Ask how big the tokenized real-world asset market is and you will be given a number with impressive precision and no stated definition. Ask two sources and the answers can differ by several times over; ask about 2030 and credible forecasts diverge by more than tenfold. None of that is dishonesty. It is that four separate choices get made before any total is published, each one legitimate, and whoever quotes the figure has usually made all four for you without saying so. This report is about those choices, because once you can see them you can read any number in the sector correctly. For the underlying structures, start with what real-world assets are.
At a Glance
- Honest answer — There is no single number
- Biggest swing — Whether stablecoins count
- Second biggest — Outstanding or cumulative
- 2030 forecasts — Differ by over tenfold
- What moves totals — Definitions, not flows
- Always ask — What is being counted
01 — Four choices, made before you see the number
Every published RWA total is the output of four decisions. Does a stablecoin count as a tokenized real-world asset? Does a token that cannot leave the platform that issued it count as being on-chain at all? Are you counting what is outstanding today, or everything that has ever been issued? And what qualifies as a "real-world asset" in the first place?
Answer those four one way and you get a large, payments-dominated market. Answer them the other way and you get a small, fast-growing institutional one. Both describe the same reality. The reason this matters practically is that the sector's headline figures move for definitional reasons at least as often as they move because anyone bought or sold anything — and a reader who cannot tell the two apart will mistake a reclassification for a trend.
02 — Choice one: do stablecoins count?
This is the largest single swing factor, and the convention is strange when you state it plainly. Most headline RWA totals exclude stablecoins. Yet a fiat-backed stablecoin is a token representing a claim on short-dated government debt held by a custodian — structurally the same instrument as a tokenized treasury fund, which everyone does count.
The distinction being drawn is really about function rather than substance: stablecoins are treated as the money layer and everything else as the investment layer. That convention has a firmer basis than it first appears, because the law makes the same cut — a compliant payment stablecoin is regulated as a payment instrument while a tokenized fund is regulated as a security, so the two sit in genuinely different regimes despite holding much the same assets. It is also the convention most trackers follow. But its effect is enormous. Include stablecoins and tokenized real-world assets are a substantial market overwhelmingly dominated by payment instruments. Exclude them and the same sector is smaller by roughly an order of magnitude and looks like an emerging institutional niche. Neither framing is wrong. Quoting either without saying which one you used is where the misleading starts.
03 — Choice two: what counts as being on-chain?
In November 2025 the sector's most-cited data provider split its figures in two, and the idea behind the split is the most useful measurement concept in the field. It asks whether a token can actually leave the platform that issued it and move directly between wallets. Assets that pass are counted as distributed; assets that fail — because they are locked inside an issuer's system or restricted to a closed list of holders — are counted as represented. Distributed became the default view.
No token was redeemed and no investor withdrew anything, but a significant part of the market moved out of the headline figure overnight. That reclassification was not a loss of value; it was an admission that the two things are not the same. A token that cannot be transferred is a database record with additional steps, and almost everything tokenization claims to add — composability, settlement outside market hours, transfer without an intermediary — requires the token to move. When you see a total, the first question is which of the two it counts.
04 — Choice three: outstanding, or ever issued?
Tokenized private credit, the largest category outside government debt, is routinely reported two ways. One figure is the loans currently active. The other is cumulative originations — everything ever written on those platforms since inception. Because loans are repaid over their lifecycles, the cumulative number runs at roughly double the active one.
Both are real, and each answers a different question: cumulative volume tells you how much business has flowed through the rails, while the active figure tells you how much exposure exists right now. Only the second is a market size. Quoting cumulative originations as though it were outstanding value is the most common way sector figures get inflated, and it is easy to miss because the larger number is usually the one in the headline. The same trap appears wherever a category reports throughput rather than balances, and the mechanics behind it are covered in our note on tokenized private credit.
A related inflation is harder to spot: the same underlying asset counted twice. Tokenized funds are increasingly held as reserve backing for other tokens, and tokens are sometimes wrapped again to move between chains. When both the wrapper and the thing inside it appear in the same total, one pool of government bills has been counted as two tokenized assets. Nobody is falsifying anything — each token genuinely exists — but the total no longer measures how much real-world value has been tokenized. As tokens are used more heavily as collateral for other tokens, this problem grows rather than shrinks, and it is the least-disclosed adjustment in the sector.
