Bitcoin cold-wallet attack affects 4,585 addresses, drains 1,367 BTC
The Bitcoin cold-wallet attack has spread to 4,585 addresses, with a total of 1,367 bitcoin observed drained and aggregate losses nearing $89 million. Those totals represent observed activity across all three waves of the incident, which together affected thousands of addresses. One observed tranche drained 208 bitcoin from 1,912 addresses between Friday midday and Saturday morning UTC as part of the incident’s early activity.
The flaw traces to a March 2021 Coldcard firmware build that routed seed generation to a predictable software randomiser instead of the chip’s hardware one. That rerouting produced a bounded set of possible seeds and therefore a bounded set of possible private keys. The limited key space is an attribute of the firmware change rather than the hardware. The description locates the vulnerability in the device’s seed generation pathway.
The practical consequence is that the set of possible keys could be reproduced without accessing a physical device when the underlying disclosure is known. “The flaw traces to a March 2021 firmware build that routed seed generation to a predictable software randomiser instead of the chip’s hardware one, leaving a bounded set of possible keys that anyone with the disclosure and enough compute can reproduce offline, without ever touching a device.” The quoted wording identifies reproduction offline as feasible given disclosure and compute. The statement locates the security failure in firmware-level entropy sourcing.
The account confines the technical root cause to firmware-level randomness routing. The reproduced keyspace and the offline reproducibility are the implications specified in the quoted statement. These sentences restate the technical characterization above.
The three observed waves of the Bitcoin cold-wallet attack together account for 1,367 bitcoin drained from 4,585 addresses, with aggregate losses nearing $89 million. The July 30 opening wave moved 1,083 bitcoin from 1,196 addresses in 41 minutes, averaging close to a full coin per address. One observed tranche drained 208 bitcoin from 1,912 addresses between Friday midday and Saturday morning UTC. These figures represent observed activity across all three waves of the incident.
The attack batched an average of six victims into each sweep, whereas wave one took exactly one at a time. Wave three sends each victim’s coins to its own destination rather than the handful of shared collector addresses that made the first two easy to map, and parks them in pay-to-witness-script-hash outputs. The first two waves used a handful of shared collector addresses, which made those waves easier to map on-chain. Galaxy said it is confident each wave is internally one operator, and will not link the three.
The sweeping activity continued almost three days after initial observations, and the falling average haul reported by observers indicates the most profitable portion of the exploitable key space has already been picked over. The incident has implications for Bitcoin wallet security because it traces to a firmware-level randomness routing issue described earlier in this report.
