On-chain investigationXRP Ledger13 years5.06bn transactions

Who actually uses the XRP Ledger?

The XRP Ledger opened in 2012, two years before Ethereum, and has since settled more transactions than Ethereum has. We counted all of them, from the first block in January 2013 to this September, and then went looking for the people sending them.

At a glance
The XRP Ledger is a public blockchain built in 2012 to move money between banks. We counted every transaction it has settled since 2013, all 5.06 billion of them, and then counted who sent them. In August 2026, 793 accounts sent 93% of the traffic, and nearly half of all active accounts sent a single transaction and stopped. The median account holds 12 XRP. Strip out the machines and the spam and what remains is 0.8% of the ledger, a figure that has not grown since January 2018. Every fee the chain has ever charged comes to about $7.6 million, and nobody receives it. On 23 August a phishing campaign began that reached 1.28 million wallets in 17 days, for $160 in fees.
93.2%
Of August traffic came from 793 accounts
0.80%
Of traffic is real, human-scale payments
12 XRP
Median account balance, about $25
$7.6m
Every fee the ledger has ever charged

01 · The busiest account on the ledger

Thirteen million orders in a month, and almost none of them traded

On the XRP Ledger there is an account that placed nearly 13 million orders in August. It worked every day of the month, about five orders every second, and never stopped for long. Almost none of them did anything. Out of nearly 13 million transactions, 882 ended in a trade.

That one account was a sixth of everything the ledger did that month. It paid about the price of a phone in fees to do it.

The XRP Ledger is a public blockchain that opened in 2012, two years before Ethereum. It settles in about four seconds, costs a fraction of a cent to use, and was built for one job: moving money between banks without the chain of other banks that normally sits in the middle. It has its own coin, XRP, one of the most valuable in crypto. Ripple, the company that built the ledger, still holds a large part of the supply and releases it on a schedule.

Bitquery indexes every XRP Ledger transaction from the first block in January 2013. We counted all of them, then counted the accounts behind them, then went through what those accounts hold, what they send, what the messages attached to their payments say, and how many of them ever come back. This ledger has settled more traffic than Ethereum, and it did so with a crowd that would fit in a football stadium. The features Ripple built for banks have user counts you can hold in your head. Somewhere along the way the busiest thing on the chain became spam, and last month it became a phishing campaign.

One line runs through all of this and belongs up front. XRP the coin is widely owned and heavily traded. Almost all of that trading happens inside exchange accounts, where it leaves no record on the ledger. We are counting the ledger here, and the coin is a separate question.

02 · A few hundred machines

The busiest thousand accounts have taken most of the chain every year since 2013

Last August the ledger settled about 82 million transactions. Just over a fifth of a million accounts sent them. Group those accounts by how hard each one worked and the chain comes apart.

Six groups of accounts, and what each group actually sentAugust 2026, every transaction on the ledger
SHARE OF ALL 81,568,239 XRP LEDGER TRANSACTIONS, AUGUST 2026793 accounts (0.37%) sent over 10,000 each93.2%819 accounts (0.38%) sent 1,001-10,0003.65%4,567 accounts (2.1%) sent 101-1,0002.10%33,499 accounts (16%) sent 6-1000.73%70,197 accounts (33%) sent 2-50.24%104,191 accounts (49%) sent exactly one0.13%The 174,388 accounts that sent 5 transactions or fewer produced 0.37% of the chain between them.Source: Bitquery XRP Ledger archive. Counts sending accounts only, so dust spam cannot inflate them.
The 793 accounts at the top sent more than 10,000 transactions each. The 104,191 at the bottom sent one. Between them sit almost no ordinary users, which is the shape of a machine network rather than a crowd.

Nearly half of every account that touched the ledger in August sent a single transaction. Four fifths of them sent five or fewer, and all of those people together produced about a third of one percent of the chain.

This is not a recent turn. We ran the same count for every year since the ledger opened. The share of traffic taken by the busiest thousand accounts has never once fallen below seven tenths, and in one year it reached almost the whole chain.

