Trezor raises its passphrase limit to 128 characters after Coldcard thefts

• Trezor’s October Suite update raises the maximum passphrase length from 50 characters to 128. Both the app and the device firmware have to be updated for it to work.
• The company points to the Coldcard thefts, where some devices generated wallet backups that attackers could recreate offline, as the reason people asked for it.
• Trezor says passphrase users were protected. That is a company blog statement with no numbers attached, and the same post describes the effect in weaker terms elsewhere.

Hardware wallet makers do not usually change a character limit and write a blog post about it. Trezor has just raised the longest passphrase you can set on its devices from 50 characters to 128, and pointed directly at the Coldcard thefts as the reason people kept asking for it.

What actually changed

The update is small in technical terms. “The limit was 50 characters, and it’s now 128 characters,” Trezor wrote in the company blog post announcing the October Suite release. Devices still running older firmware keep the 50-character ceiling, so getting the longer option means updating the Suite app and the firmware on the device itself.

None of the language in that post is attributed to a named person. It is company marketing copy, which matters for the claims that follow rather than for the setting change, which is straightforward enough.

Passphrase, PIN, recovery words: three different things

These get used interchangeably and they do completely different jobs.

The PIN unlocks the physical device. If someone steals the hardware wallet off your desk, the PIN is what stops them using it. It protects the object, nothing more.

The recovery words, usually twelve or twenty-four of them, are the wallet. Write them down, type them into any compatible wallet anywhere in the world, and you have the coins. Anyone who sees that list has your money, whether they ever touch your device or not.

A passphrase sits on top of the words. Add one and the device derives an entirely different wallet from the same recovery list: same twenty-four words, different passphrase, different set of addresses with different coins in them. The device does not store it. You type it in each time. That is why it survives a leaked backup, and why a single typo silently drops you into an empty wallet with no error message to tell you what went wrong.

Why Coldcard is in the announcement

In July, a firmware flaw in some Coldcard devices caused wallet backups to be generated with far less randomness than intended. Coinkite, which makes the devices, has attributed the root cause to a build error dating back to a 2021 firmware migration that left seed generation running on a general-purpose software random number generator rather than the hardware one. That account comes from Coinkite’s own technical backgrounder as relayed by Unchained, and we have not read the original directly.

The practical effect was that attackers could recreate the weak backups offline and simply spend the coins. Galaxy Research’s Alex Thorn, posting on X, put the total above $100m, roughly 1,596 BTC from around 7,300 addresses across three confirmed waves and fourteen smaller incidents. Trezor’s post cites no dollar figure at all and says only that funds were drained from “thousands of addresses”, so the two accounts of scale have not been reconciled by anyone. We covered the first wave of those thefts as the numbers were still climbing.

The claim, and what sits under it

Here is the part Trezor is selling. “Those with a strong, unique passphrase were protected,” the post says, because “an attacker needed both the weak backup and the passphrase to reach their wallet.” The company also positions itself as the originator of the feature: “Trezor brought passphrase protection to hardware wallets, and it’s now standard across the industry.”

The mechanism is sound. The evidence offered for it is not quantified anywhere in the post: no count of users, no number of wallets, no value saved. And the same post describes the outcome in notably weaker language elsewhere, saying a passphrase bought extra time to migrate funds before the wallet was wiped. Delay and protection are not the same thing, and the gap between them is the difference between noticing in time and not.

Nor is there anything in the announcement suggesting that 128 characters specifically was needed to stop a real attack. Trezor frames the change as an answer to “the question we kept receiving” from users, which reads as product management with a security story attached rather than a fix for a demonstrated weakness in the old limit. Fifty random characters is already far beyond what anyone can brute force.

The trade-off nobody puts in the press release

A passphrase has no reset. No support ticket, no recovery flow, no company that can help. Forget it and the coins behind it are gone in exactly the way a thief would have taken them, except you did it to yourself.

That is the risk a longer ceiling quietly raises. Something you can memorise and retype under stress is worth more than something long that you half-remember, and anything written down now has to be stored as carefully as the recovery words themselves and in a different place. If you are weighing any of this up, our self-custody guide covers how the pieces fit together, and the Trezor versus Ledger comparison covers how the two handle it differently.

What to watch

Whether anyone publishes an actual count of Coldcard owners whose passphrase kept their bitcoin intact. On-chain analysts can see which weak addresses were drained and which were not, so the number is knowable even if no manufacturer has an interest in producing it.

And whether other wallet makers follow. A run of device-level failures this year has pushed the industry towards defence in depth rather than trusting any single component, and a passphrase limit is the cheapest possible way to look responsive to that. The more interesting responses will be the ones that cost the manufacturers something.