CEOInterviews.AI
Start App
Alex Thorn
Head of Firmwide Research, Galaxy Digital

Tracking the Coldcard Hackers: Inside the $120M Bitcoin Theft with Alex Thorn

📅 Aug 07, 2026 Bitcoin Magazine 68 MIN 4426 VIEWS 50 SEGMENTS · 2 SPEAKERS
Over $120 million in Bitcoin has been drained from Coldcard hardware wallets and the attack is still ongoing. Alex Thorn, Head of Firmwide Research at Galaxy, joins Spencer Nichols to break down exactly how the critical entropy bug in Coldcard's 2021 firmware was exploited, how he's tracing stolen funds on-chain, and what victims must do right now to have any chance of recovery. Thorn also explains why multisig and collaborative custody may be the only responsible path forward for Bitcoin self-custody. ๐Ÿ”ถ Host: Spencer Nichols โ€” Bitcoin Magazine ๐Ÿ”ถ Alex Thorn โ€” Head of Firmwide Research, Gal...

Questions asked in this interview

12
  1. 0:19Alex, welcome to the show. How are you doing today?
  2. 0:35You know, sounds like last week users started knowing funds going missing, but when did you begin to enter the picture here?
  3. 9:13This is Rob Hamilton of Anchor, correct?
  4. 10:51Have you seen any evidence that this has been ongoing prior to it gaining really this public legibility?
  5. 23:46Is that an operation that's being undertaken?
  6. 26:51What insights can you offer us there, and what conclusions can we draw and can't we draw at the current time?
  7. 31:29How do you, how do you think, like how do you read what shift has taken place?
  8. 37:56How this is going to just change the Bitcoin culture?
  9. 50:46I mean, I just wonder, you know, is closed-source frontier model AI, like is that even a path worth going down?
  10. 52:07I don't know if that's true, but I know they do, right?
  11. 52:30You know, I don't know, Spencer, have you ever looked at videos of what it was like in the 1990s?
  12. 1:00:56Is it Trump coin? Are they related?
Alex Thorn 0:00 ↗
Wave 1's identity, attacker identity may be known to law enforcement. Wow. So that could, and that wave 1 is by far the biggest chunk of Bitcoin that we've found so far. It's 1,082.65 BTC, and all of it is still held in the attacker address.
Spencer Nichols 0:19 ↗
Hello and welcome back to the Bitcoin Magazine podcast. I'm your host, Spencer Nichols, and today we are joined by a very special and very busy guest, the one and only Alex Thorne, head of firmwide research at Galaxy. Alex, welcome to the show. How are you doing today?
Alex Thorn 0:33 ↗
I'm great, Spencer. Thanks for having me.
Spencer Nichols 0:35 ↗
Of course. Well, I appreciate you taking the time. You have been a very busy man with the latest in the Coldcard exploit that is still ongoing. To date, just over 120 million US have been drained from affected users' Bitcoin wallets. And we're going to discuss today how we got to this point and really what the future holds for the future of Bitcoin self-custody, the culture around it, but also just digital security generally. I think there's a lot of lessons to unpack. But I think, you know, for viewers who are not familiar with what you've been doing, if you could just kind of walk us through the fact pattern of what happened over these last few days. We're recording here on Wednesday, August 5th, for those listening as well. But Alex, you know, where did this all kick off? You know, sounds like last week users started knowing funds going missing, but when did you begin to enter the picture here?
