Large-scale Coldcard compromise (1596 BTC stolen so far)

[Cricktor]

...

I briefly followed the circus and drama with nvK and am aware of the grotesque pictures they paint. Well, let's put it this way: karma caught up on that company and its clowns.

It's just very sad that the wrong people, their customers, suffered from the consequences of an evil attitude and arrogance of those who are to blame at Coinkite.

There were times when (on paper!) I liked Coldcard, Opendime and Blockclock. Prices not so much, but maybe I would've swallowed it. I'm glad, I never did. Once I learned and followed the malice of nvK, any of their products became a no-go for me. I simply refuse to give my hard earned money or coins a company with such morons at the helm. In the end it was the right choice.

...

The concept of a stateless signing device has some charm, although it's only comfortable, aka not annoying, when you can easily import your wallet with SeedQR or similar. A DIY aspect is also nice because it opens the opportunity to get this stuff without the crypto coin label attached to it (avoiding a crypto coin purchase trail which is nice for privacy).

But I also see the SeedQR as a disadvantage. I'm only comfortable with it when it's the encrypted variant which again has the burden to not ever loose the encryption secret. Hell, safety is such a pain sometimes or shall I say: most of the times?

...

I've already quality-control concerns regarding their coding when they've done presumably "a lot of re-implementations". Sounds a lot like marketing bs from them...

[bitmover]

I imagine there is likely more than one class action being cooked up.

Would they have deep enough pockets for this? I can imagine they spent a lot of money on marketing, sales and the actual product, while they didn't sell that many devices. So if all buyers want their money back, they don't have that amount. If the victims want their losses compensated, they won't have that kind of money it either.

I wonder if they will be able to find the users who now hold those 1500 btc. One small mistake when spending, even in a kyc free exchange (like an ip leak),  might expose their identities

Tell will tell if someone from coldcard is involved

[Meuserna]

SeedSigner does not save any keys on the device.

Keys as in buttons? A Seedsigner has navigation keys on the left and three buttons on the right of the case. There are other casings that change the left-side navigation buttons and turn the device into a GameBoy Advance-looking thing. But there is no keyboard, though.

Keys, as in, keys to coins.

SeedSigner does not save the seed, the wallet, or anything else on the device. That's not a weakness. That's a strength.

SeedSigner is stateless. Every time you shut down or reboot, the device wipes the seed and the wallet.

On a ColdCard, when you boot the device, you have to enter the PIN to unlock the device. When a SeedSigner boots, there's no PIN to enter because there's no wallet on the device. You have to scan a seed QR to load the wallet.

ShieldSigner is a SeedSigner fork that adds features like encrypted seed QR, passphrase QR, and smartcard capability. ShieldSigner is incredible. I'd definitely recommend it to anyone who liked ColdCard. ShieldSigner is so much better.

I greatly prefer using a stateless wallet. If the device gets stolen, no worries. There's nothing on it.

[Meuserna]

...

The concept of a stateless signing device has some charm, although it's only comfortable, aka not annoying, when you can easily import your wallet with SeedQR or similar. A DIY aspect is also nice because it opens the opportunity to get this stuff without the crypto coin label attached to it (avoiding a crypto coin purchase trail which is nice for privacy).

But I also see the SeedQR as a disadvantage. I'm only comfortable with it when it's the encrypted variant which again has the burden to not ever loose the encryption secret.

It's not a burden at all.

What's the first thing you're supposed to do when you get a seed phrase? Write it on paper and make a metal backup. So, if you lose your encrypted seed QR decryption key, no worries. Just enter your seed phrase manually and create a new one.

[tvbcof]

...
I greatly prefer using a stateless wallet. If the device gets stolen, no worries. There's nothing on it.

On top of that, it's much more comfortable to take it through security at the airport, and it doesn't matter what device one uses so there could exist a devices and spares wherever one might need oneself (or one's associates in the case of multisig or whatever) to be working.

