Tokenisation is described as putting assets on a blockchain, which is impossible in the literal sense and misleading in every other. What goes on the chain is a record of a claim, and everything about whether the exercise works depends on what stands behind that claim rather than on the chain it is recorded on.
A blockchain can record with certainty that a particular address holds a particular token. That is the whole of its power, and it is genuinely useful. What it cannot do is make anybody deliver anything: if the token is supposed to represent a share of a property, a bond, or a quantity of gold, the connection between the record and the object exists entirely in a legal document, enforced by courts, in a jurisdiction.
This is the first thing to establish because almost every discussion of the subject skips it. The technology solves the transfer problem, which is real and was expensive, and it solves nothing about ownership. A token holder whose issuer refuses to honour the claim has a legal dispute, not a technical one, and the chain will faithfully record their continued holding of a worthless entry throughout.
The consequence is that tokenised assets carry all the counterparty risk of the structure behind them, plus the technical risks of the chain, and the marketing typically describes only the benefits of the second while implying the first has been eliminated. It has not been moved or reduced; it has been left exactly where it was, underneath a new record-keeping layer.
Putting an asset on a chain does not put the asset on a chain. It puts a claim there, and the claim is only as good as the entity that has to honour it.
Three things improve, and they are not trivial. Settlement becomes near-instant and final, replacing systems that take days and involve several intermediaries each taking a fee and each capable of failing. Trading hours stop existing, since a chain does not close. And ownership becomes divisible to arbitrary precision, which makes assets that could only be held in large units accessible in small ones.
For an asset whose main friction was settlement, these are substantial. Moving a bond position between counterparties has historically involved a chain of custodians, several days, and reconciliation work at every step, and a large part of the financial system's cost structure exists to manage exactly that. Compressing it to a single atomic transfer is a real economic improvement rather than a repackaging.
What none of them touch is the question of what the asset is worth and whether the holder can get it. That question is answered by the legal structure and by whoever holds the underlying, and it is the same question it was before the token existed.
The most successful application by a wide margin is tokenised money market instruments, and the reason it works is instructive. The underlying is homogeneous, its price is unambiguous and continuously observable, the custody arrangements are established and regulated, and redemption is a routine operation that the issuer performs constantly rather than an exceptional event.
That combination means the claim is easy to value, easy to verify, and easy to exercise, which are the three properties that make a token behave like the thing it represents. The token trades at the value of the underlying because arbitrage between the two is straightforward, and the arbitrage is straightforward because redemption actually functions.
It is worth noticing what carries the weight in that sentence. The technology contributes the settlement speed and the divisibility; everything that makes the instrument trustworthy comes from the regulated structure underneath, which existed before and would work without the token. The token is an improvement to distribution, which is a real thing to be, and it is not what makes the asset sound.
Tokenised real estate has been attempted many times and has produced very little that survived, and the failures share a shape. A building is not homogeneous, so its value is an estimate rather than an observation. Its legal ownership sits in a registry that the chain has no connection to, so the token represents a share in a company that owns the building rather than the building itself, adding an entity between the holder and the asset.
Redemption is the part that breaks. A holder cannot exchange their token for their fraction of a building, because a building cannot be divided. What they can do is sell the token to somebody else, which requires that somebody else exists, and the market for a fractional interest in one specific property in one specific city is thin by construction. The instrument becomes a claim that can only be exited through a market that does not reliably exist.
None of this is a technology problem and no chain will solve it. The difficulty is that the asset is illiquid, indivisible and hard to value, and tokenising an illiquid asset produces a tokenised illiquid asset. The divisibility that the token provides is nominal: you can hold a thousandth of it and you cannot sell a thousandth of it to anybody in particular.
A contract that needs to know the price of an asset that trades on a chain can read it directly. A contract holding a claim on something outside the chain cannot, and this is the same limitation that governs every oracle, arriving here in its most difficult version because the underlying frequently has no continuous market price at all.
For a money market instrument this is manageable, because the value is computable from published rates and a known maturity. For an asset whose value is an appraisal, the number written onto the chain is somebody's professional opinion, updated occasionally, and any mechanism that acts on it, including any lending against it as collateral, is acting on an opinion with a timestamp.
