truly smart homes and connected experiences, we will need internet connectivity and cloud computing infrastructure. And, for that, we need decentralized cloud infrastructure as well.

Using blockchains, devices with spare computing capacity and network connectivity can run more complex network level applications.

Want to manage your home power consumption based on the state of the grid? Sell power at the best times or use generative AI systems for a conversational interface? All those things consume a lot of computing power and bandwidth and, if we want a sustainable business and price model, then we must be able to do that without needing lots of new data centers.

The good news is that the smart home devices have become absurdly smart. This isn’t because we really need the intelligence of a smartphone in our lightbulbs. It’s because it turns out that it’s cheaper to put an entire smartphone-level brain in a lightbulb than it is to make a highly customized lightbulb specific smart-chip. Chipmaking is a volume business and building a standard overly-smart chip and using software to make it do things like handle a light or manage a refrigerator is cheaper and more scalable than customizing each device.

The upshot is a lot of idle connected computing power that can be put to work in building a blockchain-linked decentralized cloud computing infrastructure. Your smart home and car can “pay its own way” when it comes to compute power, selling excess capacity when you’re not using it and using more from others when needed. The result should be a sustainable network infrastructure that doesn’t need constant injections of capital from the original product sellers to keep working. If the cloud is, as the T-shirt says, just someone else’s computer – maybe it could be your neighbor's refrigerator?

There are many ways of building decentralized computing infrastructure. But there’s a reason that, when I started down this path more than a decade ago, I chose blockchain and not some other technology: payments and contracts.

It’s very simple: if we want a system where smart devices transact with each other to provide computing services, we need accounts, ledgers and agreements. Blockchains come with those baked in.

For more than a decade, I’ve been hoping to see cloud computing, blockchain and the Internet of Things get together. We might, at long last, be approaching that time.

Note: The views expressed in this column are those of the author and do not necessarily reflect those of CoinDesk, Inc. or its owners and affiliates.