This market has been reclassified faster than it has grown. Before treating any change in the total as a trend, check whether the definition moved instead.
05 — Choice four: what is a "real-world asset"?
The obvious members — government debt, private credit, gold, equities, property — are not in dispute. The boundary is. Does a tokenized bank deposit count, given that it is a bank liability recorded on a permissioned ledger? Does a conventional fund that has added a blockchain-recorded share class, transferable only among approved institutions, count as tokenized or merely as digitised? What about carbon credits, invoices, or claims on future revenue?
Different trackers draw these lines differently, and the categories that grow fastest are frequently the ones with the loosest boundary. There is a related trap worth naming: concentration. A single issuer can account for the majority of a whole category, so a "category" figure is sometimes a description of one company's balance sheet rather than of a market. Before drawing conclusions from a category's size, check how many issuers are actually in it — a point that also governs how to read the dominant category, tokenized treasuries.
06 — Why the forecasts disagree by more than tenfold
Published 2030 estimates from serious institutions run from the low single-digit trillions to several tens of trillions. A spread that wide between competent analysts is not a disagreement about adoption speed. It is the four choices above, compounded, plus a fifth: some forecasts project the value of tokenized assets, others the size of the business opportunity around tokenization, and others total addressable demand — three different quantities routinely quoted as though they were one. Several of the largest figures include the money layer that the current-market figures exclude, which alone accounts for much of the gap.
Two habits follow. Treat any forecast that does not state its definition as marketing rather than analysis, whoever published it. And check the base rate: earlier optimistic projections for this sector have already been missed and quietly revised down, which is the ordinary fate of technology forecasts and no reason for either derision or belief.
07 — What the number is actually good for
Less than it appears, and that is fine — the total is a poor instrument, but its components are informative. Useful questions it can answer: which categories are compounding and which are flat; whether growth came from many new participants or from a handful of large institutional mandates; and whether distributed value is growing as fast as represented value, because if it is not, adoption is narrower than the headline suggests.
Useless uses are just as clear. Comparing totals across sources without checking their definitions produces nonsense. So does quoting a figure without its date in a sector growing this fast. And the most common error of all is the comparison to the notional size of the asset class being tokenized — the hundreds of trillions of global assets that could in principle be put on a chain. That is a statement about ambition, not a measurement of anything, and it belongs nowhere near a market-size discussion.
08 — Reading any RWA figure in thirty seconds
Six questions, and they work on any number in this sector. Are stablecoins in or out? Is it distributed or all assets? Is it outstanding balances or cumulative volume? What date is it from? Whose data is it, and does that party sell tokenization services or issue the assets being counted? And how concentrated is the category — one issuer, or many? A figure that cannot answer those is not a measurement, it is a headline.
The durable conclusion, which will outlast every number in circulation: on the strictest definition — transferable tokens, outstanding balances, stablecoins excluded — this market remains small, both against the stablecoin float beside it and against the conventional markets whose assets it references. It is also compounding quickly from that small base, and concentrated in the safest, most standardised asset there is. That combination, rather than any total, is the honest description. Watch the mix and the rate; treat the headline as the least interesting thing about it.
"Thou shalt not have in thine house divers measures, a great and a small." — Deuteronomy 25:14
Methodology & Sources
This report is about how tokenized real-world asset market figures are constructed, not about their current level. It deliberately contains no totals, category sizes, growth rates, forecasts or percentages, because every such figure in this sector dates within months and because the report's argument is that quoting one without its definition is the error under discussion; readers should consult current tracker data directly and apply the six questions in the final section. Descriptions of measurement conventions reflect the published methodologies of the principal data providers as at the date of publication, including the distinction between transferable and platform-restricted tokens introduced in November 2025 and the parallel reporting of active balances alongside cumulative originations in the private credit category; those conventions are set by commercial data providers rather than by any standard-setter, are not uniform across sources, and may change again. The characterisation of forecast divergence reflects the range of published institutional projections and the differing quantities they measure; no forecast is endorsed, and none should be read as a prediction by this publication. The observation that some published figures come from parties that also issue tokenized assets or sell tokenization services is a general caution about incentives, not an allegation of misconduct against any provider, and no data provider, issuer, platform or institution is named. Statements about relative scale are directional and jurisdiction-agnostic. Nothing here is a recommendation regarding any product, issuer, data provider or transaction, and nothing here is legal, tax or investment advice.