The Next Crypto Exchange Test Is Whether Customers Can Get Their Money Back

Proof of reserves can show that an exchange controls assets at a particular moment. It cannot, by itself, prove that the assets are unencumbered, that customer liabilities are complete, or that withdrawals will continue when everyone wants out. The next serious exchange failure will test whether the industry has built a credible system of settlement, or simply a better way to publish reassuring snapshots.

A balance sheet is not a withdrawal promise

The central weakness in crypto exchange transparency is easy to describe. An exchange can publish wallet addresses containing billions of dollars in bitcoin, ether or stablecoins and still fail to meet customer withdrawals.

That is not a contradiction. It is a balance sheet problem.

Assets can be pledged to lenders, lent to affiliated trading firms, locked in staking contracts, held in wallets that the exchange does not control directly, or concentrated in tokens that lose liquidity when confidence breaks. At the same time, liabilities can be understated if an exchange does not disclose its full customer obligations, corporate debts, derivatives exposure and contingent claims.

The result is a gap between proof of assets and proof of solvency.

That distinction became unavoidable after the collapse of FTX in November 2022. The exchange had visible on-chain activity, a large customer base and a reputation that had attracted substantial institutional and retail capital. Yet the bankruptcy proceedings revealed a severe mismatch between the assets available to the estate and the claims made by customers and other creditors. The failure was not caused by a lack of blockchain data. It was caused by the inability to establish who owned what, where the assets were held, and whether customer funds had been used elsewhere.

The lesson for exchanges is not merely to publish more wallet addresses. It is to prove that customer claims are recorded accurately and matched to assets that can be delivered under stress.

That requires a much more demanding form of disclosure.

What proof of reserves actually establishes

A conventional proof-of-reserves exercise usually combines wallet verification with an attestation from an accounting firm or other third party. The exchange identifies addresses it says belong to the business. The auditor or verifier checks balances at a specified time, and in some cases confirms control through a cryptographic signature or a small transfer.

This can answer a useful question: did the exchange control a certain amount of identified digital assets on the date of the review?

It does not necessarily answer several more important questions.

First, are those assets customer property or corporate property? An exchange may hold tokens that belong to itself, its founders or an affiliated market maker. Those tokens can appear in the reserve total while customer liabilities remain unsecured.

Second, are the assets free of claims? Bitcoin deposited as collateral for a loan is still visible on-chain. So is ether committed to a staking contract. A reserve report that counts both without describing the encumbrance can create a misleading picture of available liquidity.

Third, are all liabilities included? An exchange can publish reserves without publishing a complete customer liability figure. Without the denominator, a large reserve number has little meaning. A platform holding $10 billion in assets could still be insolvent if it owes more than that, or if the assets are not legally available to satisfy customer claims.

Fourth, can the assets be liquidated at the stated value? The market value of a thinly traded exchange token may look substantial during normal conditions. During a run, selling a large position can push the price sharply lower. A reserve report that values a token at its last quoted price may overstate the money that can actually be recovered.

Finally, can customers withdraw the assets in the form they expect? An exchange might be solvent in aggregate but unable to meet demands for a specific coin if it has lent that coin, converted it into another asset or placed it in an illiquid protocol.

Proof of reserves therefore provides a point-in-time view of asset control. It is not a complete liquidity test, and it is not a substitute for a regulated financial statement.

The denominator is where the risk lives

The most important number in any exchange solvency analysis is not the value of its wallets. It is the total value and structure of its liabilities.

Customer liabilities are difficult to report without exposing personal information. Exchanges need to prove that millions of individual account balances are included, while keeping account identities private. Merkle tree systems can help with this problem. They allow a customer to check whether an account balance was included in an aggregate liability calculation without publicly revealing every account.

But cryptographic inclusion is not the same as accounting completeness.

A customer may be able to verify that an account balance appears in a Merkle tree while having no way to know whether the exchange omitted another liability, altered the balance before the snapshot, or included negative balances from related accounts to reduce the total. The design of the data set matters as much as the mathematical proof.

A serious process would disclose the methodology for calculating customer liabilities, including whether it covers spot balances, margin accounts, derivatives, staked assets, pending withdrawals, loans and accounts connected to affiliated entities. It would explain how disputed balances, frozen accounts and collateral are treated. It would also identify whether the exchange has obligations to market makers, lenders and payment providers that rank alongside customer claims.