Alex Thorn 1:22 ↗
Yeah, it's been quite a few, gosh, what, five days now, I think, since Friday of last week when I first really started working on this. The actual wave of exploits began shortly after midnight, July 30th, UTC time. So that would have been last Wednesday, July 29th, in the evening Eastern time, in the, you know, earlier evening and afternoon Pacific time. We were throwing an event, I was hosting an event, you go on my X, you'll find I posted Twitter's from it at Pubkey DC, and I was in DC all last week, you know, working on Clarity Act related policy advocacy, assistance, etc. And we threw the summer policy reception at Pubkey DC last Thursday night, and there was a Bitcoiner there who I know, of course, and he came up to me about an hour before the event started. You know, I'm like, you know, getting a poster up on the easel, like whatever, right? Getting ready for the event, and told me that he was hearing some stuff about there might be a problem with Coldcard, there might be a vulnerability. He ended up spending most of that night working on it, or I don't know if he had to move his own funds or whatever, but like not really participating in the reception, though I, as the primary host of the party, was totally mobbed by attendees and going around talking to people, right, and playing host. So, and I didn't fly away from DC and get home until Friday morning. So that's sort of when I started really looking in depth. I believe this is still before Coinkite ever posted. Block Inc. engineers posted that and wrote a post about that they had identified a pattern of transactions that looked to be an attack on Coldcard, and I took their markers of that pattern, which I could talk about, but you know, it's things like a fixed fee rate, like a bunch of coins that had never interacted before that were all created after the exploit was introduced in March 2021, that were long dormant, being spent together and sent to a hop address and pulled, and like, you know, looks like an attack, right? And that's how they sort of identified it.
I took those markers and reran it on my powerful proprietary Bitcoin analytic stack and also found that, and I started calling that wave one, and that's what I've been calling wave one, and that's the only one of the waves. I'm now up to wave two. One, wave one, two, and three we're very confident are hacks. Wave four I've talked about, and we'll get into this I think a little bit, but is still like I haven't promoted it into a like definitely a hack yet, but that like the pattern Block engineers found I have found and think also looks like it could be a pattern of attack. I've also got much smaller patterns that I call footprints, and we're up, those are lettered episodes. We're up, I think the last time I posted yesterday an update, we're up to footprint N. I can tell the audience we're now at P in footprints, and all of those except for my perspective wave four, which is I've not promoted into the set yet, I'm waiting on victim confirmation.
Other than wave one, all of the promoted sets were identified thanks to victim reports to me. Right? So they send in and say, "Well, mine was drained." And I say, "Huh, you're not in wave one, right?" Like, "Let me..." So I'm adding your stolen coins to the totals, but let me put this in a triage on a list of transactions that look like yours that are not part of a known first wave, like wave one. And then when I have some time in between ingesting new victim reports and, you know, trying to update numbers, I then go and spin analysis on the open chain. I basically look for markers in their transactions, anything about the shape of the transaction, where it was spent, when it was spent, were there others that look like that, right? And with high confidence, we found two full other waves based solely on that, right? Getting a couple victim reports, first one, then a couple others that look similar, then saying, well, now let's just go and look at every transaction that remotely looks similar, right? And then we've been able to identify dozens of new attackers. Not a dozen, sorry. At least a dozen attackers, I think.
I, you know, just probably not the same person. Could be, of course, we don't know, but like thanks to victim reports. So I've talked now to at least 150 individual victims directly who have reached out to me. And that's what we're doing. The attack is ongoing. We continue to get new victim reports. You know, the impact, the blast radius, so to speak, it keeps growing. And we continue to take the addresses that we identify as containing stolen funds, so attacker addresses, and widely propagating them across the Bitcoin and crypto industry and law enforcement. So, you know, US federal law enforcement, cyber crime and cyber crime investigation bureaus, crypto exchanges, so that if the funds ever show up, they can be frozen, etc. And then yeah, and then I've been providing a comprehensive forensic report back to each victim that they can use to file with their local police or authorities and the FBI's IC3 and whatever the correct like cyber crime anti-fraud center in their jurisdiction is, that shows everything about what happened to their coins, where they're currently sitting, so they can formally establish their victimhood and thus have a chance of recovery if the funds ever show up at a centralized intermediary. That's what we've been doing. I would say that I don't know, is that that's the timeline. I guess we continue to do that. I haven't slept more than like, I think last, we're recording on Wednesday, August 5th, you pointed out, Tuesday was the first night since Friday that I'd slept more than like three hours. So, I've been staying up till two, waking up at six, you know, and that, you know, so and I gotta say that the stories that I've heard from victims are incredibly sad. So, I'm just trying to do anything, you know, I have this capability. There are other people doing a lot and also to help in various ways. You know, a couple I just wanted to name, Wicked Bitcoin, Wicked Smart Bitcoin on X has been running like marathon self-custody Q&As. I've heard him help probably dozens of even Coldcard users like use Sparrow wallet for the first time to get their coins out of their Coldcard or answer other questions about the actual self-custodying of Bitcoin. And many others have been in those rooms like Shinobi and D++ and other sort of well-known technical Bitcoiners to help people work on the security of their self-custody setup. Rob Hamilton and Cala and James O'Beirne and Portland Huddle and many others are basically red teaming like all Bitcoin software, not even specifically protocol code, though I'm sure they have all other hardware wallet firmwares, all open source libraries, anything that anyone ever has ever used for Bitcoin. I think one of the takeaways from this attack is that the software is not audited enough, and so they're just doing a full sweep of the software that Bitcoin users use to, you know, identify and patch any other bugs. There is no, Rob has confirmed publicly that none of the bugs they have found and vulnerabilities that they have reported.
