Insurance matters less than people expect. What decides the outcome is whether the custodian’s own records can identify what was yours. Two insolvencies with almost identical pooling produced opposite results: Cryptopia’s account holders kept their coins because the database showed who held what, and Prime Trust’s customers did not, because the internal ledger had been manipulated.
Part of our guide: Digital Asset Custody.
The short version
- FDIC insurance does not cover crypto. It covers deposits at insured banks (FDIC). SIPC does not either; it covers securities and cash at failed brokerages (SIPC).
- Private crime or specie insurance may exist, and its limits, exclusions, and who the named insured is decide whether it reaches you.
- Pooling is normal and survivable. Pooling without a reconcilable per-customer record is what turns an insolvency into a dispute.
- Whether assets are treated as customer property or as the estate’s is the question everything else hangs on, and it is answered from records and account terms.
- The time to ask these questions is at onboarding, when you have leverage.
The insurance most people think they have
Two federal schemes get invoked, and neither applies.
FDIC insures deposits at insured banks. Crypto held with a custodian is not a deposit. Where a custodian holds dollar balances at an insured bank, those dollars may be covered under pass-through arrangements; the crypto is not.
SIPC protects customers of failed member brokerages, covering securities and cash. Digital assets held at a crypto custodian sit outside that regime.
What can exist is commercial coverage, usually crime or specie insurance on the custodian’s own storage. Three questions decide whether it means anything to you: what the limit is against total assets held, what the exclusions say, and who the named insured is. A policy covering the custodian for theft is not a policy covering you for the custodian’s insolvency.
The question that actually decides it
Whether the assets are treated as belonging to customers or to the failed company’s estate. Everything else, including insurance, sits downstream of that.
Two collapses make the point better than any general explanation, because the facts were similar and the outcomes were opposite.
Cryptopia, a New Zealand exchange, pooled customer holdings. The court nonetheless found account holders retained a beneficial interest, and the reason was administrative rather than technical: the company’s own database recorded who held what, so beneficiaries could be identified. Pooling did not defeat the claim. Max Avery’s account of what the judgment actually decided covers why the usual summary of it is wrong.
Prime Trust pooled in a comparable way and its customers fared very differently. The internal record, on the estate’s own forensic accounting, could not be relied on, and the account terms disclaimed the kind of relationship customers assumed they had. Max’s write-up of the custodian that could not open its own wallets sets out how it unwound.
Same structural feature, opposite results. The variable was the quality and integrity of the internal ledger, which is precisely the thing no marketing page discusses and no insurance certificate addresses.
What to ask before you need to know
Who reconciles the internal ledger against on-chain balances, how often, and does anyone outside the company check? This is the single most predictive question, and it is rarely asked.
Are assets segregated, omnibus, or a mix? All are workable. Only the answer combined with the reconciliation practice tells you anything.
What does the account agreement say about title? Custodial language and general-obligation language read similarly to a non-lawyer and behave very differently in an insolvency.
Which legal entity holds your account, and is it the chartered one? Groups often have one regulated entity and several affiliates.
What does the latest SOC report say, including the exceptions? The exceptions are the informative part.
What is the insurance limit against total assets under custody? A large-sounding policy across a large book can be a small percentage.
What I actually see
The diligence stops at the reassuring answer. A firm says it is a qualified custodian and carries insurance, and the conversation ends there, when both facts are compatible with a bad outcome.
The questions that predict how an insolvency goes are unglamorous and operational: who reconciles, how often, and who checks them. Firms with good answers give them immediately, because they are proud of the process. Firms without good answers redirect to charters and coverage.
The other pattern worth naming is concentration. Families that would never hold one stock hold their entire digital position with a single custodian, because custody feels like infrastructure rather than a counterparty. It is a counterparty.
Where this goes wrong
The failure arrives and the questions that would have mattered were never asked.
The specific ways: coverage assumed to be FDIC or SIPC when neither applies. An account opened with an affiliate rather than the chartered entity, discovered during the insolvency. Account terms that disclaim the relationship the customer believed existed. And the central one, an internal ledger nobody outside the company ever verified, which converts an ownership question into litigation.
There is also a self-inflicted version. Customers who kept no independent record of their own holdings and balances have to accept whatever the estate reconstructs. Keeping your own statements and periodic balance snapshots costs nothing and makes you a creditor who can substantiate a claim.
The decision rule
- Assume no federal insurance applies. Confirm what commercial coverage exists, its limit, and the named insured.
- Ask who reconciles the ledger and who checks them. Treat vagueness as the answer.
- Read the account agreement on title and insolvency before signing.
- Confirm the entity holding your account is the chartered one.
- Keep your own records: statements, balance snapshots, and transaction history exported on a schedule.
- Do not concentrate the whole position with one custodian.
Where this sits
Custodian failure is the risk that custody structure exists to manage. What “qualified” means is the threshold question and it settles less than people expect. Custody for an entity is where titling gives you evidence a third party created. Your own records are what let you substantiate a claim when the custodian’s records are the thing in dispute.
The lesson that carries across all of it: the record is not paperwork about the assets. In an insolvency the record is the thing that decides who owns them.
Sources
- FDIC, Deposit insurance
- SIPC, What SIPC protects
- 17 CFR 275.206(4)-2, Custody of funds or securities of clients by investment advisers
- FINRA BrokerCheck
- IRS, Digital assets
- CFTC, Digital assets
Related
- What is a qualified crypto custodian?
- Crypto custody for LLCs
- What records should a crypto LLC keep?
- Private key succession planning
- Can a Wyoming LLC own a crypto wallet?
- Crypto custody
Last updated: 3 August 2026.
This article is general education, not legal, tax, or investment advice. Custody arrangements can reduce certain risks but do not eliminate them, and outcomes depend on your facts, your provider, and your agreements. Talk to a qualified attorney about your own situation.