So we sorted the 793 by what they actually do. Every one of them falls into a machine category, and they account for the whole of that top band.

What the 793 busiest accounts areAccounts, and their share of the whole chain
DEX order bots283 accounts, 39,254,428 transactions, 48.1% of the chain
Payment sprayers233 accounts, 19,685,565 transactions, 24.1%
Mixed, mostly NFT offer spam262 accounts, 11,662,568 transactions, 14.3%
Check spam5 accounts, 3,436,654 transactions, 4.2%
AccountSet no-op spam2 accounts, 1,366,095 transactions, 1.7%
Escrow spam8 accounts, 578,602 transactions, 0.7%
Exchanges among themNone

Bots and outright spam are 78.8% of the ledger. Not one exchange appears in the group, and the two accounts labelled no-op spam send a transaction type that changes a setting and moves nothing at all.

03 · The growth is dust

The median XRP payment in 2024 was the smallest amount the ledger can express

For most of the ledger's life its busiest transaction type was an order placed on the built-in exchange. In 2024 that changed. Payments overtook orders, counts roughly tripled, and the charts that track XRP Ledger activity turned sharply upward.

Then we looked at how much money those payments moved.

The median XRP payment, year by yearlog scale, so each step down is ten times smaller
MEDIAN XRP PAYMENT, BY YEARlog scale — each gridline is 10x1 drop1k drops1 XRP1,000 XRP2019105 XRP2020100 XRP2021137 XRP202292 XRP202339 drops20241 drop202538 drops202686 dropsA drop is one millionth of an XRP. In 2024, half of every XRP payment on the ledger delivered that or less.Source: Bitquery XRP Ledger archive, delivered value on successful payments of native XRP.
From 2019 to 2022 the middle payment on the ledger was around 100 XRP. In 2024 it was a single drop, one millionth of one XRP. Nothing about the economy of the chain changed in between. The spam arrived.

About two in every five XRP payments this year have delivered ten drops or less. A few thousand senders produced all of them, and they went out to millions of wallets. It is the same trick other chains see. Fill a stranger's wallet with an amount too small to matter, and your address is now in their history.

XRP payments in 2026 so farCount
Delivered 10 drops or less52,756,051, from just 10,549 senders, to 2,408,137 wallets
Delivered under 1 XRP59,224,857 more
Delivered 1 to 100 XRP15,067,256
Delivered over 10,000 XRP721,877

This matters for anything you read about XRP Ledger activity. A figure that counts wallets receiving a transaction absorbs all of that spam. Every number here counts the account that sent and signed, which spam cannot inflate.

04 · What the dust actually says

A phishing campaign that reached 1.28 million wallets in 17 days

XRPL payments can carry a memo, and the memo is readable text. So we read them. In August, 11.7 million payments carried one. The largest single category is not a business.

What arrived in 1.28 million wallets, verbatim from the ledger
Incoming Escrow for 259.37 XRP Expires Soon. Finalize Escrow via xrp-escrows.com

🎁 You have been sent a free XRP CHECK worth $48,000! Visit - usdonxrp.com - to redeem it.
1,280,036
Distinct wallets reached in 17 days
~7
Accounts running it after the first three days
69.6 XRP
Total fees the whole campaign paid

Nothing of this kind existed on the ledger before 23 August 2026. It started that day, ran to 8 September, which is the last day of our data, and was still going when we stopped counting. On its busiest day it sent more than 1.2 million messages. After the opening burst the fleet narrowed to about seven accounts, each pushing a few hundred thousand payments a day.

The whole thing cost 69.6 XRP in fees, roughly $160. That is what it costs to put a message in front of one in every seven accounts that has ever existed on this blockchain.

What the memos on August's payments sayPayments, and who sent them
Phishing bait4,597,309 payments from 1,879 accounts
x402 agent payments2,645,514 from 142 accounts
Token airdrop spam2,236,145 from 18 accounts
Everything else1,878,909 from 26,051 accounts
Memos reading "Horizon Volume Booster"312,756 from 21 accounts

Two lines in that table deserve a note. The last one is a service that names its own purpose in the memo field of every payment it sends. The second is the one genuine surprise in this data: x402, the payment standard built for AI agents, is live on the XRP Ledger and stamps its payments Facilitator by t54 labs. It is the only growing use on this chain that is growing for a reason other than spam. We measured the same standard across six other chains and found a similar gap between transaction counts and money.

