Analyzing Stellar RWA Market

01 · The ticker tells you nothing

393 accounts on Stellar issue something called USDC

Open a Stellar wallet and you see a list. A short code, a balance, maybe a logo. The code says USDC and the balance says 40, and that reads like 40 dollars.

There are 393 separate accounts on Stellar that issue an asset called USDC. One of them is Circle's. The other 392 are somebody else, and in a wallet list they look the same.

This is not a bug. Stellar was built this way. An asset here is a pair, a short code plus the account that minted it, and the chain treats the same code from two accounts as two separate assets. That works fine when you know who you are dealing with. It works badly when a ticker is the only thing you check.

Bitquery indexes every Stellar ledger from the first one in 2015. We took the full record, 4.3 billion payments and 2.2 billion order book trades, and matched each one to the account that issued the asset it moved. Then we matched those accounts against the network's public directory to see which of them belong to an organization anyone can name. The question underneath was simple: on a chain built for moving money, how much of what moves is a claim on something real, who issues it, and is it reaching more people or fewer.

One limit belongs up front. An account nobody has named is simply unnamed. That is a weaker thing than proven fake. The public list of Stellar accounts is kept by hand, and our first pass at this census missed several real issuers because of it. Section 06 says which ones and how the numbers moved.

02 · What Stellar was for, and what turned up

A currency rail that asset managers started using

Stellar is a payments chain. It is old for crypto, cheap to use, and built around one job: letting somebody hold a claim on real money and send it to somebody else, who turns it back into their own currency at the far end. The firms that issue those claims are called anchors. A payments company in Tanzania holds shilling in a bank and issues a matching token. Somebody in Dar es Salaam sends it to Nairobi, and a Kenyan anchor pays out shilling there. The Stellar blockchain API records every step of that.

A real-world asset is the same idea, cast wider. A token that is a claim on a thing that exists off the chain: a dollar in a bank account, a share of a fund, a state bond. It is not a coin whose price floats on its own. Somebody has promised to hand you the real thing back.

For most of Stellar's life the real assets on it were currency and nothing else. RWA tokenization on Stellar meant a bank account, and that was the whole of it. Then it changed. Franklin Templeton put the share register of a US money market fund on this chain, and it has run there since 2021. Ondo brought a Treasury note in September 2025. Etherfuse issues state debt tokens for Mexico, Brazil, the United States, the United Kingdom and South Korea. WisdomTree runs a set of fund tokens across 17 issuing accounts. Paxos issues PayPal's dollar here.

Count every firm with a live claim on value held off the chain and you get 29 of them, across 68 issuing accounts. Between them they moved roughly an eighth of all the non-XLM payments Stellar carried this year.

03 · Big money, few holders

The largest assets on Stellar are held by a few hundred accounts each

Rank Stellar's assets by size and the top of the list is not payment money. USDY from Ondo leads, and the Franklin fund is second. Circle's dollar is third.

Rank the same assets by how many accounts hold them and the order turns over. USDC sits in more than 680,000 wallets. USDY sits in 685. The UK gilt token from Etherfuse sits in two.

A second split runs through the same table, and it sorts the funds from the money. Compare what an asset holds in ordinary wallets against what the ledger says exists in total, and the gap is what sits parked in market maker pools instead of with a person. Franklin's BENJI is almost all in wallets. So is USDC. Close to half of the Argentine peso token and half of Circle's euro live inside pools.

Fund tokens get held. Currency tokens get traded. That is what you would expect, and it is a useful check on any supply figure you read about this chain.

04 · Growing and emptying at once

Payments up 47%, the accounts receiving them down 24%

Take every payment carrying one of these 29 organizations' assets, and count the same window each year, 1 January to 8 September. The count of payments has gone up every year without exception, and this year it went up by nearly half again.

Now count the distinct accounts on the receiving end of those payments. That number peaked in 2024, held flat through 2025, and this year it fell by about a quarter.

The clearest place to see why is Circle's dollar, which carries most of the volume. USDC on Stellar moved about $1 billion in a month at the start of last year, and close to $9 billion by this August, the last full month we have. That rise is real. It is still not a growth story.

