Community Trust ScoreVerified
A critical flaw in the Coldcard hardware wallet led to the theft of more than 1,596 Bitcoin — worth over $100 million. The bug hit entropy generation, the process that creates the randomness needed to build secure private keys. And without solid randomness, those keys aren’t really secure at all.
Coldcard’s maker, Coinkite, went public with the flaw on July 31. Attackers had already exploited it across multiple devices by then. What made the bug particularly nasty wasn’t that it broke Bitcoin’s cryptography — it didn’t. The underlying math held. What broke was the randomness feeding into that math. Entropy, in this context, means unpredictability. A private key is only as safe as the randomness used to generate it. Weaken that randomness and an attacker can potentially narrow down the possible keys, guess the right one, and drain the wallet. That’s exactly what happened here, apparently at scale.
Core Lightning developer Dustin Dettmer put forward one theory: the flaw may trace back to firmware changes made in 2021 that could have inadvertently disabled the hardware’s random number generator. If that’s accurate, the vulnerability sat undetected for years. Coinkite has said it plans to release a full technical postmortem, though no date was given for that disclosure. Unclear yet how long that will take.
How Rival Wallets Handle Entropy
The incident put every other hardware wallet maker on the spot. How do they generate entropy? Ledger relies on a certified Secure Element chip with a true random number generator baked in at the hardware level. Trezor takes a different angle — it combines randomness from the device itself with entropy pulled from the host computer, and it runs checks on that entropy before using it. Foundation’s Passport wallet blends randomness from multiple hardware components and publishes open-source firmware so anyone can verify what’s actually happening under the hood.
Three different approaches. All of them, in theory, better than having a single point of failure. But the Coldcard situation showed how quickly theory and practice can diverge. Manufacturers agree entropy matters. They just don’t agree on the best way to protect it — and that debate is getting louder now.
Security expert Jameson Lopp weighed in, noting that weak random number generators have hit various cryptocurrency wallets and libraries before. The specific problem they create is almost invisible: a compromised RNG can still spit out numbers that look random. Standard checks won’t catch it. You’d need targeted testing designed specifically to probe entropy quality, which most users and even many developers never do.
Calls for Independent Audits and Bug Bounties
Nick Percoco, a security expert, pushed for new assurance standards in the wake of the incident. His proposal: independent validation of entropy sources, with certification tied to specific hardware and software versions. Not a blanket stamp of approval for a product line — a version-specific sign-off that would force manufacturers to revalidate every time firmware changes. That kind of process probably would have caught a 2021 firmware change disabling an RNG. Maybe.
Zach Herbert from Foundation has been vocal about open-source development as a security culture, not just a feature. The argument is straightforward — if the code is public, researchers can find problems before attackers do. It’s not a perfect system, but it’s a layer of scrutiny that closed-source firmware simply can’t offer.
Coinkite’s handling of the disclosure drew criticism on both counts. Some developers questioned the company’s past responses to reported vulnerabilities. The absence of a traditional bug bounty program also came up. Bug bounties are pretty much standard practice in software security at this point — they give external researchers a financial incentive to report flaws responsibly rather than sell them or exploit them. Without one, the pipeline for outside scrutiny narrows considerably.
For people who own a Coldcard right now, Coinkite’s guidance is to migrate funds away from at-risk wallets. That’s the immediate fix. Following the migration steps the company laid out is crucial — sitting on a potentially compromised wallet while waiting for the postmortem isn’t a great strategy.
Broader than Coldcard, the incident is pushing a harder conversation about what self-custody actually requires from manufacturers. It’s not enough to build a device that works. The security architecture has to anticipate failure modes, and entropy generation has to be treated as a critical system — not an afterthought buried in firmware. Single points of failure in that chain are, it turns out, catastrophic.
The community is also watching how Coinkite communicates going forward. Transparency around security disclosures has always been uneven across the hardware wallet space. Some companies publish detailed vulnerability reports quickly. Others don’t. The pressure to do better is real now, and it’s coming from developers, researchers, and users who lost funds.
Coinkite’s full technical postmortem is still pending.
Frequently Asked Questions
What caused the Coldcard hardware wallet vulnerability?
The flaw involved a weakness in entropy generation — the randomness used to create private keys. Dustin Dettmer suggested it may trace back to firmware changes in 2021 that could have disabled the hardware’s random number generator.
How much Bitcoin was stolen in the Coldcard exploit?
Attackers stole more than 1,596 Bitcoin, valued at over $100 million, by exploiting the entropy generation flaw across multiple Coldcard devices.