Spencer Nichols 9:13 ↗
This is Rob Hamilton of Anchor, correct?
Alex Thorn 9:16 ↗
Yes. Rob Hamilton had confirmed, I believe, that none of the bugs they found would create similar types of vulnerabilities. So this looks very, this specific vulnerability is limited specifically to Coldcard on-device key generation on firmware that was released March 17, 2021 or later. So if you generated keys on a Coldcard device on such firmware, and you still haven't, you should migrate your funds as soon as possible to a fresh seed. Don't feel like you're not a pure enough Bitcoiner if that means just sending it to Coinbase or River or a centralized exchange that you trust, right? You can build a new self-custody setup later. It's important that if you are holding coins on a Coldcard MK3, MK4, MK5Q or an MK2 that you upgraded to the firmware, if you bought one of those devices like after March 2021, you should assume that the key generation lacks sufficient entropy to be secure and you should migrate your funds off of that. There's a bunch of nuance, you know, if you use it in a multisig like in different ways or if you added a strong passphrase or if you rolled dice and added your own entropy. There are different instances when you may or may not be at risk. But I would just say like the high level, you should be, you know, if you really don't want to migrate, then you better figure out whether you fit into one of those exceptions, right? But in general, the average user, if you made keys on a Coldcard after March 2021, you should be looking really carefully right now about whether you need to move your funds. And I would recommend personally just go ahead and do it.
Spencer Nichols 10:51 ↗
Yeah. And I think something I want to double click on is obviously this has gained a lot of notoriety as of the end of July leading now into August 5th. But at the same time there are people that have reported, hey, I had an issue with my Coldcard being drained, as I believe I've seen people referring to 2024 as early as, and you know, this is obviously following the date of that firmware update. Have you seen any evidence that this has been ongoing prior to it gaining really this public legibility? I've seen some reports and screenshots of like Reddit posts of users saying that they had, I think in some cases maybe even a couple years ago, that they had created a seed on a Coldcard and sent funds there and the funds have been drained.
Alex Thorn 11:33 ↗
We do know with certainty that the bug has existed since March 17, 2021. So how it went five years before anyone found it at scale and disclosed it is beyond me. And I think that's one of the big lessons to take away here for the Bitcoin community, and it's sort of what Rob and others with that red team are spearheading, right? I don't know. I mean, it's absolutely possible that someone else was occasionally doing this because the bug existed. I've seen some reports. I haven't seen many. I've seen some screenshots of people who claim that that's the case. So I can't confirm that. None of the people that have reached out to me reported that yet that I got it triaged more. I mean, I think I've got now 220 victims. When I say 150, like I haven't replied now to 70 because I've, you know, I have a day job, so I'm getting to those. Don't worry, I will reply to everybody. Maybe there is one that will say definitively it was a Coldcard MK3 with no passphrase and no entropy, but this happened nine months ago. Like, I mean, I'm looking for those. If you've got them, you tell me. You know, so but yeah, I can't offer any confirmation.
