Categories
Bitcoin Mining

A mining operation generates Bitcoin every day, which makes custody an operational problem, not a one-time choice. How miners actually secure and sell it.

Buying a hardware wallet once and writing down a seed phrase is a reasonable answer to securing a single purchase of Bitcoin. It is not a complete answer for a mining operation, which generates new BTC continuously, has multiple people who need some level of access, and eventually has to convert a meaningful fraction of it back into fiat to pay the power bill. Custody for a mining business is an ongoing operational system, not a one-time decision, and most of the guidance written for individual holders does not transfer cleanly.

This is a different question from whether to hold or sell what you mine. That strategic choice is covered in treasury management for miners. This article covers what happens after that decision is made: where the coins actually sit, who can move them, and how a sale gets executed once you decide to make one.

Where coins sit between payout and decision

Pool payouts land somewhere automatically, and that default destination is usually the weakest link in the whole chain because it was chosen for convenience during setup rather than security at scale. Three tiers are worth separating deliberately rather than leaving to whatever the pool dashboard defaulted to.

Hot wallet, operational float. A small balance, software-controlled, used only for near-term needs – typically the amount you expect to convert to cover costs in the coming weeks. Convenience here is the point; this tier should never hold more than you are comfortable losing to a compromised machine or a phishing mistake, because it is the tier most exposed to both.

Cold storage, the core holding. Keys generated and kept on a device that is never connected to an internet-connected computer for signing, with the signing step happening offline. This is where the bulk of anything not immediately needed should sit, and it is the tier that most individual-holder advice is written for. For a business, the question that advice does not answer is who holds the device and what happens if that person is unavailable – which is exactly the single point of failure the next tier is designed to remove.

Multisignature, for anything material. A multisig setup requires more than one key to authorize a transaction – commonly two of three, or three of five – with keys held by different people, in different physical locations, ideally using different hardware. No single person, device loss, or coerced employee can move funds alone. For a mining operation with more than one stakeholder, this is not an advanced option; it is the baseline that prevents a departing partner, a disgruntled employee, or a single burglary from becoming a total loss.

The July 2026 incident covered in the Coldcard firmware vulnerability is a useful illustration of why device choice and firmware diligence matter even within the cold-storage tier itself – a hardware wallet is a mitigation, not a guarantee, and the devices holding your keys need the same patching discipline as anything else in your security stack.

Institutional custody as an alternative to self-custody

Self-custody is not the only professional option, and for some operations it is not even the right one. Regulated custodians hold keys on a client’s behalf under insurance and compliance frameworks that an individual operator cannot replicate alone – useful for a business that wants an institutional counterparty for audit, financing, or insurance purposes, or that simply does not want key management to be a function its own staff performs.

The tradeoff is counterparty risk in exchange for operational risk: you are trusting the custodian’s security and solvency instead of your own team’s discipline. For a company large enough to negotiate real terms – segregated accounts, proof-of-reserves attestations, insurance coverage that names your holdings specifically – that trade can be a sound one. For a smaller operation, the custodian’s minimum balances and fee structure often make self-custody the more practical choice regardless of the security argument either way.

Many operations land on a blended model: self-custodied multisig for the core long-term holding, an institutional custodian or exchange account for the working balance that needs to be liquid, and clear, written rules for how much sits in each at any time. The rules matter more than the exact split – a policy that nobody follows is not a policy.

Executing a sale without moving the market against yourself

Once the decision to sell has been made, how the sale is executed is its own problem, particularly for the size of transaction a profitable mining operation can generate at a single payout.

Selling directly on an exchange order book works fine for small amounts. Past a certain size, a market order walks the book and fills at a progressively worse average price, and even a series of smaller orders over a day can move the price against you if the pattern is visible. This is where an OTC desk earns its spread: it fills a large order at a single negotiated price, off the public order book, without the market impact of the same size hitting an exchange directly. The cost is the spread itself, which is the price of certainty and discretion rather than a hidden fee.

Three things are worth checking before picking a counterparty for either route. Settlement time – whether fiat arrives same-day or takes several business days, which matters if the sale is covering a near-term obligation. Counterparty risk – whether you are trusting an exchange’s solvency between trade and settlement, which is the lesson most directly drawn from past exchange failures. And price verification – getting an independent reference price before accepting an OTC quote, since an OTC desk’s price is negotiated, not posted, and negotiated prices are only as good as the benchmark you compare them against.

Access control is a personnel problem, not just a technical one

Most custody failures in mining operations are not sophisticated hacks. They are a departing employee who still has wallet access, a single founder who was the only person who knew the recovery process, or a seed phrase written down somewhere a cleaning crew could find it. A written access policy – who can view balances, who can initiate a transaction, who must co-sign, and what happens to each of those permissions when someone leaves – prevents more loss than any additional layer of encryption.

That policy should be reviewed whenever headcount changes, and it should assume that any one person might become unavailable without warning. A custody system that depends on one irreplaceable person is not actually secure; it is fragile in a way that looks secure until the day it is tested.

Ongoing hygiene, not a one-time setup

A custody system set up correctly on day one degrades quietly if nobody maintains it. A short set of recurring habits catches most of the drift before it becomes a loss.

Test transactions before large transfers. Sending a small amount first and confirming it arrives correctly, every time a new destination address or a new piece of signing hardware is involved, catches a mistyped address or a malfunctioning device before it costs the full amount rather than a trivial test sum.

Rotate signers when access changes. A multisig scheme is only as current as its signer list. When a co-signer leaves the company, loses a device, or changes role, the key set needs to be reconstituted rather than left as-is with an informal understanding that the departed person “won’t use it.” Informal understandings are not a security control.

Verify reserves periodically, not only when asked. If funds sit with an institutional custodian or exchange, a periodic proof-of-reserves or balance confirmation, checked against your own internal ledger, catches a discrepancy while it is still small and explainable rather than after it has compounded into something that triggers a crisis.

Keep the recovery process documented and tested, not just written down once. A recovery procedure nobody has actually rehearsed tends to have a gap that only appears during the real recovery attempt – a missing passphrase, an outdated firmware requirement, a backup stored somewhere that changed. Testing the recovery process on a small amount periodically is cheap insurance against discovering the gap when it is not small.

What this does not replace

None of the above changes the hold-versus-sell math. A well-secured wallet holding coins through a downturn is still exposed to the price risk covered in treasury strategy for miners, and a poorly timed OTC sale is still a poorly timed sale no matter how cleanly it was executed. Custody and execution decide whether the coins you intend to keep are actually safe and whether the coins you intend to sell actually convert at a fair price. They do not decide which coins belong in which category – that is a separate decision this article deliberately leaves to the strategy it links to.

Custody is rarely the exciting part of running a mining operation, and that is precisely why it gets underinvested relative to the engineering that produces the coins in the first place. A facility that spends carefully on redundant power and carefully sized generators, and then leaves its wallet access to an informal arrangement nobody has reviewed since launch, has simply moved the single point of failure from the power system to the balance sheet.

If you want a second opinion on how your operation’s custody and settlement setup compares to what other mining businesses run, get in touch.

Explore Rax Mining

Categories