A good question to have, however, would be how to be SURE the device really is stateless and not subverted to simply make one believe that it is?  In other words, how to be sure that the seed material, once loaded, is not tucked away somewhere on the circuitry to later be exfiltrated?  This could be a bigger problem with remote devices left on-site for on-site work.  It's getting pretty 007-ee by this time though.  OTOH, some of these devices have 10's or 100's of millions of dollars tied to them so the motivation to find a way could be significant.

[Meuserna]

...
I greatly prefer using a stateless wallet. If the device gets stolen, no worries. There's nothing on it.

On top of that, it's much more comfortable to take it through security at the airport, and it doesn't matter what device one uses so there could exist a devices and spares wherever one might need oneself (or one's associates in the case of multisig or whatever) to be working.

A good question to have, however, would be how to be SURE the device really is stateless and not subverted to simply make one believe that it is?  In other words, how to be sure that the seed material, once loaded, is not tucked away somewhere on the circuitry to later be exfiltrated?  This could be a bigger problem with remote devices left on-site for on-site work.  It's getting pretty 007-ee by this time though.  OTOH, some of these devices have 10's or 100's of millions of dollars tied to them so the motivation to find a way could be significant.

A SeedSigner is just a Raspberry Pi with a camera and an LCD hat. It has no storage except for the micro SD card slot. And the firmware is 100% open source.

You're right about taking it through security. Put a micro SD card with a video game on it instead of SeedSigner firmware. Even if somebody forces you to plug it in and turn it on, all they see is a video game... because that's what it is if that's what's on the micro SD card.

On the other hand, I always laughed when people talked about ColdCards being covert. "It looks like a calculator!" And it has a hardware wallet name on it. So much for OpSec. LOL.

[Pmalek]

But I also see the SeedQR as a disadvantage. I'm only comfortable with it when it's the encrypted variant which again has the burden to not ever loose the encryption secret.

You shouldn't look at it as a disadvantage in my opinion. The SeedQR is just another variant of your 12/24-word seed, represented in another way. You write down the seed physically on paper, metal, or whatever. It's not encrypted and you are supposed to hide it somewhere safe. Add a passphrase on top of that to increase security and hide that somewhere else. If it works for recovery phrases, the same approach is fine for seedQRs. The encryption is just another optional layer on top of everything else.

Keys, as in, keys to coins.

SeedSigner does not save the seed, the wallet, or anything else on the device. That's not a weakness. That's a strength.

SeedSigner is stateless. Every time you shut down or reboot, the device wipes the seed and the wallet.

On a ColdCard, when you boot the device, you have to enter the PIN to unlock the device. When a SeedSigner boots, there's no PIN to enter because there's no wallet on the device. You have to scan a seed QR to load the wallet.

Oh, so that's what he meant.

I like stateless modes in signing devices as well. I have a Jade that works similarly. What I don't like about the first model is that it requires USB or Bluetooth connection to a computer to install or update the firmware. Bluetooth can be turned off on a firmware level (you install a firmware that doesn't have it), but that just leaves USB connections. The first model doesn't support SD cards to get data to and from the signer. They added that functionality on Jade Plus - their second model.

[examplens]

I imagine there is likely more than one class action being cooked up.

Would they have deep enough pockets for this? I can imagine they spent a lot of money on marketing, sales and the actual product, while they didn't sell that many devices. So if all buyers want their money back, they don't have that amount. If the victims want their losses compensated, they won't have that kind of money it either.

I wonder if they will be able to find the users who now hold those 1500 btc. One small mistake when spending, even in a kyc free exchange (like an ip leak),  might expose their identities

Tell will tell if someone from coldcard is involved

Well, there are indications that the hacker made a mistake and exposed information about himself. Activities were recognised where the operator used a paid account on a "well-known" blockchain service. A pattern of querying the original addresses was observed during the attack and this was repeated.

https://www.cryptotimes.io/2026/08/19/block-and-galaxy-research-give-lead-to-law-enforcement-in-111m-coldcard-bitcoin-theft/

[tvbcof]

I am not a lawyer, but according to the official story (which studiously avoids any 'conspiracy theories') I'm not really sure I see a crime here.  Or at least not one which anyone is likely to get more than a slap on the wrist for.  We have:

 - a simple mistake from some device maker/seller,

 - someone(s) found they could obtain shared control of keys represented in a public blockchain and availed themselves of the discovery.

