Coldcard Hack Explained: How a 2021 Firmware Bug Enabled More Than $130 Million in Bitcoin Theft

How the seed-generation flaw enabled remote theft

A recovery seed is effectively the master secret from which a Bitcoin wallet’s addresses and signing keys can be regenerated. It must be unpredictable: if attackers can narrow the possible seed space, they can search for the correct one without ever seeing the original device.

For affected Mk3 wallets, reporting based on Coinkite’s technical analysis put effective entropy at roughly 40 bits instead of the intended 128 bits. That reduced the number of possibilities enough for specialized computation to become practical.

The reported attack sequence was:

1. Researchers and attackers identified the predictable seed-generation behavior.
2. Software generated large numbers of candidate seeds based on the flawed process.
3. Candidates were checked against publicly visible Bitcoin addresses.
4. When a candidate matched a funded wallet, the corresponding private keys could be used to authorize a transfer.

This explains why victims could follow sensible self-custody practices—never sharing a seed, keeping the device offline, and storing it securely—and still lose funds. The weakness was present at seed creation, not in the later storage location.

How much cryptocurrency was stolen?

The reported totals changed as the campaign developed. An early accounting attributed 594.48 BTC to roughly 500 single-signature wallets, while later reporting linked the campaign to more than 1,300 BTC and losses above $130 million.

Those figures should not be treated as one fixed final total. Different reports covered different waves of theft, time periods, wallet types, and Bitcoin prices. Galaxy Research’s later assessment described the losses as “$130m and counting,” underscoring that the investigation was still developing by mid-August.

Who appears to be behind the Coldcard theft?

The perpetrators have not been publicly identified. Available reporting described an active, opportunistic campaign involving at least a dozen unidentified hackers or groups, with the possibility that multiple actors were exploiting the same weakness after it became known.

That makes attribution different from proving the root cause. The technical cause is tied to Coldcard’s seed-generation implementation, while the identities, organization, and location of the people draining wallets remain unresolved in the available reporting. The evidence does not establish a breach of Bitcoin itself or a single confirmed criminal group.

How this fits the wider rise in crypto hacks in 2026

The Coldcard incident arrived during a broader period of heavy cryptocurrency security losses. By early August, reporting cited more than 200 hacks of crypto companies in 2026 and more than $950 million in losses.

The comparison needs context: the wider figure covers attacks across crypto companies and systems, while the Coldcard losses came from a specific weakness affecting wallet-seed generation. Still, the incident illustrates a recurring self-custody risk: strong cryptography and offline storage cannot compensate for a flawed key-generation or implementation layer.

Why Coldcard was considered safer than a hot wallet

Hardware wallets are intended to keep sensitive signing material isolated from internet-connected devices. A hot wallet, by contrast, typically operates in software that is exposed to a broader range of online threats. Cold storage therefore reduces the risk of malware, remote account compromise, and accidental exposure of a seed during ordinary use.

That security model did not fail because attackers remotely logged into every Coldcard. It failed because some wallets had been created with seeds that were not sufficiently random. Once a weak seed exists, keeping the device in a safe protects the hardware—but not necessarily the mathematical secret that can be reconstructed from the public blockchain and the known generation pattern.

The distinction matters: this was a product implementation vulnerability, not evidence that every offline wallet is inherently insecure or that Bitcoin’s address and signature system has been defeated.

What victims experienced

Jonathan Goodman said approximately C$1.6 million was taken from his Coldcard-controlled holdings. He said he had never shared his seed phrase and had kept his devices physically secured, yet those precautions did not prevent the theft.

His account captures the most unsettling part of the incident: operational security was not enough when the secret itself had been generated from an insufficiently unpredictable process. A carefully protected copy of a weak key is still a weak key.

What Coinkite advised users to do

Coinkite disclosed the issue on July 30, 2026, and published fixed firmware and a security advisory. Its core remediation was migration, not merely updating:

1. Install the appropriate fixed firmware from an official Coinkite source.
2. Create an entirely new wallet seed on the updated device.
3. Move funds from the old wallet to addresses controlled by the new seed.
4. Treat the old seed as compromised unless it meets Coinkite’s stated exceptions.

Coinkite specifically warned that updating firmware does not retroactively make a seed generated on vulnerable firmware safe. Existing funds must be moved to a wallet controlled by a newly generated seed.

The advisory said funds were at risk when the seed had been created without at least 50 independent, private dice rolls and the wallet was not protected by a strong, unique BIP-39 passphrase. Users should follow the current official advisory for model-specific firmware versions and migration instructions, and should never provide their seed words or passphrase to anyone claiming to offer support.

The broader lesson for hardware-wallet users

The Coldcard case shows why hardware-wallet security has several layers:

• Physical isolation: keeping the device and backup away from attackers.
• Randomness: generating a seed with enough unpredictable entropy.
• Firmware integrity: ensuring the device’s code uses the intended security components.
• Recovery hygiene: replacing an affected seed rather than assuming a software update repairs it.
• Transaction security: verifying destination details before authorizing transfers.

The most important practical point is simple: if a wallet seed may have been generated through a vulnerable process, the seed—not just the device—must be replaced. In this incident, the attackers reportedly did not need to steal a Coldcard. They only needed to reproduce the secret it had generated.