Spencer Nichols 12:43 ↗
Yeah, and I think, you know, very clearly time is of the essence for those who have yet to migrate. And I also think that for those who want to be able to recover lost funds, being able to report this to the authorities so that if it does enter an on-ramp to an exchange those funds can be frozen and people can take steps to try to recoup that. So, you know, thank you for all the work you've been doing on that front and helping these individuals interface with law enforcement. But I also think that, you know, we're kind of assessing the damage and trying to pick up the pieces, but now we're also at a point where people are asking both how did this happen and who caused this to happen. I think that's been really interesting. And so you've mentioned there's a number of waves and a different profile of the attacker or attackers throughout those waves. What can you tell us about your perspective on where this may have originated from in terms of the personhood? And, you know, perhaps we could also discuss to what degree Coinkite and the services that they offered and sold, you know, are they at risk for the liability of these damages?
Alex Thorn 13:42 ↗
Yeah, I mean, I'll take your last question first. I don't know if Coinkite, what level of liability, you know, they're a Canadian company. I haven't read their terms of service. I don't know what representations they have made. Those questions will be asked, and I think it's a safe assumption that this will be litigated as well. You'll see lawsuits and plenty of that. I don't know the answer to that. I will tell you about, I know, I think it's obvious, and they've said like they caused the bug, right? This is their fault. This is Coinkite's fault. I want to be very clear that this is not the victim's fault. They didn't do anything wrong. They didn't lose their money because they speculated on memecoins. They didn't accidentally post their seed phrase on the internet. They didn't lose their money because they, you know, went on DeFi and down to some longtail risk like place chasing yield, right? These are people who bought a reputable hardware wallet, generated keys, stored their savings, and DCA. And the Coldcard demographic, as I think the Bitcoin Magazine audience will know well, these are the philosophical Bitcoiners. These are people that are working hard and saving in sats. And the average coin age that was stolen is like four years. The average amount is like seven BTC, right? So like it's not rich people who are speculating. It is average people who are saving, and that's partly what is so sad about this. I know we'll talk about the impact on, you know, self-custody, Bitcoin culture, etc. In terms of the attackers, okay, and why I say plural is I'm very, very confident there are multiple threat actors operating now using this vulnerability. The vulnerability is known, like we, it's been explained, Coinkite has explained it, Wizard Sardine, Blockstream, Bitcoin Optech, like people have explained exactly what the bug is. You don't even have to be that sophisticated to exploit it now with the help of AI. Of course, the AIs can understand what the bug is and how it might be exploited. It's really just a matter of compute. You've got to grind out the seed phrases that this weak entropy would create. But the world has never had more compute than it has today. A lot of people have access to compute. Certainly, bad actors do as well. So, it's a matter of time before all theoretical vulnerable addresses are broken and the funds are stolen, right? They will. I'm pretty confident at some point you're going to get a GitHub post that has literally every address and every private key, like...
And then that's when we'll know the attack is done finally. Like, but like this information will be open source eventually, like because you can't, like this is another reason it's so sad too. Before I talk about the attackers, like, and it's sad because every people did nothing wrong and they did everything right, but it's also, it's impossible to know how many coins are at risk, right? I mean, without replicating the attack and grinding all the seed phrases, we don't know, we don't know how many. It's not like a normal crypto hack where like a centralized exchange is just hacked and we all see the funds leaving that centralized point and dispersing, and the forensics is all about tracing where they go. Here, we have to trace where they come from as well. And that's why the victim reports are so essential in tracking this hack. And why I really thank and have so much love and empathy for the victims who have, for all of them, but especially those with the courage to come forth to me and provide their Bitcoin drained addresses and transaction IDs. I don't need any other personally identifiable information. I just need to know the addresses that you have that were drained. And I can give you a lot of help if you do that. But also you help the world by doing that because we are constantly identifying new attackers, right? And then I can take those addresses and send them to authorities and crypto exchanges and Chainalysis and TRM Labs and all of those people, right? So that they can hopefully be frozen and we can, you know, at least if nothing else, make it harder for the attackers to try to profit off the attack, right? So it's very important. So the attackers, so I would guess waves one and wave two look very similar. They both consolidate many victim addresses into one holding address and then send forth to a second hop holding address where all the funds currently sit inert. And also there are many instances of people whose addresses are in wave one not being fully drained and then in wave two it comes back and drains the rest. Right? So plausible that wave one and two are the same actor. You know, I hope that that's the case because I don't know if you saw Block engineers also, I have no idea how they did this, but they claimed that they were able to identify like a mainstream blockchain infrastructure provider, like a node RPC provider, right? If you don't run your own node or your business who needs a better node than like an at-home node in Bitcoin, like you just pay for an RPC access to like all Bitcoin information, right? Somehow they say that they identified, Clay, I think from Block Inc. said that they had identified that the wave one hacker had used, like queried like Bitcoin data from a centralized provider and had used a paid account there, and that they had been working with this provider to identify and talk to authorities. So, it's, I haven't heard more information. I think he posted that a couple days ago. I haven't heard any more information about that, but if that's true, wave one's identity, attacker identity may be known to law enforcement. Wow.
