AI Security Agent Found XRP Ledger Flaw That Could Have Created 18.45 Trillion XRP

An artificial intelligence security agent developed by Veria Labs identified a vulnerability in the XRP Ledger that could have allowed an attacker to create approximately 18.45 trillion XRP in a single transaction. The flaw bypassed protections designed to prevent unauthorised creation of the cryptocurrency, but the XRP Ledger reported no evidence that it had been exploited on a public network.

The vulnerability involved two integer-overflow bugs in the ledger’s transaction-processing software, xrpld. Veria Labs’ agent identified the problem on September 21, 2026, and the research team developed a working proof of concept the following day. The findings were reported through the XRP Ledger’s bug bounty programme on September 22, according to the company’s technical account.

The XRP Ledger’s developers reproduced the vulnerability and classified it as critical. They issued an emergency software release, xrpld 3.4.1, on September 25, before publishing a formal disclosure on October 9. Veria published its detailed account of the discovery on October 10.

The first bug occurred when the payment engine aggregated amounts from multiple offers on the ledger’s built-in decentralised exchange. The software used 64-bit integer arithmetic without adequately checking whether the combined value exceeded the permitted range. An overflowing sum could wrap around to a much smaller number, causing the system to calculate a payment that was substantially lower than the amounts actually credited to recipients.

A second overflow affected the XRPNotCreated invariant, a safeguard intended to reject transactions that increase the total XRP supply improperly. Because this check also accumulated balance changes using vulnerable integer arithmetic, a sufficiently large increase could produce a misleading result and evade the protection designed to detect newly created XRP.

Combining the two weaknesses allowed the researchers to demonstrate how a transaction could distribute XRP to recipient accounts without a corresponding payment from the sender. The proof of concept reportedly generated 18,446,744,073,709.551872 XRP, equivalent to approximately 18.45 trillion XRP. That is about 184 times the original supply of 100 billion XRP.

Veria’s demonstration used multiple accounts to distribute the newly created tokens. The company’s account explains that the setup relied on carefully constructed exchange offers and did not require validator privileges or special access. The XRP Ledger’s disclosure confirms that the researchers demonstrated the behaviour in a controlled environment and verified that the resulting XRP could be spent in a subsequent transaction.

The vulnerability threatened the ledger’s fixed-supply model. XRP’s original supply was established at 100 billion tokens, and transaction fees are burned rather than used to create new XRP. An exploit capable of bypassing that restriction could have undermined confidence in the asset and exposed existing holders to substantial economic damage.

The developers addressed the problem by adding overflow checks to payment aggregation and widening the accumulator used by the XRP supply invariant. They also strengthened other arithmetic operations involving balance calculations. The emergency fix was delivered outside the ledger’s usual amendment activation process, which normally requires support from more than 80% of trusted validators over a two-week period.

Veria Labs reported receiving a $250,000 bug bounty for the discovery on October 8. The reward amount appears in the company’s account of the incident.

The XRP Ledger’s official disclosure states that there was no evidence of the supply-inflation vulnerability being exploited on a public network. The reported 18.45 trillion XRP figure represents the result of a controlled proof of concept, not tokens confirmed to have been created on the public ledger.