05 · Nothing is built on it

Payment channels: 58 accounts. Oracles: 16. Permissioned domains: 15.

A blockchain earns the word platform when other people build on it. The XRP Ledger has spent a decade adding the parts you would need: a built-in exchange, escrow, payment channels for streaming tiny payments, NFTs, an AMM, price oracles, digital identity, credentials, and a token type aimed at real world assets.

We counted the accounts that used each one over the twelve months to the end of August.

Every feature on the ledger, and how many accounts used it in a yearunique signing accounts, September 2025 to August 2026, log scale
UNIQUE ACCOUNTS THAT USED EACH FEATURE, 12 MONTHS TO 31 AUG 2026Sent a payment1,568,580Held a token (TrustSet)179,541Traded on the native DEX59,541Bought an NFT16,109Added AMM liquidity11,245Minted an NFT4,383Created an AMM pool3,861Created an escrow1,682Issued an MPToken (RWA)76Set a DID68Opened a payment channel58Issued a credential42Set an oracle16Set a permissioned domain15Every feature built for institutions ends in double digits. The native DEX averaged 5,150 orders per account.Source: Bitquery XRP Ledger archive. Log scale. Counts the account that signed each transaction.
Payments reach 1.57 million accounts. The built-in exchange, the ledger's oldest feature and its main claim to being a DeFi chain, reaches 59,541 and averaged 5,150 orders each. Everything Ripple built for banks and funds lands in double digits.

Payment channels are the feature that made the ledger famous with developers, the piece that was going to carry tiny streaming payments across the internet. In a year, 58 accounts opened one. Ripple's token type for banks and funds was used by 76 accounts. Permissioned domains, the compliance rail sold to regulated finance, were set by 15.

The built-in exchange deserves its own line. It carried hundreds of millions of orders in the year, which sounds like a market. Strip August down to what actually settled and both the exchange and the NFT market come out small.

What actually settled in August 2026Rows, and the accounts behind them
Fee burns81,562,652
Payments20,317,655, from 202,643 accounts
Trades on the built-in exchange2,970,922, from 12,153 accounts
NFT trades106,768, from 3,840 accounts
NFT mints218,793, from 1,038 accounts

06 · What the accounts hold

The median account has about $25 in it

Counting transactions tells you what accounts do. Counting balances tells you whether anyone has anything at stake. So we took the current XRP balance of every account ever created on the ledger.

What every XRP Ledger account actually holdscurrent balance of all 8,972,159 accounts ever created
WHAT EVERY XRP LEDGER ACCOUNT ACTUALLY HOLDSSHARE OF ACCOUNTSSHARE OF ALL XRP ON THE LEDGERZero. Emptied or deleted9.58%0.00%Under 1 XRP2.31%0.00%1 to 1026.94%0.01%10 to 10037.82%0.12%100 to 1,00010.35%0.51%1,000 to 10,0009.20%4.0%10,000 to 100,0003.43%11.8%100,000 or more0.36%83.5%The median account holds 12 XRP. 76.7% hold under 100. Account share is drawn at 2x scale so the small bands stay visible.Source: Bitquery XRP Ledger archive. Current balance of all 8,972,159 accounts ever created, measured in 8 partitions and summed.
Three quarters of all accounts hold less than 100 XRP. The 32,437 accounts holding 100,000 or more hold five sixths of every XRP on the ledger. An account needs 1 XRP just to exist, so a large part of the bottom is the reserve and little else.

The median XRP Ledger account holds 12 XRP, about $25. Nearly two fifths hold less than 10. At the other end, roughly a third of one percent of accounts hold more than four fifths of the XRP that sits on the chain.