So that could, and that wave one is by far the biggest chunk of Bitcoin that we've found so far. It's 1,082.65 BTC, and all of it is still held in the attacker address. So, you know, knock on wood. Hopefully more to learn there. If wave two is also wave one hacker, that's another 76 Bitcoin. So, you're looking at like about 1,150 Bitcoin between wave one and two. Those patterns look very similar. So, I think they could be the same attacker. So wave three's pattern looks very different, and it doesn't mean it can't be the same attacker as wave one or wave one and two, but if it is, they would have dramatically changed their on-chain operational execution methodology. So whereas waves one and two flow many victim addresses into one, into just a couple collector addresses and then sit in just a three, just I believe just three in wave one, like holding addresses that the attacker controls. In wave three, the funds are downstream sitting in 293 individual pay-to-witness-script-hash addresses. Other ones are not script hash addresses, right? And wave one and two. So that's one difference. But also those 293, it appears from victim reports that the vast majority of the 293 hold only one victim's funds. So they're not pooled, right? So, it's like if you grind a seed, if you're the attacker, and you find you derive Bitcoin addresses and keys that do hold funds, they're kind of like each instance they find those, they're then siphoning them, right? And actually, they did this all as one blast. So, they, you know, they wrote them all down and constructed them as separate transactions, right? And so I did find a couple, maybe six of the 293 that are plausibly also a little bit co-mingled, but like the vast majority are one victim, one attacker address, right? And then all of these others like what I call these footprints, right? These are just different patterns, and I would never have found them. Well, some victim says I was drained, and here's my drained address and the transaction ID, and it doesn't match any of the known waves. So, it gets its own label, and then I find others who are also saying they're part of it, right? And so now I'm like, well, now it's worth going back and saying, well, you know, person A, person B, and person C all said that they were drained. When I go and look, like they were either drained in the very same way. Anytime, by the way, that they're drained in a transaction with co-spends, I assume all the inputs are also victims, and we promote that into the set. That's basically the confirmation I need, I feel that I need.
There are many though where the transaction shapes look similar, the times are similar, but there's no direct fund flow comingling. So I can't like definitively say, but it's like, you know, I get three or four and they're all in the same couple blocks that have the same pattern topography that, you know, output into the same address type, have the same lock time set, right? Like have similar fee, like either literal hard-coded similar fees, right? That's what I call these footprints. And some of them, the more victim reports I get, the more confident I get in them and in the other ones that look like it, but I don't have a report for. And I've promoted some of those as well. And there will be a lot more promotions coming. So, I think right now we currently, as of yesterday, have been saying that we're very confident 1,600 Bitcoin has been stolen across like 7,500 addresses. Very confident. Those are the ones that we say are very confident are the ones that we're sharing the attacker addresses with law enforcement and crypto exchanges. I'm gonna, we're going to put out a new updated figure because we've now internally promoted other addresses into the verified, some any that are verified by victims are getting added anyway. But we've got high enough confidence now in some additional patterns, and I'm at like pattern A through P at this point. So like that's why, and all of those individually also many of them look very different from each other, not just different from waves one, two, and three. This is why I'm saying like that the threat environment now contains multiple different operators, right? Because the vulnerability is known, so other people are popping up copycats to exploit this vulnerability, and I think, like I said, you know, they'll keep popping up until they're all grinded and trained.
Spencer Nichols 23:46 ↗
Yeah. Do you know anyone who has been like white-hatting this side of things, trying to get to the coins before someone else does and trying to hold them for safekeeping? Is that an operation that's being undertaken?