This is where independent auditors matter. An on-chain analyst can verify a wallet balance, but cannot determine from blockchain data alone whether a private loan, an off-chain bank deposit or a legal claim exists. Auditors can test internal records and reconcile customer ledgers to bank and wallet balances. Regulators can examine whether the exchange has segregated customer property and maintained controls over transfers.

None of those functions replaces the others. Blockchain evidence, independent assurance and regulatory inspection cover different parts of the risk.

Segregation changes the meaning of custody

The strongest protection for customers is not a more frequent reserve snapshot. It is legal and operational segregation.

If customer assets are held in identifiable wallets or accounts for customers, and the exchange cannot freely use them for its own trading or lending, a corporate failure should be less likely to turn customers into ordinary unsecured creditors. Segregation also makes reconciliation more direct. The exchange can show that the assets held for a customer group match the balances recorded on its internal ledger.

In practice, segregation can be complicated. Exchanges may pool assets to reduce transaction costs, use omnibus wallets, or move funds between hot and cold storage. Staking can place assets with validators. Custody arrangements can involve third-party providers. These structures do not automatically create misconduct, but they must be disclosed and governed by clear legal rights.

The key question is whether an exchange can identify the customer entitlement to a particular pool of assets and prevent corporate creditors from claiming that pool. A reserve report does not establish that legal separation.

The same issue applies to stablecoins and cash. An exchange may advertise dollar backing while holding customer funds in a mixture of bank deposits, money market instruments and stablecoins. If a bank account is in the exchange's name, customers may not have a direct claim on it. If reserves are held through a stablecoin issuer, customers also inherit the issuer's redemption and counterparty risks.

The movement of capital during a crisis makes these distinctions visible. When customers become nervous, they usually seek the most liquid and widely accepted assets first, such as bitcoin, ether, major stablecoins or fiat currency. An exchange that has transformed those assets into longer-term loans, volatile tokens or affiliated investments may have enough value on paper but not enough immediately deliverable liquidity.

Rehypothecation creates hidden connections

Rehypothecation is one of the most dangerous sources of uncertainty because it connects customer assets to a chain of other balance sheets.

In a simple example, a customer deposits bitcoin on an exchange. The exchange lends that bitcoin to an affiliated market maker, which pledges it to a lender to obtain dollars, which are then used to trade other assets. The original bitcoin still exists, and a wallet may still show a related claim. But the customer cannot necessarily withdraw it immediately. Recovery now depends on the market maker, the lender and the value of the collateral.

This structure can work during stable conditions. Borrowers repay, markets remain liquid and withdrawal requests arrive gradually. It becomes fragile when confidence disappears. Customers do not ask for an accounting claim. They ask for the asset itself.

Affiliated entities make the problem harder to detect. An exchange may report strong reserves while a related trading firm holds large liabilities or depends on credit from the exchange. Transfers between the two can blur the line between customer property and proprietary capital. The collapse of FTX and Alameda Research demonstrated how dangerous those connections can become when governance and accounting controls fail.

A credible disclosure regime should identify related-party transactions, loans, collateral arrangements and limits on the use of customer assets. It should distinguish assets held for customers from assets owned by the exchange. It should also state how quickly each major reserve category can be converted into an asset available for withdrawal.

That last measure is more informative than a single reserve total.

Real-time attestations may help, but frequency is not assurance

The industry is moving toward more frequent, and in some cases near real-time, reserve reporting. Exchanges can publish wallet balances continuously, use application programming interfaces to update customer liabilities, and provide dashboards showing reserve ratios by asset.

This is a useful direction because it can reveal large transfers, falling balances or sudden changes in the composition of reserves. Capital flows often provide an early warning. If customers are steadily moving assets away from an exchange, if stablecoin balances are shrinking, or if the platform is relying more heavily on its own token, the trend may matter before a formal failure.

However, real-time data can create a more polished version of transparency theater if it is not paired with independent controls.