That measurement has a useful check on it. The balances we counted come to 68.1 billion XRP. Ripple keeps around 32 billion locked in escrow contracts, which are separate ledger objects and are not account balances. The two add up to the 100 billion supply, which means the method reconciles against a figure derived a completely different way.

07 · Nobody stays

Two thirds of last year's arrivals never came back after their first month

New accounts do arrive. In its best month ever the ledger added more than a third of a million of them. What happens next is the question.

What happens to every crowd the ledger drawsaccounts grouped by the year they first transacted
WHAT HAPPENS TO EVERY CROWD THE XRP LEDGER DRAWSarrivedaccountsone month onlystill active in August 2026201331,45636.8%0.13%201531,66439.8%0.13%2017393,12830.8%0.15%2019342,00751.5%0.29%20211,391,70934.2%1.08%2023477,22640.7%2.21%2024827,81548.3%2.88%20251,425,07467.3%3.44%The 2025 arrivals were the biggest crowd the ledger ever drew. Two thirds of them never came back after their first month.6,995,567 accounts have ever sent a transaction. 214,066 sent one in August 2026.Source: Bitquery XRP Ledger archive. Bar scale runs 0 to 4%. Alive = signed any transaction on or after 1 August 2026.
Read the last column. No group of arrivals, in thirteen years, has kept more than 3.5% of itself. The groups from before 2020 have kept about two accounts in every thousand.

Nearly 7 million accounts have ever sent a transaction here. Fewer than a quarter of a million sent one in August, which is about one account in every forty that exists. More than half the accounts alive today first appeared this year, and 46% of all funded accounts have not been touched since the end of 2024.

That is the shape of a treadmill. The chain takes on hundreds of thousands of new accounts a year and keeps almost none of them, and the total holds steady because the arrivals replace the ones that quietly stopped.

08 · What the institutions actually do

A treasury schedule, a stablecoin held by ten addresses, and no banks

The case for the XRP Ledger has never really been retail. It is that banks and payment firms use it. So we went looking for them.

The largest flow on the chain by value is Ripple's own escrow. Every year since 2018, between nine and twelve billion XRP has been locked into escrows and released again, on a fixed monthly cadence. That is a treasury release schedule for the company's own holdings. It is the biggest thing here and it is Ripple paying Ripple.

The one thing genuinely working is RLUSD, Ripple's dollar stablecoin. It opened at the end of 2024, and the amount moving through it has grown by more than seventy times since. That growth is real and it is the best signal on the ledger. Its ownership is another matter.

RLUSD, Ripple's dollar stablecoinMeasured across its whole life
Outstanding$1,024,338,708
Addresses holding at least $114,226
Held by the largest single address18.4%, or $188,048,186
Held by the top 10 addresses81.6%
Held by the top 100 addresses99.0%
Wallets receiving a plain transfer in a normal monthroughly 1,000 to 2,500

Then the banks. XRPL accounts can publish a domain, which is the closest thing the ledger has to a register of the businesses on it. We searched every domain ever declared, across thirteen years, for anything reading as a bank, a payment company, a remittance firm or a fund.

What comes back is memecoin token files and a handful of small XRPL sites. No bank appears. No payment company appears. Not one, in thirteen years.

Every bank-like domain ever declared on the ledgerAccounts declaring it
Memecoin token files on one launchpadSeveral hundred accounts across dozens of subdomains
getripple.org50
ripple.com26
xrp.moneymindedapes.com19
socialcredit.finance11
An actual bank, payment company or fundNone, in thirteen years
Check the issuer before you count a stablecoin
On the XRP Ledger a token is named by a four-letter code plus the address that issued it, and anybody may issue a code. At least a dozen addresses issue a token called RLUSD. Several use vanity addresses built to look like the real thing, including one beginning rLUSDtyk and another beginning rLUSDfDx. Ripple's genuine issuer is rMxCKbEDwqr76QuheSUMdEGf4B9xJ8m5De. Any tool that groups by currency code alone folds the fakes into the real figures. A second trap sits behind it: an XRPL partial payment lets a sender declare an enormous amount while delivering almost none of it, so the declared amount measures nothing. Summing it gave us RLUSD volumes around 10 to the power of 100.
12+
Addresses issuing a token called RLUSD
$5.34bn
Delivered by the genuine issuer in August 2026
10^100
The number you get if you trust declared amounts

09 · Take the machines away

What is left has not grown since January 2018

Every number so far describes a chain crowded with machines. The obvious question is what sits underneath them. So we removed them.