Alex Thorn 23:55 ↗
I'm aware of at least one effort to do this. I'm not in touch with the people doing it, but I have people I trust telling me that it's happening. I don't know the scale. I don't know. I'm not helping or anything. You know, we're not doing that. And the interesting thing for me is, you know, my, the thing that I can help with that I'm good at and that I have the resources to help with is on-chain forensic tracing. We're experts at that. I'm an expert at that. I've been, and we have the tools and the derived databases and tables of the entire history of the Bitcoin blockchain in many different formats and ways and the tools to traverse it. So that's what I've been doing. And if there is white hat activity, it's indistinguishable from black hat activity as far as the blockchain data is concerned.
So I don't know. Gosh, I guess I hope so. I think the white hats here have a tricky, if there are some, have a real tricky situation on their hands. It's why we encourage victims to come forth and file reports to formally establish their victimhoods. It's why I say never to destroy or throw away or wipe your Coldcard devices even after you've either been drained or migrated the funds, because in the apps you won't be able to prove your ownership with cryptographic proof. As I said before, like every private key of these vulnerable addresses is going to be posted eventually, it's going to be public data set, like...
Spencer Nichols 25:29 ↗
Anyone can come forward and say they had it at one point.
Alex Thorn 25:31 ↗
Correct. So, you know, establishing your victimhood early, keeping any evidence you may have, is going to be crucial to proving your ownership in the event that they can be recovered, including by a white hat, right? Or if they're frozen on an exchange, which I hope will happen if they ever move to an exchange. And I'm, you know, doing my part to help that happen if it can. So, that's why it's so important. There is an interesting thing, too. You know, if you bought your coins, if the upstream tracing of your coins that were drained emanates from a crypto exchange where you performed know-your-customer onboarding, that will be good evidence as well that you are the rightful owner of the coins, right? Because, and by the way, I think if that's the case, I don't know how long different crypto exchanges or Bitcoin brokerages store that information, but it's a reasonable step to send them an email or a letter saying, "Hey, please preserve all the records related to my account. Because it's my, later I was involved in this cyber incident and we're going to need those for future law enforcement or litigation reasons." I think it's a reasonable thing to do because that is one very good way to prove your ownership if the address that was drained was funded by a KYC crypto or Bitcoin brokerage account.
Spencer Nichols 26:51 ↗
And kind of going back to this who-done-it exploration, I think a lot of people have been looking at, okay, we understand that this was a firmware update pushed in 2021 by the Coinkite team. And they're looking, okay, well, who would have pushed such a thing? And I think that there have been people exploring, you know, who would be responsible for this. And then I think the obvious Bitcoiner question is, you know, what is your incentive to do such a thing? And this comes down to, I mean, two individuals are, I think, top of mind for scrutiny. I'm not alleging anything here, to be very clear, but obviously you have NVK, CEO of Coinkite, and then you have, oh gosh, Peter Gray, CTO of Coinkite. Peter Gray himself, as far as I understand, was responsible for this libngu firmware, I guess, source of this critical vulnerability. What can you tell us there as far as I think there's people who are incredibly skeptical, particularly of Peter Gray, who is responsible for this update. What insights can you offer us there, and what conclusions can we draw and can't we draw at the current time?
