How Big Is the Tokenized RWA Market? $35 Billion or $367 Billion, Same Dashboard

Nobody is lying to you about the size of the tokenized RWA market, and that is exactly what makes it hard to read. Open RWA.xyz this week and one figure says $34.94 billion. Another figure, in the column immediately beside it, says $366.97 billion. Same site, same day, same sector, roughly ten times apart. Neither is a typo, neither is a marketing number, and the small grey header above each one explains the entire gap to anyone who stops to read it.

You have probably already been caught by this without noticing. A post says tokenization has passed $400 billion. A research note the same week calls it a $34 billion market and treats that as the bullish case. You read both, assume one of them is talking its book, and quietly stop trusting the category. That reaction is reasonable and it is also the trap, because both writers were reading a real number off a real dashboard.

So "how big is the tokenized RWA market" is not yet a question you can answer. It is a question with a second question hidden underneath, and until you ask the one underneath, you are collecting figures that cannot be laid next to each other.

This is a Kodex walkthrough with Eunha. She lives between structure and emotion. Her working habit is to meet a claim with a question about what it counts, which is the one habit this sector rewards.

Eunha does not start with a definition of real-world assets. She puts four numbers on the desk in front of you and asks which one you would be willing to defend in writing.

Four totals, all current, all defensible

She writes each figure with its date attached, because a number without a date is a rumour with a decimal point.

$34.94 billion is the distributed asset value on RWA.xyz in late July 2026, up 3.92% over the previous thirty days. $366.97 billion is the represented asset value on that same dashboard on that same day, down 0.56% over the same thirty days. $298.75 billion is the stablecoin total, tracked in its own row. And $11 billion was the record set by tokenized US Treasuries alone back in March, the week CoinDesk reported Circle's USYC passing BlackRock's BUIDL at the top of that category.

Fold the stablecoins back into the distributed figure and you land near $334 billion. That is a fifth number, and it is also correct.

You can feel the reflex arriving. One of these has to be the real one, and the others must be inflated, stale, or sloppy. Eunha lets that sit for a moment before she takes it away. The reflex is the actual problem: it assumes the sources are competing to measure one thing.

The disagreement is not arithmetic. It is scope.

One dashboard, two columns, ten times apart

The two columns have names, and the names are the whole answer. RWA.xyz defines distributed assets as tokenized assets that can be moved to wallets outside the issuing platform and transferred between wallets, including those carrying whitelist or eligibility controls. It defines represented assets as tokenized assets that cannot be moved to wallets outside the issuing platform or transferred between wallets, whether by design or because a regulator requires it. The firm published that framework for a blunt reason it states itself: everything was still being labelled a tokenized asset, a term too broad to carry the difference.

Eunha's version of the test is physical, and she wants you to answer it out loud. "Do not ask what it is called," she says. "Ask whether it can reach you." Can you receive it? If the token can land in a wallet you control, sit as collateral somewhere the issuer never approved, and be sold to a counterparty the issuer has never met, it is distributed. If it lives and dies inside one institution's system, it is represented, and RWA.xyz describes that bucket plainly as a recordkeeping and reconciliation layer.

Figure Technologies' tokenized home equity lines are the clean case. They are real loans, tokenized in full, sitting on a ledger, and they belong in the represented column because they cannot leave the platform that issued them. Figure has said it is working to make its wider portfolio compatible with DeFi protocols. Those same assets would then change columns without a single loan changing terms. The asset would be identical. The number it belongs to would move by billions.

This is the same shape as a token the custodian still holds. The tokenization is real; what it upgraded was the record, not the custody.

Once you see that assets can change columns, a second effect falls out of it. An asset that becomes transferable does not add to the sector. It subtracts from one column and adds to the other. The two columns are an order of magnitude apart, so the same migration reads as a rounding error on one side and a visible jump on the other. A month in which serious value became tradeable for the first time can therefore show up as a shrinking represented figure, which is the opposite of what happened. Eunha treats any single-column trend line as unfinished for exactly this reason: you cannot tell growth from migration while looking at one of them.

There is a detail in the thirty-day change that makes the point harder than any definition can. Distributed value rose 3.92% while represented value fell 0.56%. Two columns describing one sector, on one screen, in one month, moved in opposite directions. Quote a single total and you would have reported growth or contraction purely as a function of which column your eye landed on first.

Those are not two views of one number. They are two numbers.

Does a stablecoin count as a real-world asset?

This is where the swing is largest, and there is no clean answer available, only a disclosure you can look for. $298.75 billion sits in stablecoins. The distributed non-stablecoin total is $34.94 billion. Putting one asset class in or leaving it out moves the sector's headline by roughly a factor of ten. That is a larger effect than every growth rate published about tokenization this year.

Both conventions defend themselves properly. A stablecoin is a tokenized claim on a dollar, and a dollar is about as real-world as an asset gets, so counting it is consistent with the category's own name. Against that, a stablecoin is held to be spent and moved, not held as an allocation. Folding $299 billion of tokenized cash into a sector total buries every other category underneath it. CoinGecko's research separates the two for that reason, and the separation is why its figures look small beside the ones that do not.

Eunha's rule here is short. The careful sources say "ex-stablecoins" somewhere on the page, and the ones that stay quiet about it tend to be quoting the bigger number.

If you have already worked through what a market cap actually counts, this is that lesson at sector scale. And a stablecoin total is a moving figure, not a floor. A shrinking stablecoin supply can drag a sector headline down in a month when nothing inside the sector shrank.

The biggest tokenized Treasury fund changed hands on a $6 million margin

You would have read that BlackRock was no longer the largest, and nothing in that sentence was wrong. Scope problems get more interesting when the number is small enough to look precise.