At this point there is no provable evidence that either Coinkite deliberately sabotaged their hardware with the intent to defraud, or that there was any collusion between Coinkite and the various parties who picked up the dropped private keys.  Indeed, it's kind of a weird sin-like thing to even think that way here in 2026.

I personally think it likely that both deliberate sabotage and collusion to engineer a lifting of BTC probably did occur, but that means nothing.  Even if it did, it doesn't seem to me like something which would get anyone into serious trouble...at least in a formal court of law.

Actually, foreknowledge of the hack would possibly yield greater monetary benefit in terms of various kinds of 'insider trading', or in terms of getting compensated (or just a pat on the head for doing God's work) by entities who benefited (e.g., coin custodians who won keys as the whole system shifted away from faith in self-custody), but nobody gets in trouble for that kind of stuff these days.

[bitmover]

I am not a lawyer, but according to the official story (which studiously avoids any 'conspiracy theories') I'm not really sure I see a crime here.  Or at least not one which anyone is likely to get more than a slap on the wrist for.  We have:

 - a simple mistake from some device maker/seller,

 - someone(s) found they could obtain shared control of keys represented in a public blockchain and availed themselves of the discovery.

At this point there is no provable evidence that either Coinkite deliberately sabotaged their hardware with the intent to defraud, or that there was any collusion between Coinkite and the various parties who picked up the dropped private keys.  Indeed, it's kind of a weird sin-like thing to even think that way here in 2026.

I personally think it likely that both deliberate sabotage and collusion to engineer a lifting of BTC probably did occur, but that means nothing.  Even if it did, it doesn't seem to me like something which would get anyone into serious trouble...at least in a formal court of law.

Actually, foreknowledge of the hack would possibly yield greater monetary benefit in terms of various kinds of 'insider trading', or in terms of getting compensated (or just a pat on the head for doing God's work) by entities who benefited (e.g., coin custodians who won keys as the whole system shifted away from faith in self-custody), but nobody gets in trouble for that kind of stuff these days.

You don't see a crime in transfering 1500 BTC (millions of dollars) from random people to your own wallet? What are you talking about.

It is the same as hacking a bank app and transferring money to your own account.

[avp2306]

Would they have deep enough pockets for this? I can imagine they spent a lot of money on marketing, sales and the actual product, while they didn't sell that many devices. So if all buyers want their money back, they don't have that amount. If the victims want their losses compensated, they won't have that kind of money it either.

I wonder if they will be able to find the users who now hold those 1500 btc. One small mistake when spending, even in a kyc free exchange (like an ip leak),  might expose their identities

Tell will tell if someone from coldcard is involved

Well, there are indications that the hacker made a mistake and exposed information about himself. Activities were recognised where the operator used a paid account on a "well-known" blockchain service. A pattern of querying the original addresses was observed during the attack and this was repeated.

https://www.cryptotimes.io/2026/08/19/block-and-galaxy-research-give-lead-to-law-enforcement-in-111m-coldcard-bitcoin-theft/

Now that they already traced the hacker, its good to see what other action they made. Because many people are so curious about this case, they want those people involve in this hacking issue to get captured.

Maybe this is also the reason why the funds stolen has left unmoved? https://crypto.news/coldcard-hackers-leave-87-of-stolen-bitcoin-unmoved-after-114m-theft they are maybe afraid to do transaction because it can create fresh traces to the authorities.

They should act fast, so they can still recover the funds and return it back to those proven victims.

[joker_josue]

I have built a couple of these and the project is extremely impressive. It has its weaknesses and its strengths.

