Your balance can look perfectly normal right up to the moment it becomes unavailable.
That is the hard part of relying on a crypto custody provider: a wallet balance is not the same thing as immediate, legally protected access to your coins. If the provider enters bankruptcy, withdrawals can stop at once. Trading may stop too. Then the question is no longer “How much crypto do I have?” It becomes “Who legally owns the assets behind this balance?”
For retail users, that distinction often sits quietly behind an intuitive app workflow: deposit, trade, earn, withdraw. For institutions, it hides behind polished custody dashboards, omnibus wallets, and reassuring security language. But bankruptcy is where the paperwork, wallet architecture, and asset segregation practices get stress-tested.
A crypto custody provider can be excellent at cold storage, multi-signature controls, and phishing protection while still leaving customers exposed to insolvency risk. Security against hackers and protection from a custodian’s creditors are related problems, but they are not the same problem.
A custody setup is only as safe as its legal ownership structure when the company holding the keys fails.
The legal reality: your account balance may not be your property in bankruptcy
When a crypto custodian files for Chapter 11 in the United States, an automatic stay under Section 362 of the Bankruptcy Code generally freezes the company’s assets. Customers cannot simply log in, press “withdraw,” and move on. The platform’s assets are locked while the court determines what belongs to the bankruptcy estate and what, if anything, is clearly customer property.
This is the part that can feel counterintuitive. Crypto users are used to a simple rule: if you control the private keys, you control the coins. A custodial arrangement changes that. You do not hold the keys; the provider does. Your rights depend on the custody agreement and on how the provider actually handled the assets.
The Celsius bankruptcy offered the most visible lesson. In January 2023, the court found that assets placed in Celsius Earn accounts belonged to the bankruptcy estate because the platform’s terms transferred title to Celsius. A later November 2023 ruling reached a similar conclusion for retail borrower accounts.
That outcome was not about whether customers thought of the assets as “their crypto.” They obviously did. It was about the contractual terms they accepted and the legal ownership they gave up in exchange for the platform’s service.
This is why a digital asset custodian should never be assessed only by its security marketing. “Cold storage,” “institutional-grade protection,” and “insured infrastructure” can all be useful features. None automatically answers the ownership question.
Here is the practical split:
| Custody feature | What it helps protect against | What it does not automatically solve |
|---|---|---|
| Cold storage | Remote hacks, hot-wallet compromise | Custodian insolvency or creditor claims |
| Multi-signature controls | A single compromised key or employee | Whether customer assets are legally segregated |
| Proof of reserves | A snapshot of on-chain assets held | Liabilities, asset ownership, bankruptcy treatment |
| Insurance arrangements | Certain defined theft or operational losses | A broad guarantee against a company bankruptcy |
| Account terms saying “you own crypto” | Can support a customer-property argument | The damage caused by commingling or weak legal structure |
The workflow is less frictionless than “not your keys, not your coins,” but it is more useful. You need to understand both the technical custody layer and the legal layer.
Why commingling and terms of service can cost you
Commingling means customer assets are mixed together, or mixed with a company’s own assets, without a sufficiently clear legal and operational separation.
An omnibus wallet is not automatically a red flag. Many exchanges and institutional crypto custodian services use pooled wallet infrastructure because it is efficient. The problem begins when the provider cannot show whose assets are whose, whether the customer assets were held in trust, and whether the company treated them as distinct from its own property.
Prime Trust became a sharp example of that risk. After its Chapter 11 filing in August 2023, the legal fight centered on mixed customer fiat and crypto assets. In July 2025, a Delaware bankruptcy judge ruled that commingled customer funds and crypto were property of the bankruptcy estate rather than customer-owned trust property. The ruling pointed to the absence of a clear trust relationship in the agreements and what was described as “hopeless commingling.”
That phrase should land heavily. Once records and wallets are muddled, a customer may have a claim against the bankrupt company rather than a direct claim to specific coins.
For a user, the difference is enormous:
1. Directly segregated assets may be easier to identify as customer property. If the custody architecture, contracts, and books all point in the same direction, the customer has a stronger foundation for seeking return of assets.
2. Commingled assets can turn customers into unsecured creditors. In that scenario, customers stand in line with other unsecured claimants and may face a long, uncertain recovery process.
