Should Everything Go On-Chain?

Money has started to move without waiting for people. Autonomous and AI-agent systems now originate, evaluate and route capital continuously, and the records meant to account for that movement still arrive weeks after the fact, assembled by hand. Anyone responsible for that capital is operating in the gap between what happened and what they can see.

Tokenization is the loudest answer on offer, and it has moved from conference-stage provocation to institutional roadmap. BlackRock runs a multibillion-dollar tokenized Treasury fund. The DTCC, which custodies more than $114 trillion in securities, ran its first live tokenized trades this month ahead of a full launch in October. The value of tokenized real-world assets on public blockchains passed $30 billion this year, up from roughly $5 billion at the start of 2025.

The reflexive move is to pick a side. Either everything belongs on-chain, and the holdouts are managing decline, or the whole exercise is a solution in search of a problem. Both camps are arguing about the wrong variable.

Put an asset on a ledger, and you've changed where the data sits. You haven't changed whether anyone can trust what the data says. That distinction determines whether tokenization addresses the accountability gap or merely relocates it.

The maximalist case is winning the room

Concede the momentum in full. McKinsey's base case puts tokenized financial assets near $2 trillion by 2030. BCG's early estimate ran to $16 trillion before a 2025 revision brought it closer to $9 trillion. The forecasts disagree on magnitude and agree on direction. Six separate asset categories, including private credit, commodities and U.S. Treasuries, have each crossed $1 billion in on-chain value. The GENIUS Act gave issuers a federal perimeter to design against, and U.S. banking regulators have clarified that a tokenized security that carries the same legal rights generally receives the same capital treatment as its conventional form.

When the largest asset managers, the largest securities custodian and federal regulators all move in the same direction within eighteen months, the direction of travel is settled. The forecasts still gloss over the harder part.

A faster ledger can still be a mess

Much of the tokenized value on-chain today reflects issuance rather than activity. Assets get minted, allocated in large institutional batches, then mostly sit. The token exists. The market around it stays thin.

The thin market is a symptom of a deeper gap. A token is a claim about something in the world: a loan, a fund share, a warehouse of equipment. The chain can prove the token moved. It can't prove the claim behind the token is true, current, or complete. If nothing connects the token to the state of the underlying asset and the capital flowing through it, tokenization has reproduced the old opacity at higher speed.

A permanent record of an unverifiable claim is still an unverifiable claim. It's just harder to delete.

Private markets already know this failure mode. Capital deploys, and the account of what it's doing arrives quarterly, assembled by hand, weeks after the fact. Moving that structure on-chain without fixing the record gives the same blind spots a faster venue.

The record is the product

Programmable rails change one specific thing. They make it possible for the record of what's happening to capital to update in the same motion as the capital itself. Every transfer and every repayment can write its own entry the moment it settles.

Everything else on the tokenization brochure is real but secondary. Fractional ownership, around-the-clock settlement, custody efficiency and composability are all real. None of it is load-bearing. An asset earns its place on-chain when moving it there produces a record participants can act on: a lender who can see collateral status without waiting for a report, an auditor who can trace a dollar without a data room.

That's the sorting principle the debate lacks. The question worth asking about any asset is whether a live, verifiable record of its movement changes what someone can do. Where the answer is yes, tokenization is infrastructure. Where the answer is no, it's decoration.

Capital now moves at machine speed

The velocity problem deserves its own accounting. Agentic workflows compress decision cycles that ran weeks into hours. A treasury operation that once rebalanced monthly can rebalance continuously, with software executing against policy around the clock.

Quarterly reporting was a tolerable convention when capital moved quarterly. When capital moves at machine speed, a record that lags by ninety days becomes a standing liability, and the gap between the movement and its accounting widens with every cycle.

This is the honest case for on-chain records. The chain's contribution is a record produced at the speed of the movement itself rather than reconstructed after it.

What a working record looks like

W3.io builds against exactly this problem. W3 workflows are built so that every movement of capital writes its own record, a Programmable Capital Record, the moment it settles. The movement produces the record. Nobody assembles it later. Enterprises can compose and deploy these workflows in a day, not months.

The division of labor stays clean: W3 produces the record, the underwriter still underwrites and the compliance perimeter holds. What changes is what the person responsible for the capital can see: a verifiable account of where every dollar sits and how it got there, current as of the last movement, not the last reporting cycle. That visibility is what oversight looks like once the capital is automated.

The clearest proof that the record is the real product is that it has no yield in it at all. Consider a donated dollar. There's no investor and no return anywhere in the flow, and the accountability question only gets purer. The donor wants to know the dollar arrived where it was promised, and the organization owes an answer better than an annual report. Any capital that moves on someone's behalf owes them a record of the movement.

The honest answer to the on-chain question

Should everything go on-chain? No. Plenty of assets sit idle and pass through few hands. A live record of them changes nothing for anyone, and migrating them is cost without a customer.

What belongs on-chain is the capital where the record is the point: capital that moves fast and increasingly moves itself. That's where a real-time, verifiable record stops being a nice property and becomes the thing that makes the movement governable.

This is the category W3 calls autonomous finance. W3 is the operating system for it, the layer where agent-driven capital workflows run and every movement writes its own verifiable record.

Tokenization matters where it delivers that record. Everything else is migration for its own sake. What the market is building toward is a trust layer for money that moves itself.