Alex Thorn 27:58 ↗
Well, we don't know if they knew there was a vulnerability. We don't know that, Coinkite has denied that they knew, right? They said that they had tested and looked and blah blah blah and couldn't, didn't know. We know that they are responsible for the vulnerability. We know they pushed the code that updated their own firmware. We, I think, can see in the repository because their code is not fully FOSS, free and open source, but it, I think, can be looked at, like I think it's like a read-only. I'm not the expert on this part of it, but I'll tell you what I know and what we can't, what we don't, what I don't know. They pushed the, they merged the code. They said it runs. That was the note. I've seen people who have analyzed their repository that say it is pretty sloppy repository. Not nearly enough commenting. Most commits just merge to main right away without any kind of review. So, here's what I can say. I definitely know they're responsible for the bug. They introduced the bug. I don't think as today there's any evidence it was intentional. I'm certainly reserving judgment about the Coinkite team on this. I think it's probably the most likely thing is that they make really awesome hardware and they're really bad software developers. Because every hardware wallet manufacturer knows and wallet, like good Bitcoin or crypto wallet knows that entropy is essential to developing and generating strong public-private key pairs. These, you didn't need AI to find that bug. These bugs have existed in the past in other wallets. Like this is well known that like this is a crucial part of the code, you know. You know, if the device breaks but it makes good seeds, like okay, it might be a crappy device, but like the crypto is strong enough, right? So they are sloppy. They were sloppy with the most mission-critical part of their software. So that we do know. I think that's well established now, right? So, I'm not going to speculate about like malfeasance. I have to say I doubt that. I doubt that. I don't know the folks at Coinkite or NVK well. I believe they are, they were well-meaning Bitcoiners. I believe they are talented hardware builders. I mean, I've had a Block Clock over my shoulder on my podcast for the last 265 episodes. I love the Block Clock. I've used Coldcard, right? Like I have an OpenDime. I have OpenDime on my keys with me. I have an OpenDime on my keychain, right? I love their hardware. It turns out their software is not good. Critically so. I'd guess that it was not malicious and that it was negligence that caused this.
And I think this is what Rob and many others on the red team are doing is they're saying, I mean, if we, if the community, they were supposed to, you know, don't trust, verify. Well, people didn't verify either. I think it, you know, we have a little bit of a tragedy of the commons here where I think, I guess everybody assumed someone else was verifying. Well, clearly they weren't. So, that's going to change. That's changing now. You've got basically in terms of how Bitcoin software development is audited and vetted, you've got two eras now, before and after the Coldcard vulnerability.
Spencer Nichols 31:29 ↗
Yeah, I think a lot of people, and especially just within my firm, Bitcoin Magazine, BTC Inc., are asking, you know, how did we get here? I think there was just like, you know, Coldcard extremely popular amongst, you know, the quote-unquote Bitcoin maximalist crowd, the self-sovereign maxis who I think were trying to do everything right. They're like, "Okay, let me get, you know, a hardware wallet from this highly reputable brand, take self-custody of my Bitcoin, get it off the exchange." And it turns out that if you weren't rolling your own dice, you ended up putting yourself in a much, much worse place than you would have been had you trusted a custodian to begin with. And really there was a trust in a hardware wallet manufacturer that I think was obviously quite misplaced. But yeah, this has been, I think, really a soul-searching moment for the Bitcoin community. How do you, how do you think, like how do you read what shift has taken place? Like I, it's, I feel like everyone's just questioning their priors at the current moment, to be honest.
Alex Thorn 32:25 ↗
Yeah, I think, look, I mean, I think your audience will know well, the Bitcoin community is reeling from this. This is a tragedy. It's different. Like I said, you know, they weren't speculating. They weren't, didn't lose it on meme coins. They didn't like going deep into the realms of DeFi and do something risky, right? Like it's the exact opposite. In fact, these are basically the most conservative cultural demographic in the entire cryptocurrency ecosystem, right? Like these are average everyday people saving in Bitcoin and storing it in the way that the Bitcoin philosophy preaches, right? And that's what hurts so much. And it really does hurt. Even by blast radius at 105 million or whatever you publicly have said the number is, and like I said, it is going to go up and we're going to publish new numbers soon that probably take it north of like 125, 130. That's still just barely in the top 20 of all crypto hacks of all time. But it's unique because it's these people did everything right. They did, they followed the ethos. They kept their keys, like right, they stack sats. Like I said, the average loss here per person is like seven Bitcoin or per address. I'll