3. Terms of service can override assumptions created by the interface. A dashboard calling something “your BTC” does not settle the legal question if the agreement says title transferred to the platform.
4. Operational behavior matters alongside written promises. A provider can write a customer-friendly clause, then undermine it by pooling assets poorly or using them in a way that blurs ownership.
This is where we should be realistic about proof of reserves as well. Proof of reserves can be a helpful transparency signal, particularly when an exchange publishes wallet data and liabilities methodology. But reserves are not a complete custody opinion. They do not, by themselves, establish that customer assets are bankruptcy-remote, unencumbered, properly segregated, or protected from creditors.
A reserve report may tell you that coins exist. Bankruptcy asks a more uncomfortable question: whose coins are they under the law?
The 90-day surprise: withdrawals can face clawback claims
Even users who withdrew before a failure may not be completely outside the blast radius.
Under Section 547 of the US Bankruptcy Code, a bankruptcy estate can try to claw back certain transfers made in the 90 days before a filing. These are known as preference claims. The general theory is that one creditor should not receive a better deal than similarly placed creditors shortly before bankruptcy.
Crypto makes this messier because withdrawals often feel final. On-chain, they are final. Legally, the transaction can still be challenged.
The Prime Trust estate has pursued this route. In May 2026, the Prime Trust Litigation Trust sought roughly $970 million from Swan Bitcoin and about $29.5 million from Strike in clawback lawsuits tied to pre-bankruptcy transfers. Those filings do not mean every customer withdrawal will be reclaimed, and defenses can matter significantly. But they show why “I got my coins out in time” is not always the end of the analysis.
For ordinary customers, the practical lesson is not to panic about every withdrawal. It is to understand that a failing custodian has a legal tail. The closer a provider is to insolvency, the more its transaction history may be examined later.
In crypto, an on-chain withdrawal can be technically complete and still become legally disputed.
This is also why a sudden rush of withdrawals is not a clean safety signal. It may be the right instinct when you want control of your assets, but it does not repair a weak custody structure that existed for months or years before the crisis.
What bankruptcy-remote custody is supposed to change
Bankruptcy-remote custody is the phrase worth learning if you hold material balances with a third party.
In simple terms, it describes a structure designed to keep customer assets outside the custodian’s bankruptcy estate. The provider may administer the assets, but its creditors should not be able to treat them as property available to pay the provider’s debts.
In the US, one route involves Article 8 of the Uniform Commercial Code. Crypto can potentially receive more robust treatment when a custodian acts as a “securities intermediary,” agrees to treat the crypto as a “financial asset,” and maintains it in a designated securities account.
That may sound like legal plumbing. It is. But it is exactly the kind of plumbing that matters when the lights go out.
A provider claiming Article 8 treatment should be able to explain, in plain language:
- whether its customer agreement explicitly treats the digital assets as financial assets;
- whether the customer receives a designated securities account rather than an undefined platform balance;
- whether the custodian acts as a securities intermediary for that account;
- how customer positions are recorded, reconciled, and segregated;
- whether assets are pledged, rehypothecated, lent, or otherwise exposed under any service tier;
- what happens if an affiliate, sub-custodian, or exchange in the custody chain fails.
For larger holders, this is where the difference between an exchange account and a dedicated custody relationship becomes meaningful. An exchange is optimized for trading liquidity and a fast workflow. A custody provider may be optimized for governance, segregation, settlement controls, and institutional reporting. Neither model is automatically safer in every respect. But they solve different jobs.
If you trade frequently, keeping a working balance on an exchange can be efficient. If you are holding strategic treasury assets, long-term reserves, or client assets, the custody integration needs a higher standard. You want documented ownership, controlled withdrawal rules, clear sub-custody disclosures, and a clean path to asset return.
A good provider does not make you decode this through legal fog. It gives you an intuitive explanation, then backs it with contracts and operating evidence.
A practical custody review workflow
When we review a crypto custody provider, we would put these questions ahead of flashy yield offers or a sleek mobile app:
1. Who holds legal title?
Read the specific custody terms, not the homepage. Look for language around ownership, title, beneficial interest, trust arrangements, liens, and the provider’s rights over assets.
2. Are assets segregated in practice?
Ask whether segregation is on-chain, ledger-based, account-based, or merely promised. A pooled wallet can work, but the reconciliation and legal structure need to be strong.