On 22 January 2026, Circle's USYC held $1.69 billion against BlackRock's BUIDL at $1.68 billion. The gap was about $6 million, roughly a third of one percent. It was enough to move the phrase "largest tokenized Treasury fund" from one company to the other in every headline that week. USYC had grown 11% over the preceding thirty days while BUIDL slipped 2.85%, so the crossing was real and the direction was real.

Underneath the ranking sat a fact the ranking could not see. Binance held around $1.43 billion of USYC, about 94% of its entire supply. BUIDL, at effectively the same size, had 103 unique holders. One fund's leadership rested on a single counterparty's balance; the other's was spread across a hundred institutions. A league table reads the total and nothing else.

By March the picture had both settled and grown. CoinDesk put USYC near $2.2 billion against BUIDL's roughly $2 billion. The whole tokenized Treasury category hit a record $11 billion, and BUIDL's share fell to 18% from a 46% peak the previous May.

Eunha's point is not that the January ranking was wrong. It was accurate on the day it was taken. A leadership claim is a snapshot of a margin. When that margin is a third of a percent held by one counterparty, the claim has a shelf life measured in weeks. It also means far less than the headline implies.

Did private credit really overtake Treasuries?

Two people can look at tokenized private credit, quote real figures, and land billions apart. The claim is the column problem wearing different clothes, which is why the argument about it never resolves.

Distributed private credit stood near $5 billion in March 2026. Broader counts that include represented and platform-locked positions run to $18 billion and above, and RWA.xyz's own May AUM view landed between them, with Treasuries around $10 billion and private credit around $8 billion inside a $22 billion total. Nobody in that spread is quoting a bad figure.

The reason the wide count runs so much wider is Figure again. A large share of tokenized active loans sits on permissioned infrastructure. Real credit, real borrowers, real repayment schedules, and none of it able to move to a wallet you control. Count what exists and private credit is enormous. Count what can move and Treasuries still lead comfortably.

So the overtake question has no answer, and the absence of one is the answer. It is not a dispute about data quality between rival providers. It is one category being measured on two different axes by people who each think they are stating a fact, which they each are.

The same split shows up inside single assets and not only across whole categories. A receipt for private shares is a tokenized claim whose transferability is the entire question, and the answer decides which column it lands in.

What each figure is actually counting

Eunha's complaint about all four is the same: they travel without their headers attached. That is the part you are really quoting, and the part that gets left behind.

No row here is truer than the others. They answer four different questions, and the sentence you are about to write is what decides which one you needed.

The four questions to ask before you quote a number

You already have the habit now. Eunha compresses it into four questions, ordered by how much time each one saves you.

Are stablecoins in or out? This is the largest single swing available, worth roughly a factor of ten. If the source does not say, assume they are in, and read the figure as tokenized cash plus everything else rather than as a sector total.

Can the asset leave the platform? Distributed or represented. This is the difference between a market and a ledger entry, and it is where the tenfold gap lives.

Is this a fund, a category, or the whole sector? "Largest tokenized Treasury fund" and "the tokenized RWA market" sit three orders of magnitude apart. A headline will move between them inside one paragraph without telling you.

What date, and is it a level or a change? In one thirty-day window this July, the distributed column rose 3.92% while the represented column fell 0.56%. A growth rate means nothing until you know which level it was measured against. A figure from four months ago is not wrong either. It is a different measurement, and setting it beside today's is how two accurate numbers combine into one false trend.

None of this makes the sector smaller or less worth watching. Tokenized value is being created, and the distributed column is growing on its own merits. The represented column measures how much conventional finance has already been rewritten into a new format. What it does mean is that "how big is the tokenized RWA market" is a question you cannot answer on someone else's behalf. An honest answer starts with what they intend to do with the number.

Eunha closes the dashboard and leaves the four questions sitting on the desk. The next figure you meet will arrive without them attached. That part you supply.

FAQ

How big is the tokenized RWA market?

There is no single figure. In July 2026, RWA.xyz reported $34.94 billion of distributed assets, $366.97 billion of represented assets, and $298.75 billion of stablecoins as three separate lines. Which one is "the market" depends on what you mean: assets that trade freely, assets written onto a ledger, or both plus tokenized cash.

Why do different sources give different RWA numbers?

Because they are measuring different things and each is being accurate about its own scope. Four variables do the work: stablecoins in or out, assets that can leave the platform versus assets that cannot, fund versus category versus sector, and the date.

Are stablecoins part of the RWA market?

Both conventions are in use. A stablecoin is a tokenized claim on a real-world asset, so including it is defensible. At roughly $299 billion it also dwarfs every other category and swings the sector headline by about a factor of ten. Sources that exclude them normally say "ex-stablecoins" explicitly.

What is the largest tokenized Treasury fund?

Circle's USYC passed BlackRock's BUIDL on 22 January 2026, $1.69 billion against $1.68 billion. The margin was roughly $6 million. By mid-March it had widened to about $2.2 billion against $2 billion. At the January crossing, Binance held about 94% of USYC's supply while BUIDL had 103 unique holders. The two funds were the same size in very different ways.

Is private credit bigger than tokenized Treasuries?

It depends on the axis. Distributed tokenized private credit was near $5 billion in March 2026, while counts that include represented and platform-locked positions reach $18 billion and above. RWA.xyz's May AUM view had Treasuries near $10 billion against private credit near $8 billion.

Does tokenized mean tradeable?

No, and this is the distinction that produces the largest disagreements. A represented asset has been tokenized as a record but cannot be moved to a wallet outside the issuing platform. A distributed asset can. Both are correctly described as tokenized, and only one of them is something you could receive.

Eunha's version of homework: open the simulator, put a tokenized stock and gold side by side on the $5,000 paper account, and hold both through a full session. The labels will tell you what each one is called. Only the session tells you what each one does.

Start free →