$70M Coldcard Hack: If Your Seed Wasn't Random, Your 'Hacked-Proof' Wallet Isn't

What happened on July 30

This was not another routine exchange hack. On July 30, an attacker swept 1,082.65 Bitcoin from 1,196 wallets in 41 minutes, draining roughly $70 million in a fast, systematic sweep. The problem was not a bridge exploit or an exchange insider threat. It was a seed-entropy failure: wallets were created with private keys that were not generated with sufficient randomness, turning the old self-custody pitch into a real danger when keys become guessable private keys were not generated using sufficient entropy.

Why this matters beyond one firmware bug

Cold storage is supposed to remove online attack surfaces. But if the seed was generated with weak randomness, the weakest link is no longer the network path; it is the moment the wallet was created. On affected firmware, Coldcard fell back to a weak software PRNG instead of the hardware true random number generator. That makes the problem more unsettling than typical exchange risk because weak seeds can expose old, dormant wallets, not just actively used ones.

The situation also looked unresolved as coverage expanded. Coinkite first warned Mk3 users and only later advised later-device owners to take precautions while investigators were still looking into reportedly ongoing drains. That makes this feel less like a closed postmortem and more like an event still being fully mapped.

Why weak randomness breaks the cold-wallet promise

On affected Coldcard Mk3 devices starting with version 4.0.1 in March 2021, the firmware fell back to a weak software random generator instead of the hardware true random number generator, producing seeds with only ~40 bits of entropy rather than the intended 128. In practical terms, that turns a problem that should be computationally infeasible into one that may simply be a race against time.

How the attack worked

The fallback was tied to the chip serial number and clock registers, which narrowed the seed space from astronomically large to something researchers could actually search. Investigators reproduced key generation on older Mk2 and Mk3 devices, and for the Mk4, Q, and Mk5 they estimated a range of about four billion possible seeds.

This does not require remote access to the device. An attacker can generate candidate seeds, derive the corresponding addresses, and compare them with the public blockchain. The victim's hardware does not need to connect during that search. The wallet mix behind the drain also matters: the sweep hit 1,183 native SegWit addresses, seven older standard addresses, and six even older addresses, suggesting systematic testing across several derivation paths.

The bigger question is confidence, not just scope

If the issue is confined to specific firmware and device lines, affected users can move funds and the broader self-custody story can survive. But the narrative damage runs deeper. Once a hardware wallet stops being a trustworthy source of randomness, confidence gets damaged fast. Coverage of the incident already raised the question of whether managing private keys has become too risky for everyday investors.

That is the real pressure point now. Even if Bitcoin itself is fine, the slogan "your keys, your coins" is less comforting if the key-generation process itself turns out to have been unpredictable.

What Coldcard users should do now

The practical takeaway is simple: if your seed was created on vulnerable firmware, treat the situation as urgent. The confirmed exposure center is Mk3 devices with firmware from version 4.0.1 in March 2021, and Coinkite has said those Mk3 units were affected.

When to move funds immediately

Treat this as a same-day migration if:

• You use an affected Mk3 running firmware from version 4.0.1 onward.
• You do not know whether your seed used strong entropy.
• You are unsure whether your wallet was among the affected derivation paths.

If any of those apply, prepare a fresh wallet on a known-clean setup and move funds rather than waiting for a perfect public audit. When the issue is weak randomness at seed generation, delay only gives attackers more time.

How later-device users can stay measured

Coinkite said Mk4, Mk5, and Q users should also take precautions, but that guidance is different from saying every later device has the same confirmed exposure.

For those users, the better posture is preparation, not panic:

• Check your exact model and current firmware.
• Have backup hardware ready and verify your recovery process before making changes.
• If you used weak seed input methods, follow official guidance to regenerate the seed on updated firmware.
• If your Mk4, Mk5, or Q setup looks clean and you used strong entropy, wait for clearer official guidance.

Why urgency depends on the exact setup

Not every Coldcard owner faces the same level of risk. The most serious exposure is tied to the confirmed Mk3 firmware window and seeds that may have been generated without sufficient entropy. Later devices were told to take precautions, and researchers also found a nontrivial seed range for some of those models, but that does not mean every later wallet has the same confirmed weakness.

Until you can confirm your model, firmware, and entropy source, the safest test is straightforward: if you used an affected Mk3 from the March 2021 firmware window onward, you should assume your wallet is not as immune as the branding implies.