XRPL-Utilities Learn

Lesson 2 of 8 XRP Ledger Education Month

How the XRP Ledger agrees without mining

Updated 6 minute read

The short answer: The XRP Ledger doesn't use mining. Trusted computers called validators compare their lists of new payments, and once at least 80% of them agree, the next page is locked in. It takes a few seconds, with no race to burn computing power.
NEXT PAGE LOCKED IN
The next page locks in once 28 of the 35 validators agree. A few can be offline and it still works.

The problem every shared ledger has to solve

In Lesson 1 we pictured the XRP Ledger as a notebook, with copies on computers all over the world. So who decides what goes on the next page?

Here's the catch. Say Alice has 10 XRP, and at the same moment she sends 10 to Bob and 10 to Carol. Only one of those can go through.

AAlicehas 10 XRP 10 XRP10 XRP BBob CCarol COPY IN NEW YORKAlice paid Bob 10 COPY IN TOKYOAlice paid Carol 10 ≠No match
Two payments, but only 10 XRP to spend. If different copies keep different ones, the ledger stops agreeing with itself.

That's called the double-spend problem. A bank solves it by being the boss. A shared ledger has no boss, so every copy follows the same rules instead. Those rules are called a consensus protocol.

How Bitcoin solves it: a race

Bitcoin settles it with a guessing race called mining. The first miner to find a guess that fits writes the next block and gets paid in new bitcoin. A winner turns up about every ten minutes. Lesson 1 shows that race.

To cheat Bitcoin, you'd need more computing power than everyone else put together. That's called a 51% attack. The XRP Ledger can't have one, because there's no mining to take over.

How the XRP Ledger solves it: compare notes

The XRP Ledger skips the race. A group of servers called validators compare notes instead.

Each server keeps a list of validators it trusts, called its Unique Node List (UNL). Most use a recommended list from the XRP Ledger Foundation or from Ripple. On November 9, 2026, both lists named the same 35 validators.

How validators compare notes
  1. CollectEach validator gathers the new payments it has heard about.
  2. ShareIt sends that list to the other validators.
  3. Trim and repeatPayments most others haven't seen wait for the next page. Then they swap lists again.
  4. Lock it inOnce 80% of the trusted validators match, the page is final.
This takes a few seconds, then it starts over for the next page.

With 35 on the list, 28 have to match. Nobody wins a prize, and no new XRP gets made. Every XRP that will ever exist was created when the ledger started.

What if some validators go bad?

A validator could crash, lose its connection, or be run by someone dishonest. Try it: knock some out and close a page.

Try it: knock out some validators

Here are 35 validators, the same number as on the recommended list today. A page locks in once 28 of them agree. Take some offline, or make them cheat, then close a page.

With every validator working, all 35 agree and the page locks in.

The white mark on the bar is 28 out of 35, which is 80%. These are the same rules the real network follows, from xrpl.org.

Trusted doesn't mean in charge

Anyone can run a validator. It only counts for the servers that choose to put it on their list.

Validators also can't move your money. Every payment has to be signed by the account sending it, and every server checks that signature for itself. A payment without a proper signature never goes in, even if every validator wanted it to.

What validators do is agree on which signed payments go on the next page, and in what order. That's the whole job.

See it for yourself

Here are the newest pages of the ledger. Each one was locked in by the validators the same way you just saw.

The XRP Ledger right now

Loading the latest pages...

    These are real pages, pulled live from the public ledger. See how each page lists the fingerprint of the one before it? That link is the "chain" in blockchain.

    You can also watch the validators themselves. Open the validator list on livenet.xrpl.org to see who they are and whether each one agreed on the latest page.

    Common questions

    Do validators get paid?

    Not by the ledger. No new XRP is created, and transaction fees are destroyed rather than handed out. The people and groups who run validators usually do it because they rely on the network working.

    Is this the same as proof of stake?

    No. Proof of stake gives a bigger say to whoever locks up more coins. On the XRP Ledger, owning more XRP doesn't get you a bigger vote.

    Does Ripple control the validators?

    No. Ripple publishes one of the two recommended lists, and the XRP Ledger Foundation publishes the other. Each validator is run by its own operator, and any server can pick its own list.

    What happens if the validators can't agree?

    The network pauses and adds nothing until they can. It's built to stop rather than write down something wrong.

    What if one validator keeps disagreeing?

    It gets outvoted. The page goes in with the validators that match, and the odd one out has to drop its version and take the agreed page. If it keeps happening, servers can take it off their trusted list. The recommended lists also drop validators that aren't reliable.

    Why 80%?

    It's high enough that a few broken or dishonest validators can't push a bad page through. It's also low enough that the network keeps going when a few of them go offline.

    Words used in this lesson

    • Double spend: trying to spend the same money twice.
    • Consensus protocol: the rules a shared ledger follows so every copy agrees.
    • Mining: Bitcoin's way of agreeing, a guessing race to write the next block.
    • Validator: a server that helps decide which payments go on the next page.
    • Unique Node List (UNL): the list of validators a server trusts.
    • Recommended list: a ready-made UNL that most servers use.
    • Proposal: a validator's list of payments it thinks should go on the next page.
    • Validation: a validator's signed copy of the finished page.
    • Supermajority: at least 80% of the trusted validators.

    Sources