More than 1,000 bitcoin, worth approximately $70 million, were drained from 1,196 Coldcard hardware wallets in a 41-minute window on July 30, according to an analysis by Galaxy Research. The figure is nearly double the amount reported when the theft first surfaced.

The incident stands out from typical cryptocurrency thefts because the attacker never needed access to the victims' devices. A firmware flaw in certain Coldcard hardware wallets made supposedly unguessable seed phrases computationally enumerable, enabling the attacker to reconstruct private keys entirely offline and sweep funds at scale.

Coldcard, manufactured by Toronto-based Coinkite, is a Bitcoin-only hardware wallet favored by self-custody advocates for its air-gapped design and open-source firmware. Its user base skews toward users who prioritize sovereignty over their private keys, which makes a seed-generation failure particularly consequential for a community that chose the product specifically for its security posture.

Security firms warn that additional wallets remain at risk because owners cannot reliably determine whether their seeds were generated on vulnerable firmware. The only reliable remediation is to generate a new seed on a device confirmed to use proper randomness and transfer all funds to addresses derived from that seed. Investigators are actively tracing the attacker through logs obtained from a blockchain data provider.

Galaxy Research Maps the Full Event

Galaxy Research published its full mapping of the event on Friday, finding that 1,082.65 BTC were swept between 01:10 and 01:51 UTC across six blocks. Three intervening blocks contained no related transactions, suggesting the transfers were broadcast in batches rather than continuously.

The proceeds currently sit across four addresses and have not moved. Early reporting captured only one of those addresses, which explains why the initially reported loss figure has since grown.

While the total value stolen is smaller than some of the largest attacks this year, the mechanism itself is what makes this exploit particularly significant.

Why the Coldcard Exploit Is Unusual

Most cryptocurrency theft involves some form of direct access: an exchange is breached, a smart contract is exploited, or a private key is phished from a user's computer. The core defense has always been distance — which is exactly what a hardware wallet is designed to provide. By keeping the key on a device that never connects to the internet, there is, in theory, nothing for an attacker to reach.

When a wallet is created, the device is supposed to select a number so large and unpredictable that guessing it is effectively impossible. That number serves as the seed, from which every address and private key is derived using fixed, publicly known rules defined by BIP39 and related standards.

Coldcard's firmware was designed to draw that number from a dedicated hardware randomness generator. However, an internal build setting instructed the device to bypass that generator. A validation check in a supporting library tested only whether the setting existed, rather than whether it was actually switched on.

As a result, key generation fell through to a basic software substitute seeded from the chip's serial number and its clock registers. The serial number is fixed factory metadata, and clock values are timing state that an attacker can narrow down or measure on a device of their own.

The consequence was that the range of keys the device could ever produce collapsed from an unimaginably vast space to one that was countable. Random number generator failures have been a recurring vulnerability class in cryptocurrency: a 2013 Android Secure Random bug compromised Bitcoin wallet keys on affected devices, and similar weaknesses have periodically surfaced across the ecosystem since.

Security teams determined that key generation could be fully enumerated on the older Mk2 and Mk3 models. On the Mk4, Q, and Mk5 models, they estimated the range at roughly four billion possibilities. Four billion is a large number for a human but a trivially small one for a computer.

An attacker can generate candidate seeds on their own hardware, derive the addresses each seed would produce, and check those addresses against the public blockchain — which anyone can download. Every step of this process runs on the attacker's machine. The victim's device is never involved and could be powered off in a safe on another continent.

Galaxy's analysis confirmed this process in action. Of the drained wallets, 1,183 used the modern native segwit address format, seven used an older standard, and six used an even older one. The fact that funds were targeted across three address formats simultaneously rules out individual victim selection — it points to systematic enumeration, with each candidate seed checked against every derivation path it might have produced.

The operator can widen the search, refine it, and return at any time. Galaxy warned that further waves of theft are likely if affected owners do not move their funds.

Critically, there is no test an owner can run against their own wallet to determine whether their seed falls within the reproducible range. Until the issue is fully resolved, anyone who generated a seed on affected firmware must assume the worst.

Affected Models and Vendor Response

Coinkite, Coldcard's manufacturer, has warned Mk3 owners and stated that its newer devices are unaffected. However, Block's investigation places the Mk2, Mk4, Q, and Mk5 models in scope as well. This discrepancy means the full extent of affected devices remains unresolved, and the gap between the vendor's and investigators' assessments is itself a risk factor for users deciding whether to act.

The Attacker's Mistake

The attacker did leave one trail. Block's Clay Garrett said on X that the operator used a paid account at a well-known blockchain data provider to query the source addresses during the sweeps. According to Garrett, the provider's internal logs matched the suspected workflow with what he called extraordinary specificity — down to the number, timing, and sequence of requests.

1/ During our investigation of the Coldcard drain yesterday, we identified an unusual pattern in the sweeps. That pattern led us to a hypothesis that has since been confirmed: the operator used a paid account at a well-known blockchain-services provider to query the source… — Clay Garrett (@clay_garrett) July 31, 2026

The provider appears to have been supplying ordinary services to requests that gave no indication of their purpose. Block has passed the information to authorities.

Broader Implications for Cold Storage

Cold storage has always promised that a private key is unguessable. The industry widely interpreted that as a guarantee that a key is also unreachable. The Coldcard exploit demonstrates that these are two separate properties, and the cost of finding and exploiting weaknesses in the first continues to fall.

In a separate development highlighting the evolving threat landscape, Anthropic published research on Tuesday showing that one of its AI models halved the security of a candidate post-quantum cryptographic algorithm in 60 hours, against a design that had previously survived two years of expert review.

Storing a key safely is now the easier half of the problem. Ensuring it was generated correctly in the first place may be the harder one.

Source: CoinDesk