After several high-profile exchange failures, proof of reserves became a standard transparency claim: exchanges publish cryptographic evidence that they hold the crypto assets they say they hold, inviting users to verify their own balances are included. That is genuinely useful, but it is also widely misunderstood. Proof of reserves shows assets, not solvency, and an exchange can pass a proof-of-reserves check while still being insolvent if its liabilities exceed its assets or if its assets are encumbered. Understanding both sides of that distinction is essential before trusting any exchange with significant funds.

A typical proof-of-reserves process uses a Merkle tree, a cryptographic structure that lets each user verify their individual balance is included in the exchange's total reported holdings without revealing other users' balances. The exchange publishes the tree's root hash and a list of wallet addresses it controls, and an independent auditor or verification tool checks that the on-chain balances of those addresses match or exceed the total user liabilities in the tree. Users can then verify their own leaf in the tree using their account data, confirming the exchange counted their funds.

This design is clever and genuinely informative: it proves the exchange controls at least a stated amount of assets at a specific snapshot in time. But snapshots are the first limitation. An exchange could borrow assets for the snapshot date and return them afterwards, a practice that audits have sometimes caught only through timing analysis. Reputable exchanges therefore publish proofs regularly, use independent auditors, and include liabilities data, because a proof that is infrequent, self-attested, or assets-only is far weaker than one that is frequent, independently verified, and paired with liability disclosure.

The critical omission is liabilities beyond customer crypto balances. Proof of reserves does not reveal the exchange's debts, its obligations to lenders, its off-balance-sheet positions, or whether the assets shown are pledged as collateral elsewhere. An exchange holding one billion dollars of customer crypto while owing 1.2 billion dollars to creditors is insolvent despite a clean proof of reserves. It also does not show the quality of assets: a reserve composed largely of the exchange's own illiquid token is worth far less in a crisis than the same nominal value in Bitcoin or stablecoins, because selling it to meet withdrawals would crash its price.

Other gaps include fiat balances, which are harder to verify on-chain and may sit in bank accounts subject to different risks, and the segregation of funds, meaning whether customer assets are truly separate from the company's operating funds. Proof of reserves also says nothing about management quality, regulatory compliance, or cybersecurity. An exchange can publish perfect proofs and still suffer a hack, a regulatory shutdown, or a bank run driven by rumours, as past failures have shown repeatedly.

Use proof of reserves as one input among several. Prefer exchanges that publish frequent, auditor-verified proofs including liabilities, that disclose the composition of reserves with a high share of liquid, non-native assets, and that submit to regulatory oversight in a reputable jurisdiction. Check whether the exchange's auditor is a recognised firm and whether the audit scope covers both assets and liabilities, not just a wallet-balance attestation. Watch for red flags such as long gaps between proofs, auditor changes, or refusal to disclose liability figures.

Most importantly, limit your exposure regardless of proofs: keep only trading balances on exchanges and move long-term holdings to self-custody. Proof of reserves reduces one specific risk, the risk that an exchange secretly lacks the crypto it claims, but it does not eliminate counterparty risk entirely. The holders who survived past exchange failures were rarely those who analysed proofs most cleverly; they were those who kept most of their funds where no exchange failure could reach them.

Proof of reserves proves an exchange held stated assets at a snapshot moment; it does not prove solvency, asset quality, or safety from hacks and bank runs. Use it as one check among several, prefer frequent auditor-verified proofs with liability disclosure, and keep long-term holdings in self-custody regardless.