What "backed" has to mean

Proof of what actually stands behind a token, onchain and off.

You are holding a claim, not an asset

If you hold a tokenised treasury, you do not own a treasury. You own a claim on one, and a claim is worth exactly as much as the evidence behind it.

So go and look for the evidence. This is where it gets interesting, because the asset is almost never where you think it is.

Where the asset actually lives

Not on a chain. The token is on a chain. The asset is somewhere else entirely, and usually in several somewhere-elses at once.

A typical issuer is spread across several legal entities. Under those entities sit bank accounts, brokerage accounts, sometimes privately held positions that do not trade anywhere. Money moves between them. And at the end of all that, a token mints or burns on-chain.

The mint is the last leg of a long journey, and it is the only leg you can see.

There is no node to ask

On-chain, you can check a fund’s books because there are two independent sources. You build the book from every transaction, you ask a node what the balance is, and you compare. The node has never heard of your book, which is exactly what makes the agreement worth something. That argument is set out in reconciling the book to the chain.

Now try it on a treasury bill sitting in a brokerage account.

There is nothing to query. No node, no contract, no public state. What exists is a statement, and a statement is a document a counterparty wrote about itself.

And here is the trap, because a statement does carry a closing balance, and it looks for all the world like the second source you were after. It isn’t. The balance and the transactions on that statement came from the same institution, out of the same system, at the same time. They cannot independently contradict each other, because they were never independent. Both ends of your rope are tied to the same post.

So you cannot reconcile an RWA the way you reconcile a DeFi position. The tool that would do it does not exist, and it is not coming, because you cannot query a filing cabinet.

What is left, and it is harder

What you have instead is a flow. And flows have a property that balances do not: they touch more than one party.

Cash leaves an operating account at one bank. It arrives at a brokerage. It buys a bill. The bill settles. A subscription is recorded against an entity. A token mints. Every one of those steps appears in somebody else’s records too, and the somebody else is not the person telling you the number.

That is where the independence comes from. Not from a balance you can ask a machine for, but from the fact that a transfer out of the bank has to turn up as a transfer into the broker, on the same day, for the same amount, or something is wrong. Nobody gets to mark their own homework, because their homework is stapled to three other people’s.

Which means the check is not does the balance agree. The check is:

Does the flow hold together, all the way through, across statements written by parties who do not talk to each other?

That is a genuinely harder problem than the on-chain one, and it is the one that decides whether the NAV is a number or a guess wearing a number.

Where it actually breaks

Not in the custody account. That fails rarely, and when it does it fails loudly, in a courtroom.

It breaks in the joins. Cash sits in transit over a weekend and gets counted twice, or not at all. An intercompany transfer between two of the issuer’s own entities looks like a subscription from the outside. A brokerage reports on a trade date and a bank reports on a settlement date, and for two days the same dollars are in two places. A redemption burns tokens on-chain on Friday and the wire leaves on Tuesday.

None of that is exotic and none of it is fraud. It is what happens when one asset’s story is told by four institutions who have never met, in four formats, on four calendars. And every one of those seams is a place a number can go missing without anybody noticing, which is why the NAV provider has to carry all of it. Every entity, every account, every venue, and the on-chain leg too, in one book that has to stay logically consistent from end to end.

Handle nine of the ten and you have not built a NAV. You have built a number that is right nine tenths of the time, which is a different product and a much worse one.

The bit we are not going to pretend

We do not attest. We are not your auditor, we do not go and look inside the custody account, and nothing here is a claim about whether the bills exist. No vendor can tell you that, and one who implies otherwise is worth backing away from slowly.

What can be built is the book: every transaction, across every entity and every venue, joined so the flow of a dollar can be followed from a bank, through a broker, into an asset, and out to a token, without the thread going slack anywhere in the middle. Then when somebody does sign the attestation, they are signing something that was checked rather than something that was merely added up.

Ask your issuer to walk you through the flow of one subscription, end to end, and to name which statement proves each hop. It is a short conversation, and you will learn most of what you need from how long it takes them to answer.