Coldcard flaw drains $40M in BTC

TLDR

A Coldcard Bitcoin hardware wallet firmware bug let attackers guess wallet seeds and steal tens of millions of dollars in BTC from hundreds of single signature wallets.

1. Reports say about 594 BTC (~3840M USD) were drained within roughly 25 minutes, with later analysis suggesting losses exceed 1,000 BTC (~70M USD).
2. The flaw affects seeds generated on specific Coldcard firmware since 2021, mainly single signature wallets without strong BIP 39 passphrases, and it cannot be fixed without creating a brand new seed.
3. The incident shakes confidence in self custody hardware wallets, highlights rising AI driven exploit risk, and may push some Bitcoin holders toward multisig setups or regulated custodial and ETF alternatives.

Deep Dive

1. Exploit Scale And Mechanics

Security reports indicate that an attacker exploited a Coldcard firmware bug to sweep around 594 BTC from roughly 500 single signature wallets between 01:31 and 01:56 UTC, moving funds in a tight three block window and consolidating most into a single address that has not moved yet. The bug caused vulnerable devices to bypass their hardware random number generator and use predictable software based key generation seeded by chip details like serial number and internal clock, turning supposedly unguessable seeds into solvable puzzles.

Further tracing by Galaxy Research found that nearly 1,200 addresses were drained for about 1,082.65 BTC (~70.2M USD), suggesting the total impact is larger than initial estimates and that the attack pattern was automated and highly optimized. Chainalysis reports that approximately 30M USD was stolen in the first 10 minutes, with larger balances targeted first.

2. Affected Users And Mitigation

The vulnerability primarily affects seeds generated on Coldcard Mk3 firmware 4.0.1 and later, and certain Mk4, Mk5 and Q versions before emergency patches, where entropy was reduced from a robust level to around 4072 effective bits, making brute forcing feasible for a determined attacker. Incident notes highlight that seeds created with a strong BIP 39 passphrase or substantial extra entropy from dice rolls are much less exposed, and multisig wallets only help if none of the keys were ever generated on compromised devices.

Critically, updating firmware does not repair seeds already created on affected versions; a weak seed remains weak even if imported into another brand of wallet, so mitigation requires generating a new seed on patched or unaffected hardware and migrating funds carefully. Coinkites advisory and follow up posts recommend upgrading to fixed firmware, creating a fresh seed, backing it up, sending small test transactions, and only then moving larger balances.

3. Broader Self Custody Impact

Commentary from wallet makers and security researchers notes that a single firmware bug in a widely used hardware wallet undermined a core narrative of Bitcoin self custody, that offline devices are practically immune to remote theft. Coinkites CEO has accepted responsibility and suggested the flaw may have been uncovered using AI assisted code review, while rival vendors stress that AI tools can find latent bugs in public code faster than traditional audits and that this changes the threat model for all open source wallets.

Despite the size of the theft, Bitcoins price has shown limited immediate reaction, but the incident is already fueling debate around whether everyday users should rely more on multisig setups, institutional custodians, or spot Bitcoin ETFs, while treating single signature hardware wallets as a high responsibility tool rather than a simple consumer product. Confidence: high because multiple independent security reports and the wallet makers own advisories agree on the exploit and affected firmware.

Conclusion

A Coldcard randomness bug turned some seeds from mathematically unguessable into practically brute forcible, allowing attackers to steal tens of millions of dollars in BTC from hundreds of wallets in minutes. The episode shows that self custody is powerful but unforgiving, and that hardware wallet security depends not just on offline storage but on high quality firmware, independent review, and user practices like passphrases and diversified setups. Going forward, the most important signals are further advisories from Coldcard and security firms, evidence of copycat attacks, and whether users shift toward more resilient custody patterns such as multisig or regulated ETF exposure.