The test is deliberately generous. Keep any account that sent fewer than 100 payments in the month, which excludes every machine in the top band and nobody else. Then require at least one of those payments to have delivered more than a token amount, which excludes dust. What remains is people and small services doing something real.

Accounts doing something real, month by monthunder 100 payments in the month, at least one of them not dust
ACCOUNTS DOING SOMETHING REAL, MONTH BY MONTHunder 100 payments in the month, at least one of them not dust0150k300k450k600k196k20182019202120222024571k2025191k2026January 2018: 196,240 accounts. August 2026: 191,761. Eight and a half years, no growth.These accounts are 89.6% of everyone active in August and produce 0.80% of the ledger's traffic.Source: Bitquery XRP Ledger archive. Selected months shown; the full series runs every month from 2018.
The line ends where it began. Between those two points sit a bull market, an NFT boom, an AMM launch, a stablecoin and a hundredfold rise in the price of the coin.

Two numbers carry the whole argument. Those accounts are 89.6% of everyone who transacted in August, and their payments are 0.80% of the ledger's traffic. Nine in ten active accounts produce less than one percent of the activity.

The count itself has barely moved. There are about as many of these accounts now as there were in January 2018, before the ledger had an AMM, NFTs, a stablecoin or a hundredfold rise in the price of its coin. The peak came in December 2024, when RLUSD launched into a rising market. Two thirds of that has gone.

10 · The chain that opened six years later

Tron has 82 times the senders and spreads them properly

A number is easier to judge beside another number. Tron launched in 2018, six years after the XRP Ledger, and chased the same use: cheap payments and dollar stablecoins. Both chains sit in the same database here, so we ran the same count on each.

Concentration on the two chains, same month, same methodshare of the month's transfers taken by the busiest accounts
HOW MUCH OF THE MONTH THE BUSIEST ACCOUNTS TOOK, AUGUST 2026Busiest 1030.5% XRP Ledger3.9% TronBusiest 10072.8% XRP Ledger9.2% TronBusiest 1,00090.2% XRP Ledger14.5% TronTron opened in 2018, six years after the XRP Ledger. In August it had 17,558,743 signing addresses against 214,066.Source: Bitquery. Both chains measured on the same database, same method, same month.
On Tron the busiest thousand addresses take about a seventh of the month. On the XRP Ledger they take nine tenths. The Tron curve is what a chain with users looks like.

That month more than 17 million addresses signed a transfer on Tron. The XRP Ledger had about an eightieth as many accounts send anything at all. Our separate work on Tron takes apart what that chain carries, which is a different problem and a busier one.

11 · What the whole thing costs

Thirteen years of fees come to about $7.6 million, and nobody gets it

Every XRP Ledger transaction destroys a fee. The XRP is burned and no validator receives it, which makes the running total two things at once: a clean measure of how much the chain has been used, and the reason nobody is paid to run it.

Every fee the ledger charged, per year, in dollarsburned, not paid to anyone — no validator earns a cent
EVERY FEE THE LEDGER CHARGED, PER YEAR, IN DOLLARSburned, not paid to anyone — no validator earns a cent$0$500k$1000k$1500k$2000k$2500k201420152016201720182019202020212022202320242025$2.43m2026$162kThirteen years, 5.06 billion transactions, about $7.6 million in fees. 2026 is running at $162,000 so far.Fees are converted at each month's average XRP price, then summed. 2013 is omitted; price coverage starts mid-year.Source: Bitquery XRP Ledger archive and Bitquery market prices.
Fees rise with congestion, so the peaks mark the years the ledger was busy enough to compete for space. 2026 has more transactions than any year before it and the smallest fee bill since 2016, because almost nothing on it is competing for anything.