3. Is there a sub-custodian?
Many platforms do not hold everything themselves. If another exchange, bank, or specialist custodian is in the chain, you are assessing more than one balance sheet and more than one agreement.
4. What does “insurance” actually mean?
Crypto custody insurance often covers specific risks such as theft, employee misconduct, or physical security incidents. It is not a substitute for insolvency protection, and digital assets held by a crypto custodian are not covered by FDIC insurance in the way bank deposits are.
5. Can you verify controls beyond marketing claims?
Look for independent security audits, proof-of-reserves methodology, published incident procedures, multi-signature governance, withdrawal allowlists, and transparent reporting on asset storage.
6. Can you move assets without unnecessary friction?
A secure provider should still have a usable exit path. Long holds, opaque approval queues, and vague withdrawal limits matter more when markets are moving quickly.
The phrase “screening test” applies here more than most people think. There is even a screening test of new TV series with aliens, but custody deserves a much tougher version: not whether the service looks polished on day one, but whether its structure still works when the company is under legal pressure.
MiCA raises the baseline in Europe, but it is not a magic shield
The European Union’s Markets in Crypto-Assets regulation, better known as MiCA, introduced a more direct framework for crypto-asset service providers. From December 30, 2024, applicable rules require CASPs to segregate clients’ crypto-assets from their own holdings.
That is a meaningful shift. It places customer asset protection closer to the center of the regulatory framework rather than treating it as a voluntary best practice.
For users of EU-regulated providers, this should improve the baseline in several ways:
- the provider must distinguish client crypto-assets from proprietary holdings;
- creditors should have a harder time reaching properly segregated customer assets if the provider fails;
- custody processes face more formal regulatory expectations;
- users have a clearer basis for asking how the provider handles client asset records and wallet controls.
Still, regulation is not a substitute for due diligence. MiCA can set rules, but customers should still ask how a particular provider implements them. Segregation can be technically complex. Wallet management may involve multiple chains, staking arrangements, liquidity providers, and third-party custody infrastructure.
The words “regulated” and “safe” should never become interchangeable. They are useful signals, not the final answer.
The security features that matter before a crisis
A crypto exchange hack is dramatic. A slow custody failure is usually not. That is exactly why users often pay more attention to the first one.
The best custody setup is built to handle both. It reduces the chance of theft while also limiting the chance that customer assets get trapped in a provider’s corporate failure.
For retail users, the sensible arrangement is often layered:
- Keep only active trading capital on an exchange.
- Use strong two-factor authentication, ideally an authenticator app or hardware security key rather than SMS alone.
- Turn on withdrawal address allowlisting and anti-phishing codes where available.
- Store long-term holdings in a self-custody wallet if you are genuinely comfortable managing seed phrase security.
- If using third-party custody for significant amounts, choose a provider whose customer asset terms you can actually understand.
For institutions, the workflow gets more formal. Dedicated accounts, multiple approval roles, withdrawal policies, audit trails, disaster recovery processes, and legal segregation should all work together. Multi-signature security is useful, but it is not the whole picture. If the same bankrupt entity controls the operational structure, the contract, and the asset records, key distribution alone does not make the assets bankruptcy-remote.
The goal is not to avoid every custodial service. That would be impractical for many traders, funds, and businesses. The goal is to know what job the custodian is performing, what risks you are accepting, and whether the arrangement remains functional under stress.
The answer, without the legal fog
If a crypto custody provider goes bankrupt, your assets may be frozen immediately. Whether you get them back promptly, recover only part of their value later, or become an unsecured creditor can depend on the custody contract, the provider’s asset segregation, its recordkeeping, and the jurisdiction involved.
A strong crypto custody provider gives you more than cold storage and a comforting insurance badge. It offers a clear legal framework, credible operational separation, transparent controls, and an exit workflow that does not become mysterious when conditions turn ugly.
For active retail traders, a reputable exchange can still be the right tool for liquidity and convenience—just keep the balance purposeful. For long-term holders with meaningful exposure, self-custody or a dedicated, clearly structured institutional crypto custodian is usually the more sensible path. And for any business holding client or treasury assets, bankruptcy-remote custody should be a core design requirement, not a phrase discovered after the withdrawals stop.