be able to get some sort of per person figures by the sample of the people that I've talked to me directly. But this isn't like, you know, a hedge fund blows up and a bunch of rich people lose money, right? Like this is everyday people. Not that that's not bad too, but like it's just different. And I think people are re-evaluating like, well, we definitely put too much trust in Coldcard, that's obvious. It's empirically true. But like, did, should people have, you know, I have heard, I know most people didn't, but I have heard when people talk about Coldcard, not all people, but some people, some people have said you should roll dice, have a strong passphrase, like, or use multisig, collaborate. I got to say, not one multisig transaction have I identified in any of the victims. Which makes sense, if you have a multisig quorum, let's say it's just two of three as a simple example, if only one of the three is a Coldcard, you're fine right now. You should rotate that one out eventually, but you know, it can't reach the signing threshold on its own, right? And
So like there are ways, you know, Casa has a great product, Unchained, Nunchuk, there's other resiliency built into things like Anchor Watch and Liana wallet, right? Like there are ways to self-custody that aren't โ€” you don't just have to hope your hardware wallet doesn't have this bug. You can diversify the risk and still keep your own coins without a centralized intermediary. There are ways. People have said that over the years, but you know, I do think we have to โ€” this is going to be, at least in the near and medium terms, a blow to self-custody as a concept. I mean, this is the scariest thing that can happen to self-custody. It's literally the worst possible. You know, a lot of people who aren't technologically sufficient and shouldn't be self-custodying their coins might lose them because they accidentally lose their seed phrase, for example, right? They don't know how to store it securely, right? Those people also shouldn't self-custody, right? But โ€” and do they know who they are? Maybe we did push the average person into self-custody unduly. I think there was a โ€” I think some Bitcoiners thought that they were not pure enough Bitcoiners, pure enough Bitcoiners, if they stored their bitcoins on an โ€” on, you know, on Coinbase or River or just owned the ETF. And I think there's going to be blowback โ€” kind of criticisms of paper Bitcoin. Um, I think there's going to be blowback because of that. I mean, on the self-custody movement, I still believe in self-custody. I believe it can be done safely. I think again, those firms that provide collaborative self-custody are very easy to use. There are people you can pick up on the phone and call that will help you. They're not that expensive and they're very good. So, um, and they help you diversify your risk and you don't even have to fully trust them, right? Like that the setup's multisig is very powerful. Like if you care about self โ€” and by the way, whether if you're good enough at it and you want to roll your own multisig, like that is possible as well. Um, just do your research. Like do not do something beyond your technical capability. But I would say the collaborative self-custody providers like Casa and Unchained are very, very good and not that hard to use for average people. Um, and so I think, you know, to the extent people self-custody still, and I still hope they will, there are ways to do it safely.
Um, I think the era of generate a single seed on a hardware wallet and store the seed in your house and the hardware wallet โ€” I think the era of single-sig self-custody is over. It should be over after this. The risk is just simply too high to trust just one hardware wallet provider. And by the way, um, the rest of crypto needs to be looking at this too. Um, I don't know a lot of multi-sig self-custody people in altcoin land, but you know, these bugs can happen in other hardware โ€” they can happen. They, you know, it's the first time it's happened at this scale. Um, I'll be surprised if it's the last time in our lifetime that we see it happen, though. So, diversify. If you're going to do self-custody, diversify and use multisig. It's an incredibly powerful tool. Learn about multisig. By the way, all the custodians use multisig and MPC. Like they're not storing your coins on a single hardware wallet in a vault somewhere. Not at all. Right.

19 more exchanges in this transcript

Sign in free to read the rest of this interview. No card required.

Sign in to read the full transcript

Cite this transcript

APA, MLA, BibTeX
APA

Thorn, A. (2026, August 7). Tracking the Coldcard Hackers: Inside the $120M Bitcoin Theft with Alex Thorn [Interview transcript]. Bitcoin Magazine. CEOInterviews.AI. https://ceointerviews.ai/interview/3265671/

MLA

Alex Thorn. "Tracking the Coldcard Hackers: Inside the $120M Bitcoin Theft with Alex Thorn." Bitcoin Magazine, 7 Aug. 2026. Transcript, CEOInterviews.AI, https://ceointerviews.ai/interview/3265671/.

BibTeX
@misc{thorn2026_3265671,
  author       = {Alex Thorn},
  title        = {Tracking the Coldcard Hackers: Inside the $120M Bitcoin Theft with Alex Thorn},
  howpublished = {Interview transcript, Bitcoin Magazine. CEOInterviews.AI},
  year         = {2026},
  month        = {aug},
  url          = {https://ceointerviews.ai/interview/3265671/},
  note         = {Speaker-attributed transcript with timestamps}
}