Across thirteen years and five billion transactions the ledger has charged about $7.6 million. Its best year was 2025, at roughly $2.4 million, and this year is running well below that. Ethereum burns more than the XRP Ledger's entire history in a normal day.

Set the total burned against the supply and thirteen years of a public blockchain have consumed about one part in seven thousand of its own coin.

12 · How we counted

The traps, and what we could not do

DecisionWhy
We count senders, never receiversDust spam fills millions of wallets with tiny amounts. Receiver counts absorb it; sender counts cannot be inflated by it.
Distributed tables onlyThe archive sits on a two-shard cluster behind a round-robin hostname. Querying the local shard tables returns about half the chain, and a different half each time. Our first pass read 2.51 billion transactions instead of 5.06 billion.
Exact medians, not approximate onesClickHouse's quantile() is a sampling estimate. It put the 2023 median at 31 drops when the exact figure is 39. Every median here uses quantileExact.
Delivered value, never declared amountAn XRPL partial payment lets a sender declare any figure while delivering almost none of it.
Stablecoins filtered by issuer addressCurrency codes are not unique here. Fake RLUSD issuers exist and use lookalike addresses.
Balances measured, not sampledThe current balance of all 8,972,159 accounts, computed in eight partitions and summed. It reconciles against the escrowed supply.
Fees converted month by monthEach month's XRP burn is priced at that month's average XRP price, then summed. 2013 is left out because price coverage starts mid-year.
Not done: StellarStellar was built by one of this ledger's founders for the same purpose and would be the sharpest comparison. Its archive sits on a host this session could not reach, so Tron carries the comparison alone.

The ledger is public and every figure here can be rerun. Bitquery's XRP Ledger API serves the same data these queries ran against, and our money-flow product and address APIs cover the tracing side.

For the same method applied elsewhere, see our Stellar real world asset census, and the tokenised stock market audit, which turns on the same split between activity and value.

13 · The record

The accounts behind the numbers

The XRP Ledger has no route on Bitquery's explorer, so these are listed without links. Any XRPL explorer will find them.

AccountWhat it did
rBTwLga3i2gz3doX6Gva3MgEV8ZCD8jjah12,961,446 transactions in August 2026, a sixth of the whole chain. 12,791,555 orders placed, 882 trades.
rHXgaJLpjDe4jsAiF8atNrRfS6eYrjPYPBPublishes xrp-escrows.com. Ran the escrow phishing bait from 23 August.
rLBSbJgwwVbbrf9eLcK23mWeVSDQwV8kYj2,571,064 dust payments to 108,844 wallets, every day of 2026.
rHjUEuGTbiSc2KowsvATgfFQwP8rWGyMvKPublishes usdxrp.net. Check spam, 1.47 million transactions in 9 active days.
rdGzD9oHEUJkiWxiuVA4UnB9f4kF7giQn725,274 AccountSet transactions in August. AccountSet changes a setting and moves nothing.
rMxCKbEDwqr76QuheSUMdEGf4B9xJ8m5DeRipple's genuine RLUSD issuer. $1.02 billion outstanding.
rLUSDtykL2NVz3HJe1Jqoc7dsxWFVcsmuKIssues a fake token also called RLUSD, from a lookalike address.

The XRP Ledger works. It has never gone down, and it settles in seconds for a fraction of a cent. Thirteen years on, the thing it has not managed to attract is people.

Run it yourself

Ask these questions in plain English

Every figure here came from queries against Bitquery's XRP Ledger archive. With the Bitquery MCP server connected you can ask the same questions without writing SQL: which accounts sent the most transactions last month, what share of payments delivered less than a cent, what the memos on those payments say, and how many of last year's new accounts are still active.

Rank the busiest accounts on any chain for any monthSplit payments by size to separate real transfers from dustRead the memo text attached to any set of paymentsMeasure how many of one year's new accounts are still active
Explore Bitquery MCP Figures measured 8 September 2026 against Bitquery's XRP Ledger archive, 2 January 2013 to 8 September 2026.