Self-Custody
BTCPay Server Wallet Drain: Check Your Own Setup
BTCPay Server hot wallets on 2.4.1 or older were swept of Bitcoin and Litecoin on November 1. Run this checklist to see if your own setup shares the exposure.
On November 1, 2026, BTCPay Server posted a warning that was short on comfort and long on urgency: reports were coming in of hot wallets on BTCPay Server instances being swept by unauthorized attackers, and every operator should run the shutdown script and keep the instance off until the investigation finished. A day later the project narrowed it down. In its incident notice, BTCPay said Bitcoin and Litecoin were stolen from hot wallets on servers running version 2.4.1 or earlier, and that the attackers used a different method to exploit the vulnerability it had already fixed in 2.4.2 in August. The BTCPay Server wallet drain is still being pieced together, but the shape of the problem is clear enough to learn from.
BTCPay Server is self-hosted, widely respected Bitcoin payment software, and plenty of serious operators run it precisely because no third party holds their money. That is exactly why this incident is worth a few minutes of your time: it shows what "non-custodial" can still miss when a hot wallet sits somewhere between a payment and your cold storage. Use the checklist below to see whether your own setup, mining or otherwise, has the same shape of exposure.
What happened in the BTCPay Server wallet drain
According to BTCPay Server's own notice and the reporting that followed, the sequence looks like this:
- At 13:46 UTC on November 1, a wallet drain described in the first incident reports took place. The project says it has not established when attackers first began exploiting the flaw.
- At 20:56 UTC the first theft report arrived by email. At 22:25 UTC BTCPay warned operators on X and asked them to run
./btcpay-down.sh. It also asked hosting providers, LunaNode among them, to take instances offline as a precaution. - Founder Nicolas Dorier said the attacker logged into affected instances as an administrator. How that access was obtained had not been explained at the time of writing.
- On November 2 the project identified 2.4.1 (released July 23, 2026) and earlier as the affected versions. It said it had seen no impact on servers already running 2.4.2 or later when the warning went out, and its own public servers on 2.4.7 were not attacked.
- The vulnerability itself is the one fixed in 2.4.2 on August 7, 2026, when attackers took Lightning credential files from older servers. This time they reached it a different way, and went after on-chain hot wallets.
- No loss total or count of affected servers has been published. BTCPay warned that some activity at the attackers' addresses predates the reported thefts, so those balances are not a confirmed total.
The honest summary: self-hosted payment software that kept spendable keys online was entered through an admin path, and whatever sat in its hot wallet could be moved by someone other than the operator. BTCPay's own note puts the other half plainly: funds in wallets whose spending keys live off the server, hardware wallets included, were not affected.
The checklist: does your setup share the BTCPay Server exposure?
Skip the question of whether you run BTCPay Server and ask whether anything in your own payment or mining setup creates the same shape of risk. Walk through this with your own stack in mind.
- A hot wallet. Does any software you run keep spendable keys online and move funds on its own, rather than just watching an address whose keys live elsewhere?
- An open admin door. Is the admin interface for that software reachable from the open internet, protected by nothing stronger than a password?
- An old version. Are you on the current, patched release, or has it been sitting untouched since the summer?
- An intermediate wallet. When value arrives, does it land in a pooled or intermediate wallet that something else later sweeps out to you, rather than at an address you already control?
- Unknown sweep keys. If that software sweeps funds automatically, do you actually know who or what holds the keys that authorize the sweep?
Answering yes to two or more of these means you carry a version of the exposure that just played out at BTCPay Server. Patch, put that admin interface behind something better than a bare login, and find out where the keys to any sweep live.
Why a payout that never touches a pool wallet is a different risk class
Mining payouts raise the same question in a narrower form. When a solo-mined block pays out, the coinbase transaction can either send the reward to an address the pool holds, which then needs a later transfer to reach you, or it can pay your own address directly on the chain, with nothing in between to sweep. Those two designs are not cosmetically different. The first one creates exactly the kind of hot wallet that was just drained elsewhere. The second one doesn't create one at all.
That second design is the premise behind NexusPool, and the technology page walks through how the coinbase is built. With a 0% pool fee and the coinbase paying the finder's own address, there is no pooled hot wallet on NexusPool's side for an attacker to find, which is a structural fact, not a security promise. The same holds on the Litecoin side of NexusPool, which matters this week because Litecoin was among the coins taken from BTCPay hot wallets. You can see the payout for yourself before it matters, by checking the exact coinbase transaction your address would receive with NexusPool's Payout Preflight tool, well before any block is found.
Removing the pool wallet from the payout path doesn't make the rest of the pool's work invisible, either. Each hour, NexusPool signs a receipt covering the shares it counted from your rig, the difficulty it served them at, and the window they landed in, using a key it publishes. That's the Glass Ledger, and it exists because "nothing to drain" and "nothing to verify" are not the same claim. You can check those signed receipts yourself rather than taking anyone's word for what was counted. The whole reward goes to your address. Check it before you mine.
None of this is a claim that self-custodial software is risk-free, that NexusPool's own source is published for inspection today (it is not), that any pool changes how often a block is found, or that mining is an investment with a return. Solo mining is still pure variance, and network difficulty sets those odds identically at every pool, NexusPool included. What changes, when a payout skips the pool wallet entirely, is which failure mode is even possible. A pool key that was never created can't be stolen. Your own wallet is still yours to secure.
If you checked yes, or if you checked no
If you checked yes to two or more boxes: treat it as this week's task, not next month's. If any of it is a BTCPay Server instance on 2.4.1 or earlier, the project's advice goes further than an update, because updating cannot remove access an attacker may already have. Sweep the remaining hot-wallet funds into a new wallet made with a new seed on a trusted device, take the old server offline and keep its logs, build a fresh server with the latest release, and replace every password and key the old machine could reach. An unchanged balance does not prove the server was never entered.
If you checked no across the board, because value already lands directly at an address only you control: good, but "no hot wallet" still isn't the same as "nothing to check." Verify the mechanics for yourself instead of taking a pool's word, or a blog's word, including this one.
Trust nothing. Verify where your own payout lands before a block is ever found.