Blockchain's Problem Isn't Tech. It's That Nobody Wants Boring

I read another blockchain market-size report this morning. This one says the market hits some absurd number by 2036. I've seen versions of this report since 2018. Different year, same fantasy, same total absence of anyone actually using the thing at scale. Meanwhile, buried under listicles about "10 cryptos under 10 cents," something real is happening in supply chain management, and almost nobody selling blockchain noticed it happened without them.

Here's the take: blockchain never had a technology problem. It has a boring problem. Every founder who raised money on "decentralizing finance" wanted to build the next payment rail, the next exchange, the next thing that moves money and gets you on a stage in Miami. Nobody wanted to build the unglamorous thing that actually needed solving: proving who touched what, when, in a chain of custody nobody trusts. That job was always there. It just wasn't sexy enough for a whitepaper.

The Real Opportunity Was Never Money. It Was Trust.

Money movement already has rails. SWIFT is slow and annoying but it works. Visa and Mastercard clear trillions without asking anyone to understand a merkle tree. The actual gap in the world, the one worth billions in efficiency, was never "how do we move value," it was "how do I know this batch of vaccines was stored at the right temperature for all 14 hours between the factory and the hospital," or "how do I prove this cobalt didn't come from a mine using child labor," or "did anyone tamper with this shipment between Rotterdam and Gothenburg." That's not a finance problem. That's a verification problem. And verification is boring. Nobody wants to build a company whose entire pitch is "we watch things happen and write it down accurately." But that boring job is worth more than every DeFi protocol combined, because every company on earth that ships a physical product needs it.

Why Token Projects Never Delivered

Look at any enterprise blockchain pilot from 2019 to 2023. IBM Food Trust, Walmart's leafy greens tracking, various shipping consortiums. Most quietly died or got shelved. Not because the cryptography failed. Because the projects tried to force a token economy onto a problem that didn't need one. You don't need a native coin to log a temperature reading. You don't need gas fees to confirm a container left a port. The blockchain layer got weighted down with financial incentive design that solved a problem nobody in the supply chain actually had. The real bottleneck was always: who verifies the data going in, and how fast can that verification happen without a human doing manual reconciliation for three weeks.

AI Agents Just Did What The Tokens Promised

This week's Unite.AI roundup on supply chain tools is the tell. It's not about tokens or distributed ledgers. It's about AI agents that read invoices, cross-reference shipping manifests, flag anomalies in customs data, and reconcile discrepancies between what a supplier claims and what a sensor recorded, all without a person touching it. That's verification. That's the exact job blockchain projects said they'd solve. The agents just do it without needing a consensus mechanism or a token to bootstrap network effects.

CIFF 2026 is building entire conference tracks around e-SCM and what they're calling TechCreate infrastructure, which is a fancy label for "software that makes supply chains actually legible." PadUp Ventures and Unicity Labs are quietly building what they call agentic commerce infrastructure. No whitepaper. No token launch. No Discord full of people asking when the coin moons. Just infrastructure that does the verification work and gets out of the way. That's the tell that this is real. When you stop needing a marketing narrative to explain what the product does, the product is probably good.

Where Blockchain Actually Fits Now

Blockchain isn't dead in this picture, it's demoted, correctly, to what it should have been the whole time: a trust layer nobody sees. If an AI agent verifies that a shipment container's seal wasn't broken, you still want an immutable record of that verification that no single party can quietly edit later. That's a real, boring, valuable job for a ledger. Not the star of the pitch deck. The plumbing behind it. This is the same shift crypto had to make with stablecoins: the interesting part was never the coin, it was the settlement rail underneath it that nobody talks about at dinner parties.

The Nordic View: We're Better at Boring Than Silicon Valley Is

Here's where being from Jönköping instead of San Francisco actually helps. Swedish industry doesn't get excited about narratives. We get excited about whether the thing works, ships on time, and doesn't need a hype cycle to justify the invoice. Sweden's manufacturing and logistics base, Volvo, Sandvik, Ericsson's supply chain, half of Gnosjö's industrial network sitting an hour from my office, has spent decades optimizing for exactly the kind of unglamorous verification problem that AI-native SCM tools now solve well.

