The sentence I buried at the bottom of the page

TLDR — Years ago I wrote a confident paper forecasting where blockchain would go. It turned out to be wrong, and looking back at it now, it was wrong in a way that is perhaps more useful than if it had been right. I backed a specific technology as the answer and buried the durable point (that customers want verifiable trust and won't pay extra for it) at the very end of the article. This is what a dead forecast teaches you about reading a technology at the top of its cycle: separate the durable principle from the perishable example, and never confuse the tool for the need.

A confident forecast I wrote at the top of the hype turned out wrong. How it was wrong is the useful part.

Some years ago, at the top of the blockchain excitement, I wrote a discussion paper forecasting where the whole thing would land. I was confident. The paper had a single, tidy thesis: blockchain would mature into a commodity trust utility, the way cloud computing had, once regulators and policy caught up. It even had the requisite confident line. "The blockchain is here to stay".

I was wrong. Not wrong that blockchain would persist, it has, but wrong about where the value would go, which is the part that actually mattered. And because I have spent this blog telling other people that the method is rarely the thing that fails them, it seems only fair to put one of my own forecasts on the table and perform the autopsy in public.

The useful news is that a forecast can be wrong in a lazy way or wrong in an instructive way. I take the view, perhaps to protect my ego, that this one was instructive, and the lesson has almost nothing to do with blockchain.

Durable and perishable

Every forecast is a mix of two things, and the trick is telling them apart.

There is the durable part: the underlying human or commercial need that will still be true in ten years. And there is the perishable part: the specific technology, product or company you bet on to serve it. Get those two confused and you will write exactly the type of paper that I wrote, one that spends too much energy on the perishable and underplays the durable and then buries it in the last paragraph, which made it seem like an afterthought that it was never meant to be.

My paper backed the perishable. It was almost entirely about permissioned enterprise blockchain for supply chains, a named list of flagship projects, and one confident analogy holding the roof up. The durable point, the one that has aged perfectly, was a single sentence near the end: customers will increasingly demand trust and traceability, and they will not pay a premium for it. That sentence needed no blockchain to be true. It needed no particular technology at all. It was the actual insight, and I let the confident infrastructure story sit on top of it and drown it out.

That is the pattern worth focusing on. I backed the method and forgot the need. The method died. The need was the durable thing all along.

Watch the flagships die

You can mark a forecast fairly by tracking the specific things it named. Mine named a lot, so here is the scoreboard.

TradeLens, the Maersk and IBM supply-chain platform, was discontinued, offline by early 2023. Tellingly, the stated reason was not the technology. It was commercial viability and the failure to get an industry to collaborate. Everledger, the diamond-provenance darling, entered administration in 2023 despite having raised more than fifty million dollars and holding serious backing. we.trade, the trade-finance equivalent, closed in 2022 citing a lack of revenue. IBM Food Trust, with Walmart, survives, but in a far more modest register than the fanfare promised, and IBM's blockchain ambitions shrank around it. Dubai's pledge to put everything on blockchain by 2020 passed quietly, the way all of these PR driven showboat targets tend to, see Net Zero targets for the latest incarnation of these!

Notice the cause of death in nearly every case. Not the sophistication of the technology. The commercial reality around it. Whether anyone actually needed it, would pay for it, and could be bothered to co-operate. The method worked fine. The mindset and the momentum around it never showed up.

And the money? It did not leave. It relocated. The institutional energy that was supposed to flow into permissioned supply-chain ledgers went somewhere the paper never mentioned: tokenisation, stablecoins and real-world assets. Tokenised real-world assets passed roughly twenty-four billion dollars by early 2026. Stablecoins processed enough volume in 2024 to overtake Visa on the headline number, though a large share of that was automated trading rather than people buying things, which is its own small lesson in reading a triumphant statistic. What unlocked it was not the technology finally maturing. It was regulatory clarity, the US stablecoin law of 2025 and the EU's rules before it. Policy, not cryptography, moved the needle. The paper had at least got that mechanism right, even as it pointed the arrow at the wrong target.

There is one name from the original that ended up on the right side of the shift, and it is instructive precisely because it did not pivot. Quant held the same unfashionable thesis in 2019 that it holds now: interoperability, the plumbing that lets different networks and rails talk to each other. It did not chase the supply-chain story. It stayed being infrastructure, and the category eventually moved towards it. Its public work is now with the regulated money world rather than the provenance world: a pioneer partner on the European Central Bank's digital euro project, technology vendor work with the Bank of England and the BIS, the UK's regulated-liability experiments. The point here is not the company. It is the direction. The value migrated to the connective, settlement layer, not the supply-chain-provenance layer my paper spent way too many of its words on.

Why the forecast was off target

Here is the uncomfortable part, and it is not about intelligence. The failure was structural, and you could have spotted it in the shape of the paper before a single prediction came due.

It was a monoculture. One thesis, one analogy, no counter-case. I never wrote down the conditions under which I would be proved wrong, so there was no way to mark my own homework later. And I never asked the one question that turned out to decide every single outcome on that scoreboard: do you actually need a blockchain here, rather than a trusted shared database and some agreements between grown-ups? That question, asked honestly, would have killed half the flagships on paper before they burned real money.

None of those failings are technical. They are failures of mindset. A confident single story is comfortable precisely because it saves you from holding two possibilities at once. It reads as conviction. It is usually just tidiness.

The fix

So here is what I should have done differently and what I would offer to you the next time you encounter something that is at the top of its cycle and everyone is very sure.

Separate the durable principle from the perishable example, on purpose, in writing. Say which sentence in your argument you would still stand behind if every named product died, and put that sentence at the top, not the bottom. Then write down what would prove you wrong, and date it, so the future version of you can actually check it. And keep the deflating question close: do we even need this specific thing, or do we need the outcome it promises?

For the customer, that outcome is the whole game, and it has not changed in a decade. People want to be able to check that a claim is true. Provenance, trust, traceability, verifiable and on demand. They want it, and they will not pay a premium for it, and they do not care in the slightest whether it arrives on a blockchain, a boring database, a QR code or an old-fashioned certification body. Decouple the outcome from its method of delivery and you stop betting the farm on any one of them.

Or, as I put it in the very last paragraph of that paper, in the one line I should have opened with:

"Customers will increasingly demand trust and traceability, and they will not pay a premium for it."

That is the sentence I buried at the bottom of the page. It is the only part of the forecast that was never in danger. And the lesson of getting the rest so confidently wrong is worth more to me now than being quietly right would have been.