A live wallet balance does not reveal whether assets are borrowed. A live liability figure can be incomplete or subject to manipulation. An automated dashboard may show a reserve ratio of more than 100 percent while excluding off-chain debts and contingent obligations. More frequent publication may simply allow an exchange to present a continuously updated but narrow view of its finances.

Real-time reporting also creates a false sense of security if customers assume that a visible balance is equivalent to immediate redemption. Liquidity can disappear between blocks. A large wallet can be frozen by a custodian, affected by a smart contract exploit or trapped in an asset whose market depth is too shallow for a mass sale.

The right standard is not whether an exchange updates its dashboard every minute. It is whether an independent party can verify the data, test the controls and challenge the assumptions behind it.

Stress testing the withdrawal queue

The practical test for an exchange is a withdrawal stress test.

Such a test would model what happens if a significant percentage of customers request withdrawals over a short period, with the heaviest demand concentrated in the most liquid assets. It would account for blockchain settlement times, banking cutoffs, redemption windows, gas fees, market depth, collateral calls and the possibility that counterparties stop responding.

The test should distinguish between assets available immediately, assets available within one day, assets available within several days and assets that cannot be liquidated without a substantial discount. It should measure the effect of selling exchange-issued tokens or affiliated assets. It should also test a scenario in which stablecoin redemptions slow or a major banking partner suspends transfers.

A reserve ratio of 105 percent may sound safe, but it offers little protection if 20 percent of reserves are locked, 15 percent are held in an illiquid token and customer claims are concentrated in bitcoin that has been lent out. The same ratio could mean something very different if nearly all reserves are segregated, unencumbered and held in assets that can be transferred immediately.

Stress tests should be independently reviewed and repeated without giving management enough time to rearrange balances for presentation. A single announced snapshot invites window dressing. Surprise examinations are more valuable because they show how the balance sheet looks during ordinary operations, not only after assets have been moved into the right wallets.

Regulators are likely to focus on claims, not just coins

Regulatory scrutiny is increasingly moving toward custody, capital requirements and customer protection. In the United States, the Securities and Exchange Commission, the Commodity Futures Trading Commission, banking regulators and state authorities oversee different parts of the digital asset market. Other jurisdictions have introduced licensing regimes that set requirements for custody and reserve management.

The details differ, but the direction is consistent. Regulators want to know whether customer assets are protected if the platform fails, whether the operator can use them without permission, and whether the company has enough liquid capital to meet obligations.

This approach reflects a broader financial principle. Solvency is not merely an inventory of assets. It is the ability to satisfy legally enforceable claims as they mature.

Crypto exchanges have often resisted full disclosure by arguing that transparency must not expose wallet security, trading strategies or customer privacy. Those are legitimate concerns. They do not justify vague reserve claims. Privacy-preserving cryptography, audited controls and standardized reporting can reveal enough to test solvency without publishing sensitive account information.

The industry also needs to explain the limits of each assurance. A report that is explicitly called a limited-scope attestation should not be marketed as a full audit. A reserve snapshot should identify its date, methodology and exclusions. Customers should not have to infer whether the figures include loans, related parties or locked assets.

The next failure will be a test of market discipline

Capital tends to flow toward venues that offer speed, yield and convenience until trust becomes more valuable than return. In a calm market, customers may tolerate opaque lending, low fees or the use of an exchange token. When confidence breaks, that preference reverses. Liquidity migrates toward self-custody, regulated custodians, bank accounts and venues perceived to have clearer legal protections.

The next exchange failure will show whether proof of reserves has changed that behavior or merely changed the language used to defend it.

A credible exchange should be able to answer five questions clearly. What assets does it control? Which of those assets belong to customers? What total claims does it owe? Are the assets unencumbered and immediately available? Can an independent party verify that the answers remain true during a withdrawal shock?

If the answer to any of these questions is unclear, a reserve total is incomplete.

The most meaningful standard is therefore not a photograph of wallets. It is a continuous, independently tested demonstration that customer claims are matched to segregated and liquid assets. That demonstration should include liabilities, related-party exposure, rehypothecation, legal ownership and stress scenarios.

Crypto was built around verifiable ownership and transparent settlement. Exchanges should be held to a standard that reflects those principles. The market does not need another reassuring number. It needs evidence that, when confidence disappears and money rushes for the exit, the exit is actually open.