Canada Accounts for 25% of Attributed Losses in Coldcard Wallet Exploit; Total Theft Hits $116M

Canadian Bitcoin holders represent the largest identifiable group affected by the ongoing Coldcard hardware wallet exploit, accounting for 25% of all losses that can be attributed by geography. Analysts link the concentration to Coldcard parent Coinkite's strong domestic presence; the company is headquartered in Toronto. Chainalysis visualization data places Australia as the next most impacted market, responsible for roughly 15% to 20% of total damages. The United States and Thailand follow, each in the 10% to 15% range. While English-speaking, early-adopting Bitcoin jurisdictions absorbed the heaviest losses, the incident has produced broad global fallout across Western Europe, Latin America, and major African crypto hubs including Nigeria and South Africa. Total assets stolen in connection with the exploit have reached $116 million. Galaxy Research traced the root cause to a March 2021 firmware update, citing the introduction of a new random number generator as the single point of failure that enabled the attack. Galaxy said the RNG was wired incorrectly and silently fell back to a weaker option, leaving users unaware that their private keys were being generated with low entropy. Galaxy reported that five years later, an attacker swept $70 million from 1,200 wallets in 41 minutes. Natalie Newson, Senior Blockchain Investigator at CertiK, said standard checks failed to catch the issue because of a specific configuration error: MICROPY_HW_ENABLE_RNG was set to zero. Newson explained that for a static guard checking #ifndef, a macro set to 0 is still considered defined, causing the safety check to evaluate as true, suppress the #error guard, and allow the build to proceed as if configurations were correct. On remediation, Newson urged wallet manufacturers to prevent silent fallbacks to software pseudorandomness by removing production fallbacks and designating a single approved RNG provider. She added that the full pipeline from entropy acquisition to seed generation should sit within a NIST FIPS 1403-defined validation boundary. Newson also warned that incident response during active automated sweeps must prioritize user communication alongside technical testing. She said teams should rapidly disclose scope, identify affected users, and provide clear mitigation steps while thoroughly validating any fix before release, arguing that transparency is as critical as the patch. For nontechnical users worried about bricking devices during emergency firmware updates, Newson recommended a strict sequence: obtain a trusted hardware wallet, generate a new seed phrase offline, validate it with a small test transaction, move remaining funds to the verified setup, then attempt firmware updates on the original device. She also advised against single points of failure by splitting holdings across hardware wallets from different manufacturers. The episode is intensifying debate over self-custody security assumptions. Critics argue that the rapid draining of dormant, long-term holdings shows offline execution alone does not guarantee protection. Nanak Nihal Khalsa, cofounder at Human.tech, said the incident highlights persistent third-party risk within hardware ecosystems, arguing that even self-custody involves outsourcing trust and warning that AI-assisted exploit vectors could amplify threats. Newson echoed concerns about single-signature setups, saying broader adoption requires designs with graceful degradation so a single user or vendor mistake cannot wipe out life savings. She characterized singlesig self-custody as offering no margin for error, because users rely on the device, its code and dependencies, and QA to catch failures. As a result, industry thinking is shifting toward multi-vendor, multisignature, or threshold signature (MPC) arrangements as a baseline. Newson said the aim is to move beyond trusting a single device so that no single compromised component or actor can move funds, with signing keys or threshold shares distributed across independent organizational and technological failure domains to prevent any one provider from reconstructing the key or authorizing transactions alone.