The Convergence of AI, Crypto, and IoT: The Machine Economy Is Quietly Assembling Itself
Tracy Thayne7/9/2026·5 min read
A founder I know runs a company that services commercial HVAC systems. Over coffee last month he told me, almost as an aside, that his newest revenue stream has no human buyer. Sensors in the units monitor their own wear, an agent evaluates service options against the maintenance contract, and the purchase order arrives signed before anyone at the customer has read an email. His best customer, he joked, is a rooftop.
He was joking. The rooftop was not. That transaction is what happens when three technologies we have spent a decade discussing separately finally click together: devices that sense, models that decide, and ledgers that let machines transact with something like trust.
The scale of this is easy to underestimate because each piece looks incremental on its own. There are now 21.1 billion connected IoT devices, growing 14% a year and headed for 39 billion by 2030 (IoT Analytics). Gartner projects that by 2028, 15 billion connected products will have the potential to behave as customers, and that by 2030 machine customers will directly influence or participate in $30 trillion worth of purchases (CIO Dive on Gartner, Forbes). In B2B specifically, Gartner expects AI agents to intermediate 90% of buying by 2028, routing more than $15 trillion in spend (DigitalCommerce360). Those are not three separate trends. They are one trend counted three ways.
Three Technologies, One Loop
The convergence is simpler than the buzzwords suggest. Think of it as a loop with three functions.
IoT is the senses. Billions of devices now know their own state: inventory levels, wear rates, temperatures, locations, usage patterns. AI is the judgment. Given that sensory stream, models can evaluate options, weigh tradeoffs, and decide what should happen next, not just report what already did. And crypto, or more precisely the cryptographic infrastructure underneath it, is the handshake. Machine-to-machine commerce needs programmable settlement, verifiable identity, and tamper-evident records of who agreed to what, because there is no human in the loop to catch a forged invoice or a misrepresented product spec.
Subscribe
Get notified by email when we publish a new post. No spam, unsubscribe anytime.
Each piece alone is a feature. A sensor without judgment is a dashboard. A model without settlement rails writes recommendations someone still has to execute. But close the loop and you get an economic actor: something that perceives, decides, and transacts on its own. That is not a smarter tool. That is a new participant in your market.
The Machine Customer Is a Buying Committee of One
For marketers, the most important word in Gartner's research is "customer."
We have already crossed the first threshold. I wrote in The AI Buyer Is Already Here about agents quietly filtering vendors on behalf of human buyers. The machine customer is the next step: the agent is not researching for a person, it is purchasing for a device, a fleet, or a system, under rules its owner set once and rarely revisits.
And a machine customer is a strange thing to market to. It does not feel urgency. It is immune to social proof, brand nostalgia, and a well-timed steak dinner. It evaluates structured claims: price, specification, availability, compliance, contract terms, verified performance history. Persuasion, the muscle marketing has spent a century building, matters less than legibility, the discipline of making your offering parseable, comparable, and verifiable by software. The uncomfortable question for 2026 planning is not "is our brand differentiated?" It is "can an agent, reading our product data cold, determine that we are the right answer?"
Trust Is the Bottleneck, and Provenance Is the Toll
Here is where the crypto layer stops being optional.
Human commerce runs on soft trust: reputation, relationships, the ability to sue someone if things go wrong. Machine commerce cannot. When an agent buys from an agent at machine speed, every claim in the transaction needs to be verifiable at machine speed too. Is this seller who it says it is. Is this performance data real. Was this data acquired with the right to use it. Cryptographic provenance, whether it lives on a public chain or in signed credentials, is how those questions get answered without a human notary in the middle.
This is the same argument I have been making about marketing data generally: owned, consented, traceable context is the asset, and everything borrowed is a liability with a delay on it. The machine economy raises those stakes again, because provenance stops being a compliance virtue and becomes a gate. An agent that cannot verify your claims does not discount them. It skips you.
What This Means for Your 2027 Plan
You do not need a blockchain initiative or an IoT division to act on this. You need three unglamorous workstreams.
First, make your company machine-readable. Structured product data, published specifications, clean pricing, schema markup, verifiable proof points. Second, get your context layer in order, because the same organized, current, owned business knowledge that makes your AI useful internally is what agents will evaluate externally. As I argued in Context Is the Whole Game, context is the operating layer, and the machine economy simply extends that layer past your walls. Third, start tracking non-human traffic and treating it as a segment: what agents ask about you, what they retrieve, where they abandon. That funnel exists today, and almost nobody is measuring it.
The convergence rewards the boring work. The companies that win machine customers will be the ones whose data was clean, structured, and provable before the agents showed up to read it.
The Takeaway
AI, crypto, and IoT were never three stories. They are the senses, judgment, and handshake of a single new thing: an economy where software perceives, decides, and transacts without waiting for us.
The numbers say this is arriving on a schedule measured in quarters, not decades. Twenty-one billion devices are already online, and the analysts are counting trillions in machine-influenced spend before the decade ends. When it lands, the marketing question will not be whether your message resonates. It will be whether your business is legible to the machines doing the buying, and whether your claims survive verification by something that cannot be charmed.
My friend's best customer really is a rooftop. Yours is coming. The only question is whether it will be able to read you when it arrives.
Tracy Thayne* is the founder of Expona, an AI-powered operational intelligence platform for B2B marketing. Read the Expona founder story or subscribe to the blog (below) for weekly insights on context, AI, and the operating model of the next decade.*