The typical payment did not change. The median stayed a shade over $10, and the count of accounts sending USDC actually fell. What changed is who was moving money at the top. In August a single account, Circle's own anchor, was close to a third of the entire month. Circle's custodian was next. Binance sat just behind, on both its deposit and its withdrawal address.

That is issuance, redemption and exchange settlement, which is plumbing rather than demand. It is a normal thing for a stablecoin to do. It is just not the same thing as more people using Stellar to hold dollars, and a headline volume figure cannot tell you which one you are looking at.

05 · The lookalikes, and what they cost

Real money changed hands for tokens whose tickers belong to somebody else

We took 27 tickers that a general reader would recognise as finance, from USDC and USDT through gold and silver to BlackRock's BUIDL and a row of listed-company symbols, and counted every Stellar account that issues one of them and has made at least one payment with it. The count came to 1,620 accounts. Eight of them belong to an organization the directory can name.

Some of the imitators put work in. Stellar addresses can be ground out until the last few characters spell whatever you want, and one fake GOLD is issued by an account ending in the letters GOLD. Another ends in USDC. Another in USDT. Somebody checking the tail of an address instead of the whole thing sees what looks like confirmation.

This matters because people buy them. On Stellar's built-in order book this year, 10,943 separate accounts sold real XLM to buy a token branded as a dollar, a metal, a Treasury fund or a listed share, from an account with no connection to that name. They gave up 3.97 million XLM doing it, worth roughly three quarters of a million dollars at the price XLM was trading when we measured.

The largest share went to metal. Tokens calling themselves gold or silver took 84% of that spend, well ahead of the fake dollars. A fake dollar is easy to check against a real one. A fake gold token has no obvious price to check it against.

The most widely held asset on Stellar under the ticker XAU is issued by an account the public directory flags as a scam outright. It has paid out to 9,338 accounts.

Stellar order book, 1 January to 8 September 2026

That public list tells you what kind of place this is. It runs to nearly 18,000 entries, and about 19 in every 20 are on it because somebody flagged the account as a scam. Fewer than 500 are on it because the account issues or anchors anything. Almost all the work of naming accounts here goes into naming frauds.

06 · What we got wrong the first time

The audit changed the headline

Our first pass at this census built the issuer list the obvious way. Rank assets by activity, check the big names by hand, publish. That found 10 issuers and concluded that Stellar carried no bond tokens, no share tokens and no metal tokens at all.

That conclusion was wrong, and the reason it was wrong is worth stating. Ranking by activity finds the assets that many accounts touch. It cannot find a program that moves half a billion dollars between a few hundred accounts, which is exactly what Ondo, Etherfuse and WisdomTree look like on this chain. Rebuilding it from the public list instead, and checking all 318 accounts it tags as an issuer or anchor, took the count from 10 to 68.

Three other figures moved in the same pass.

The general lesson sits in the first line of this piece. You cannot screen assets on this chain by ticker, and you cannot screen them by a data provider's asset identifier either. The issuing account is the only key that means anything. We got that wrong once in our own pipeline before it turned into a published number.

07 · Method and limits

What was measured, and what it cannot say

Everything here comes from Bitquery's Stellar archive, which starts at the chain's second ledger in 2015 and runs to the day we measured. Payments come from the network's payment record, order book activity from its trade record, and balances from its balance-change record. Every query is capped at that same day, so the figures can be checked again against a chain that keeps moving.

Supply and holder counts are the ledger's own, cross-checked against stellar.expert. Where we computed supply ourselves by summing balances, that method is a floor rather than a total, because assets sitting in market maker pools and claimable balances are not held by any account. The held-in-wallets table in section 03 is that gap, reported rather than hidden.

The same index sits behind the Stellar blockchain API, the address APIs for tracing a wallet's counterparties, and Coinpath for following money between them. For the same question asked across every chain at once, see our count of all 11 types of tokenized asset, and for what happens when tokenized shares meet a thin order book, see the $45 tokenized stock market and xStocks on Solana.