Can the XRP Ledger Be Frozen? What a New Network Study Found

A new academic study looked at the XRP Ledger from an unusual angle. Instead of modeling an attacker who hacks wallets or targets validators directly, the researchers asked what happens if someone removes the network nodes that hold the system’s communication together. In one scenario, removing only about 11% to 12% of the most strategically important nodes was enough to stop validators from reaching consensus.

Before that number travels too far, the video adds the context it needs. The study used network data from 2022, ran simulations rather than live tests, and found a simple change that made the model network about three times harder to disrupt.

How does XRP Ledger consensus work?

The XRP Ledger has no miners and doesn’t work like a typical proof-of-stake chain. It relies on trusted validators. Each server chooses a set of validators it trusts, called its Unique Node List (UNL). The network moves forward when enough trusted validators agree on the next ledger.

According to XRP Ledger documentation cited in the video:

• If fewer than about 20% of trusted validators are faulty, consensus continues normally.
• If more than 20% but fewer than 80% are faulty, the network may stop making progress rather than accept an invalid ledger.
• Validators exchange proposals until a supermajority agrees, with an 80% final agreement threshold.

That design is deliberate. The network would rather stop than accept a bad state.

What happens to your XRP if consensus stalls?

Your XRP doesn’t disappear, your wallet isn’t hacked and ledger history isn’t rewritten. What fails is forward progress: new transactions stop being finalized until enough of the network can communicate and agree again.

The video stresses this is still serious. Payments don’t settle and applications that depend on new ledgers can’t move forward. Availability is part of security too.

What did the study test?

Validators need the peer-to-peer network underneath them to talk to each other. That network carries transactions, ledger data, consensus proposals and validation messages. If enough key pathways disappear, some validators can no longer reach others, and once the connected group drops below the required threshold, consensus can stall. That’s the attack surface the researchers modeled.

Their setup, according to the video:

Model detail	Value
Data source	Real XRP Ledger topology data collected in 2022
Nodes	952
Connections	15,070
Average connections per node	About 31.7
Validators	34 simulated, matching the recommended UNL size at the time

Because the dataset didn’t identify which nodes were validators, the researchers simulated them. They also assumed attackers couldn’t target validators directly, since validators may have stronger defenses. Instead, the model removed ordinary nodes based on how important they were to connectivity.

How the attacks were modeled

The researchers tested two kinds of targeted attack:

1. By degree: remove the nodes with the most connections first.
2. By betweenness centrality: remove the nodes that sit on the largest number of shortest paths between other nodes. The video compares these to major highway intersections: a lot of traffic passes through them even if they don’t have the most roads.

In the baseline model, the quorum failure threshold was about 11% for the degree attack and about 12% for the betweenness attack. Remove that share of nodes in that order, and validator connectivity fell below what consensus needs.

Does this mean XRP is centralized?

Not necessarily, the video argues. Network topology and decentralization are related but different. A network can have many independent participants and still develop highly connected hubs, as happens across the internet with major routers, data centers and cloud providers. XRP Ledger documentation also notes that each server can choose its own validator list, even though default configurations commonly use a recommended list to keep overlap high.

The better question, according to the video, is how resilient the network topology underneath consensus is if important hubs disappear.

The fix: a few random connections

This is the part of the study the video considers most important. Instead of rewiring the network, the researchers added a small number of random peer connections.

With just two extra random connections per participating node, at 60% participation:

Attack type	Baseline threshold	With random connections
High degree	About 11%	About 38%
High betweenness	About 12%	About 33%

That is nearly three times the modeled resilience. The video uses a bridge analogy: if all traffic must cross one big bridge, destroying it stops everything. Add several smaller bridges in random places and the big one is no longer the only route that matters.

The method also compares well with a more aggressive approach studied earlier, which removes existing connections and replaces them. At higher participation, about three random connections per node matched or beat roughly 20 to 25 rounds of that rewiring, while leaving the existing network mostly intact. No central coordinator is needed to redesign the network.

The limits of the study

The video is clear about what the paper does not show:

• The network data is from 2022. Nodes, peer relationships and validator setups have changed since.
• It is a simulation, not a live test of today’s network.
• The real XRP Ledger already uses peer discovery, a gossip-style process for finding new peers, and important validators can use private peers and other setups to improve resilience against denial-of-service attacks.

So the claim “researchers proved XRP can be shut down by attacking 12% of nodes” isn’t supported. The authors themselves say real-world evaluation is still needed, and they released their simulation framework so others can reproduce the work.

What to watch next

The video lists four open questions:

1. Will anyone map the current 2026 XRP Ledger topology and rerun the experiment?
2. Will XRP Ledger infrastructure operators or developers respond to the paper?
3. Will some form of random peer augmentation be tested in real conditions?
4. How much resilience can be added without creating new problems with bandwidth, privacy or server limits?

Its conclusion is neither “XRP is broken” nor “there’s nothing to worry about.” The researchers found a potential weakness in an old model and a cheap way to reduce it. How resilient today’s network is remains unknown until someone tests it.