That does not make it useless and it does change what the record means. A position marked at an appraisal is marked at a number nobody has traded at, and the difference between that number and what the asset would fetch is invisible until somebody tries. Systems that lend against such collateral inherit that gap in full.
An asset with no continuous market price has no honest number to put on the chain. What gets written is an estimate, and everything downstream treats it as a fact.
Every question about a tokenised asset collapses into one: can the holder exchange the token for the underlying, at what cost, in what time, and under what conditions can that be suspended. If redemption works routinely, the token tracks the asset because arbitrage keeps it there. If it does not, the token trades at whatever the market for the token is willing to pay, which can be far from the value of what it represents.
The details that matter are unglamorous and documented. Who is eligible to redeem, since some structures allow it only for large holders or qualified investors and everybody else can only sell to another holder. What the minimum size is. How long the process takes. And what circumstances permit the issuer to pause it, which is the clause that decides what happens on the day it matters.
A holder who has not read those four things owns an instrument whose behaviour under stress they cannot predict. It is the same analysis as any redemption-dependent instrument, and the presence of a blockchain changes none of it.
The word liquid gets attached to tokenisation as though the two were connected, and they are not. Putting an asset on a chain makes it transferable, which is a technical property, and liquidity is an economic one that requires participants willing to be on the other side at a price. A token with no buyers is exactly as illiquid as the asset it represents, and it has acquired a settlement layer rather than a market.
There is a related effect that works against the promise. Many tokenised instruments carry transfer restrictions imposed by the legal structure, limiting who may hold them to verified addresses or to certain jurisdictions. Those restrictions are enforced by the contract, they are usually necessary for the structure to be lawful, and they mean the token cannot circulate freely in the way the vocabulary implies.
The honest summary is that tokenisation improves the plumbing of an asset and does nothing to create demand for it. Where demand already exists, the plumbing improvement is valuable. Where it does not, a better settlement mechanism for something nobody wants to buy is not the constraint that was binding.
Five questions, all answerable from public documents. What exactly does the token entitle you to, in the legal document rather than the marketing. Which entity owes it to you and in which jurisdiction. Who holds the underlying and who verifies that they hold it. How redemption works, for whom, and when it can be suspended. And whether transfer restrictions apply that limit who can buy the token from you.
The pattern in the answers is what distinguishes the working applications from the rest. Where the underlying is homogeneous, continuously priced, held by a regulated custodian, and redeemable routinely, the token behaves. Where any one of those is absent, the token is a claim whose price and whose exit both depend on conditions that are not guaranteed, and the chain it lives on has no bearing on either.
No. It puts a record of a claim on the blockchain. The connection between that record and the asset exists in a legal document enforced by courts in a jurisdiction. The chain can prove who holds the token and has no power to make anybody deliver the thing it represents.
Settlement speed and finality, availability outside market hours, and divisibility into small units. For assets whose main friction was a multi-day settlement chain with several intermediaries, that is a real economic improvement. None of it changes what the asset is worth or whether you can get it.
Because the underlying is homogeneous, continuously priced, held by regulated custodians, and redeemable as a routine operation. Those four properties make the claim easy to value and easy to exercise, which is what keeps the token tracking the asset. A building has none of them.
It depends entirely on the structure, and it is the question that matters most. Some allow redemption only for large or qualified holders, with minimum sizes and processing delays, and most reserve the right to suspend it. Everybody else can only sell to another holder, which requires a market to exist.
No. Transferability is technical, liquidity is economic and requires buyers willing to trade at a price. Tokenising an illiquid asset produces a tokenised illiquid asset with better settlement. The divisibility is often nominal too: you can hold a fraction and still have nobody to sell that fraction to.
Because the legal structure requires it, usually to remain compliant with securities rules in the relevant jurisdictions. The restrictions are enforced by the contract itself, limiting transfers to verified addresses. They are generally necessary and they mean the token does not circulate freely the way the vocabulary suggests.
By an appraisal, which is a professional opinion updated occasionally rather than a market price. Any system acting on that number, including anything lending against the asset as collateral, is acting on an estimate. The gap between the estimate and what the asset would actually fetch is invisible until somebody tries to sell.
How redemption works: who is eligible, at what minimum size, how long it takes, and under what circumstances the issuer may pause it. If redemption functions routinely, arbitrage keeps the token at the value of the underlying. If it does not, the token trades at whatever its own market will pay.