Compare that to the US market, where the noise-to-substance ratio in enterprise tech announcements is still dominated by companies raising Series B rounds on "blockchain-powered" everything, years after the market decided that phrase means "please don't ask what problem this solves." EY is out there this week publicly asking how telcos even do B2B growth in the current environment, which tells you something: even the advisory firms are struggling to make the old sales motion work. The old motion was: sell transformation, sell the platform, sell the vision. The new motion has to be: sell the boring outcome. Fewer errors. Faster reconciliation. Provable chain of custody. That's a sale Swedish industrial buyers actually respond to, because it's the same language they use internally. Sweden and the broader Nordic market are, frankly, ahead here by temperament, not by policy. Our industrial base was never going to buy the blockchain hype in the first place. Now that AI agents are doing the actual verification work without needing anyone to believe in a coin, it's an easier sell into Swedish manufacturing and logistics than it ever was into finance.

Where Sweden and the EU Are Behind

The regulatory side is a different story. The EU's approach to AI liability and data provenance in supply chains is still catching up to what these agentic tools are already doing in production. If an AI agent flags a shipment as compliant and it turns out to be wrong, who's liable? The tool vendor? The company deploying it? Right now that's genuinely unclear across most EU jurisdictions, and Sweden hasn't carved out anything sharper than the general EU AI Act framework. That's a gap. Companies deploying these tools today are moving faster than the legal frameworks meant to cover them, which is fine until it isn't.

Where This Goes: 2 to 5 Years Out

The path here is not subtle. Supply chain verification is one of the clearest near-term applications of increasingly autonomous AI agents, because the task is bounded, the data is structured enough to work with, and the cost of errors is measurable in real money, not vibes. As we move toward more general, more autonomous agent systems, capable of the kind of multi-step reasoning and tool use that projects like OpenHands and the newer agent harness projects are pushing toward, the boring verification layer becomes the training ground. It's low-stakes enough to deploy widely and high-value enough to fund the next round of capability. Five years out, I expect most large manufacturers and logistics operators to run agent networks that handle 80 percent of supply chain verification without a human in the loop, with a blockchain or blockchain-like ledger underneath as the tamper-proof record, invisible, unmarketed, doing exactly the job it should have been doing since 2018. The token layer, if it exists at all, will be for machine-to-machine micropayments between autonomous agents settling small transactions faster than any card network can, not for retail speculation. That's the real crypto use case hiding in plain sight, and it has nothing to do with anything currently trading under 10 cents.

What Builders Should Prepare For

If you're building in this space, stop pitching blockchain as the product. Pitch the verification outcome, and treat the ledger as infrastructure, not the headline. The market has made this extremely clear over eight years of failed enterprise pilots. Build the AI agent that does the actual work. Use blockchain where you need an immutable, multi-party trust layer, and nowhere else. That's it. That's the whole strategy.

What to Actually Look At

If you want to see where this is headed technically, a few things worth your time this week:

• OpenHands for how AI-driven development agents are architected when they need to take real actions, not just generate text. This is the pattern SCM verification tools are converging on.
• Graphify for turning messy documentation, schemas, and configs into a queryable knowledge graph without a vector store. This is exactly the kind of deterministic, explainable structure you want underneath any system that's making verification claims about a supply chain.
• ECC for anyone building agent harnesses that need memory, security, and research-first workflows. If you're prototyping an agent that verifies logistics data over time, this is the kind of infrastructure thinking you need underneath it.
• CIFF 2026's e-SCM and TechCreate sessions, worth watching even remotely, because it's one of the few conferences treating supply chain infrastructure as the main event instead of the panel nobody attends.

At HEIMLANDR we build exactly this kind of thing for clients who don't care about the buzzword, they care about whether the system catches the error before it costs them money. If you're evaluating whether your supply chain problem needs an AI agent, a proper blockchain trust layer, or honestly just better custom AI solutions wired into your existing systems, that's a conversation worth having before you buy either narrative.

The Boring Truth

Enterprise blockchain didn't fail because the tech was bad. It failed because everyone building on it wanted to be exciting, and the actual money was in being reliable. AI agents doing supply chain verification are winning right now because they never promised a revolution. They promised fewer errors and faster reconciliation, and then they delivered it. That's the whole lesson. Boring wins. It always was going to.

I'll take a system that quietly catches a bad shipment over a whitepaper that promises to reinvent finance every single time. Build the boring thing. Ship it. Let the narrative catch up later, if it ever bothers to.

Fredrik Brunnberg is the CEO of HEIMLANDR.IO, building AI and software solutions from Jönköping, Sweden. This is the daily HEIMLANDR briefing. If you found this valuable, share it with someone who builds things.