Cold Storage Turned Hot Target: How a Code Flaw Unlocked $130 Million in Bitcoin

The Irony of Secure Hardware

At DailyTechWire, we've tracked dozens of cryptocurrency breaches over the past year, but this one stands apart. More than $130 million in Bitcoin has been drained from users who did everything the security playbook told them to do: store keys offline, use hardware wallets, never touch the internet. The target was Coldcard, a device made by Coinkite and widely regarded as one of the safest places to park digital assets. Yet at least a dozen separate attackers managed to empty wallets by exploiting a single weakness buried in the device's code, one that made supposedly random seed phrases entirely predictable.

According to Galaxy Research, the thefts are ongoing, with multiple hacker groups independently discovering and exploiting the same vulnerability. Elliptic, a blockchain-intelligence firm, confirmed the loss estimate is broadly accurate. The flaw dates back to 2021, a single line of code that undermined the cryptographic foundation of every wallet generated during that window.

How Cold Storage Was Supposed to Work

The entire value proposition of a hardware wallet like Coldcard rests on isolation. Bitcoin itself always lives on the blockchain, a public ledger anyone can read. What makes it yours is a secret key, or more commonly a twelve- or twenty-four-word seed phrase that derives that key. If you store that phrase on a device connected to the internet, a laptop or a mobile app, you're vulnerable to malware, phishing, supply-chain attacks. A hardware wallet keeps the phrase on a chip that never touches a network. Even if your computer is compromised, the theory goes, the key remains locked inside the device.

Coldcard took this further by marketing itself as "air-gapped," meaning it could generate and sign transactions without ever plugging into a computer. Users could transfer unsigned transactions via SD card, sign them on the Coldcard, then broadcast the signed result. It was a fortress model: the seed phrase generated on the device, stored on the device, never exposed.

The Flaw: Predictable Randomness

Security researchers at Block, the payments company formerly known as Square, identified the issue. Coldcard's firmware included a flaw in how it generated entropy, the randomness needed to create a seed phrase. Instead of pulling from a truly unpredictable source, the code used a method that produced patterns. Those patterns meant that if you knew the rough timeframe a wallet was initialized and had access to the flawed algorithm, you could brute-force the seed phrase. You didn't need to steal the device, crack the PIN, or intercept any communication. You just needed to generate keys until you found one that held funds.

This is the nightmare scenario for any cryptographic system. The safe is intact, the locks are strong, but the locksmith used a key-cutting machine with only a few dozen templates. Attackers reverse-engineered the templates and started trying every door.

Jonathan Goodman, a Coldcard user who claims to have lost $1.6 million, described the betrayal in stark terms. He stored his seed phrase in multiple physical safes and safety deposit boxes. He never typed it into a computer. He followed every recommendation. None of it mattered, because the seed phrase itself was compromised the moment the device created it.

Scale and Attribution

Blockchain-monitoring firms have identified at least a dozen distinct wallet addresses receiving stolen funds, suggesting multiple actors rather than a single coordinated group. This is consistent with how vulnerabilities spread in the hacker economy: someone discovers a flaw, writes an exploit tool, and either sells it or shares it in closed communities. Within weeks, dozens of operators are running the same playbook.

TRM Labs, another blockchain-intelligence firm, counts more than 200 hacks targeting cryptocurrency platforms and users so far this year, totaling over $950 million in losses. The Coldcard incident accounts for roughly fourteen percent of that total, making it one of the largest single-vulnerability heists in recent memory.

What distinguishes this from exchange hacks or DeFi protocol exploits is the target profile. Coldcard users skew toward high-net-worth individuals and institutions who prioritize security over convenience. These are not retail traders leaving funds on Binance; they are users who invested in dedicated hardware and operational security. The breach cuts at the credibility of the entire cold-storage category.

Coinkite's Response

Coinkite published an advisory alerting users to the vulnerability and urging them to update their firmware immediately. The company also recommended that anyone who initialized a wallet during the affected period migrate to a new seed phrase, a process that involves generating a fresh wallet and transferring all funds. The advisory was first posted on Thursday and updated over the weekend as the scale of the thefts became clearer.

Coinkite did not respond to requests for comment. The company has not publicly disclosed how many devices were sold during the vulnerable period, nor has it offered a timeline for when the flaw was introduced or internally discovered. The lack of transparency has frustrated users, many of whom are now questioning whether other hardware-wallet manufacturers face similar issues.

Implications for Hardware Security

The Coldcard breach exposes a structural problem in the hardware-wallet industry. These devices are marketed on the promise of trustlessness: you don't need to trust an exchange, a cloud provider, or even your own computer. But you do need to trust the device manufacturer to implement cryptography correctly. And unlike open-source software, where thousands of eyes can audit the code, hardware wallets often run proprietary firmware with limited external review.

Coldcard's firmware is partially open-source, but the vulnerability persisted for years before independent researchers flagged it. That lag suggests the auditing process, even in relatively transparent projects, is not sufficient to catch subtle entropy flaws. The problem is compounded by the fact that users have no way to verify randomness after the fact. A seed phrase looks random whether it came from a secure process or a broken one.

This also raises questions about the broader hardware-wallet ecosystem. Ledger, Trezor, and other manufacturers use different architectures and entropy sources, but all face the same fundamental challenge: generating unpredictable numbers in a deterministic environment. If one major player can ship a flawed random-number generator for years, others might too.

The Economics of Theft

From the attackers' perspective, the Coldcard flaw was a golden opportunity. Unlike ransomware or phishing, which require social engineering and operational overhead, this exploit was purely computational. Run the seed-phrase generator in a loop, check each resulting address against the blockchain, transfer any funds found. The entire process could be automated and scaled across thousands of addresses.

The blockchain's transparency worked in the hackers' favor. Every Bitcoin address and its balance are public, so attackers didn't need to guess which wallets were worth targeting. They could scan for high-value addresses, generate candidate seed phrases, and test them in order of expected return. It's the digital equivalent of checking every safety deposit box in a city because you know the lock manufacturer used a faulty tumbler.

The stolen funds are now moving through mixers and cross-chain bridges, standard techniques for obfuscating the trail. Elliptic and other firms are tracking the flows, but recovery is unlikely. Once Bitcoin is mixed and converted to privacy coins or moved through decentralized exchanges, attribution becomes nearly impossible.

What Comes Next

For Coldcard users, the immediate priority is migration. Anyone who generated a wallet on an affected device should assume the seed phrase is compromised and move funds to a new wallet with a freshly generated phrase on updated firmware. That process is tedious and carries its own risks; transferring large amounts of Bitcoin requires careful attention to transaction fees, address verification, and operational security.

For the hardware-wallet industry, the incident is a credibility test. Manufacturers will need to demonstrate that their entropy sources are robust, ideally through third-party audits and formal verification. Some may adopt hybrid approaches, combining on-device randomness with user-supplied entropy, such as dice rolls or coin flips. Trezor has experimented with this model, allowing users to add their own randomness during wallet initialization.

Regulators, meanwhile, are likely to take notice. The European Union's Markets in Crypto-Assets regulation and similar frameworks in Asia include provisions for custodial security, but hardware wallets occupy a grey zone. They are consumer devices, not financial services, yet they hold billions in assets. Expect calls for mandatory security certifications and liability standards.

At DailyTechWire, we've long argued that the crypto industry's fixation on decentralization often obscures the centralized points of failure that remain. Coldcard was supposed to eliminate trust. Instead, it concentrated it in a firmware implementation that users had no practical way to verify. The $130 million loss is a reminder that in security, trustlessness is an aspiration, not a guarantee. And when the randomness breaks, the fortress is just a box.