Zilliqa Confirms 683M ZIL Theft in Ledger App Failure

Zilliqa Post-Mortem Confirms 683M ZIL Theft

The loss-impact check has a clear distinction: 683,130,969.66 ZIL is the project’s count of proven theft, not simply suspicious movement or an estimate based on an attacker’s wallet balance. Zilliqa said it identified the amount in 66 successful transactions during the 2026 attack window, while 12 failed attempts moved nothing.

That theft total is not the same as the 6,772 exposed accounts. Zilliqa said 6,763 keys were reconstructed and verified against public chain data, while nine more accounts were classified as compromised because they were drained but had incomplete historical-signature records. Only 51 of the exposed accounts were drained in the published record.

The project said 679.2 million ZIL went directly to three consolidation wallets from those 51 accounts. It traced about 80.6% of the proven theft to a single exchange receiving wallet, said the exchange account used to liquidate part of the funds had been frozen, and cautioned that recovery was unlikely and could take a long legal process.

The incident belongs to the same self-custody risk category as the Coldcard seed-generation failure, but the attack path is different. In this case, Zilliqa says an attacker could reconstruct a legacy account key from signatures already written to the chain, without touching the physical device, stealing a recovery phrase or tricking the holder.

Ledger App’s Nonce Flaw Exposed Legacy Keys

Zilliqa said the app generated a random signing nonce correctly, then copied the wrong 32 bytes into the buffer used to sign. Eight bytes of zero padding were retained and eight bytes of real entropy were discarded, leaving the top 64 bits of every nonce fixed at zero.

In plain terms, that regularity gave each affected signature information about the account’s private key. Zilliqa said four or more legacy transactions signed through the app could let a key be reconstructed in seconds on ordinary hardware from public blockchain data. The protocol’s attributed explanation is unusually direct: “The exposure is created by the act of signing, and it is permanent.”

The scope is narrower than a general Ledger-device compromise. Zilliqa said only legacy, non-EVM transactions signed through the affected app were exposed. Its incident guidance says seed phrases were not exposed, Zilliqa EVM transactions were outside the signing path, and assets on other blockchains on the same device were not at risk from this issue.

That separation matters for readers weighing broad hardware-wallet claims. A previous Ledger and Trezor malware report involved a separate Windows threat designed to scrape wallet-related data. Zilliqa’s finding concerns a historic application-level signature defect, not proof that every Ledger wallet or every ZIL transaction is vulnerable.

Zilliqa Keeps Legacy Transfers Frozen for Migration

The halt is central to the user impact. Under normal circumstances, moving funds could be the standard response to an exposed key. Zilliqa said that would not work safely here because an attacker who had reconstructed the key could watch the chain and front-run the transfer. The protocol therefore disabled legacy transactions on July 20, preventing movement by holders and attackers alike.

Zilliqa is retiring the legacy side rather than reopening it, with Zilliqa EVM set to become the only production environment. It has published an address checker and said a migration tool would undergo an external security audit, with the report expected around the beginning of September. Several exchanges have agreed to support migration, but Zilliqa did not publish a final recovery date for individual legacy holders.

The threat model also explains why official channels matter. The project says it will never ask for a recovery phrase, private key or seed phrase, and holders should treat unsolicited recovery forms and direct messages as hostile. That caution has become a recurring theme after incidents such as the SafePal customer-order data breach, where public disclosures can create fresh phishing opportunities even when wallet keys are not part of the confirmed exposure.

Broader market sentiment was risk-on despite the protocol-specific restrictions. Alternative.me’s Crypto Fear and Greed Index read 74, or Greed, in its latest update on Aug. 26.

The next proof point is the migration and recovery framework, including its audit result, exchange rollout and any further tracing update. For now, the strongest supported account is Zilliqa’s own: a legacy app flaw exposed 6,772 accounts, produced 683.13 million ZIL in proven theft and left all legacy transfers frozen while the network moves to its EVM environment.