Are you satisfied?
All the people I've talked to about this equipment speak very well.

I believe it is another model to serve as a hold. But it is still very interesting.

I think I'll have to venture out.

[tvbcof]

...

A SeedSigner is just a Raspberry Pi with a camera and an LCD hat. It has no storage except for the micro SD card slot. And the firmware is 100% open source.

You're right about taking it through security. Put a micro SD card with a video game on it instead of SeedSigner firmware. Even if somebody forces you to plug it in and turn it on, all they see is a video game... because that's what it is if that's what's on the micro SD card.

On the other hand, I always laughed when people talked about ColdCards being covert. "It looks like a calculator!" And it has a hardware wallet name on it. So much for OpSec. LOL.

I'm uncomfortable with powerful and complex OS's for very simplistic tasks (e.g., Linux).  Similarly, with the basically two monopoly-ish architectures in widespread use in the West and licensed by their authorities.  ARM and X86 are not Open ISA architectures.  For many years I have been enthusiastic about the development trajectory of risc-v for this reason.  If/when, if not already, there does seem to be a potential for smaller-scale fabs to exploit a niche market for fully auditable silicon.

The reason strong air-gap is so important to me is that it makes otherwise unsafe silicon safe...with the Achilles heal being mainly entropy generation.  Thankfully that is easy, if tedious, to overcome when it comes to simple stuff like Bitcoin.  Other common cryptographic needs, no so much.

I dug through my collection of signing devices and found a Krux.  It seems to operate in the same 'stateless' manner as the SeedSigner.  To be honest, I didn't really appreciate the distinction between stateful and stateless back several years ago when I was researching this stuff.  At least not as much or in the same way as I do now.  The Krux operates mainly on risc-v dev boards so it seems.  Still researching the operating system.  Since the timeframe of the few SeedSigners I have kicking around, their Linux OS has been cut down which is a good thing (as I see it.)

I think at this point I would be favorable to a very simple (but durable) and inexpensive stateless device running an extremely primitive OS which is always built from source by the user as a matter of course.  In order to avoid a large screen, I would hope it possible to so some sort of a long scrolling bar code as opposed to something like bbqr.

That said, I just watched a pretty interesting vid about the likely trajectory of quantum-safety in Bitcoin, and it looks likely that key management is probably going to become much more computationally expensive.  https://www.youtube.com/watch?v=E61gUCKcx2s.  Probably it does make some sense to consider some of this stuff when choosing the ultimate bitcoin management 'assist-device' which has staying power.

Edit:  Another huge win for stateless devices, which I now recognize after some experience, is that I would be far less concerned about upgrading them.  My experience with wallets is that it is always a matter of puckering-up at upgrade time to see what regressions the 'upgrade' brings along for the ride.  I cannot help but come to the conclusion that most of these open source efforts are often undertaken by people who have minimal hands-on experience with production hardware/software and very little budget for testing and QA.  Also, minimal attention spans.  I still vastly prefer Open Source, but it is a notable disadvantage which I cannot ignore.

[joker_josue]

I dug through my collection of signing devices and found a Krux.  It seems to operate in the same 'stateless' manner as the SeedSigner.  To be honest, I didn't really appreciate the distinction between stateful and stateless back several years ago when I was researching this stuff.  At least not as much or in the same way as I do now.  The Krux operates mainly on risc-v dev boards so it seems.  Still researching the operating system.  Since the timeframe of the few SeedSigners I have kicking around, their Linux OS has been cut down which is a good thing (as I see it.)

I think at this point I would be favorable to a very simple (but durable) and inexpensive stateless device running an extremely primitive OS which is always built from source by the user as a matter of course.  In order to avoid a large screen, I would hope it possible to so some sort of a long scrolling bar code as opposed to something like bbqr.

Based on your experience, would you prefer to set up a SeedSigner or a Krux (even though its development has stopped)?

Or would you prefer to set up a different type of system to generate your seed and use it to sign transactions?