Every wrapped asset or custodied stablecoin you hold on Base rests on a promise you cannot see directly. A cbBTC token promises there is a real bitcoin sitting somewhere backing it. A USDC token promises there is a dollar, or something close to it, held by Circle. You cannot open that vault yourself, so the question becomes how you know the promise is being kept. Proof of reserves is the general name for the methods issuers and exchanges use to answer that question.
The problem proof of reserves is trying to solve
A wrapped or custodied token works by having an issuer hold the underlying asset off chain, or on a different chain, and mint an equivalent amount of tokens that trade on Base. Bitcoin itself does not run on Base, so a token like cbBTC has to represent it. The arrangement only works if the issuer mints tokens one for one against something it actually holds, rather than minting more than it can back.
History has already shown what happens when that assumption breaks. Several centralized exchanges collapsed after it turned out customer deposits had been spent or lent out rather than held. Proof of reserves grew into a widely used term in the aftermath of those collapses, as exchanges and issuers looked for ways to show, rather than just claim, that customer funds were still there.
Two different approaches, with different strengths
In practice, "proof of reserves" covers two fairly different techniques, and it is worth knowing which one you are actually looking at.
Attestation reports. This is the older, more traditional method. An independent accounting firm reviews the issuer's bank statements, custody records, or blockchain balances at a specific point in time, and publishes a signed report confirming that reserves matched or exceeded the tokens in circulation on that date. Circle's USDC reserve reports, reviewed monthly by Deloitte, work this way: the firm checks Circle's bank and Treasury holdings directly with the relevant institutions and confirms the figures line up with circulating USDC. An attestation is a professional, legally meaningful statement, but it is a snapshot. It tells you reserves were adequate on the day someone checked, not that they are adequate right now.
Cryptographic and on-chain proof. This method tries to make verification continuous and automatic instead of periodic. The clearest example is how Wrapped Bitcoin's custodian, BitGo, uses a Chainlink oracle to check the Bitcoin balance in WBTC's custody wallets roughly every ten minutes and publish that figure on chain, where any smart contract or user can read it and compare it against WBTC's circulating supply. Nobody has to wait for a monthly report, and nobody has to trust a single party's word for it between reports.
For exchanges handling many individual customer balances rather than one simple total, the same cryptographic idea shows up as a Merkle tree. The exchange hashes every customer's balance into a tree structure and publishes the resulting root hash. You can then get a personal proof that your specific balance was included in that tree without the exchange having to reveal anyone else's balance, and without anyone being able to quietly edit the numbers after the fact.
Where this stands for assets you might hold on Base
USDC relies on the attestation model described above, published on a regular schedule as part of the reporting the GENIUS Act now formalizes for stablecoin issuers more broadly.
cbBTC, the wrapped Bitcoin token issued by Coinbase and usable on Base, has stated since its 2024 launch that adding Chainlink based proof of reserves is part of its plan, following the same pattern BitGo already uses for WBTC. As of this writing that live, continuous on-chain feed for cbBTC has not been confirmed as fully shipped, so if this matters to you, check Coinbase's own current documentation for cbBTC rather than assuming it already works the same way WBTC does.
Either way, you do not have to take anyone's summary of this on faith. Chainlink publishes its live proof of reserve feeds at data.chain.link, where you can see the reported reserve figure for assets that use it directly. For a token's circulating supply on Base, a block explorer like Basescan will show you the total supply of the token contract, which you can compare against whatever reserve figure the issuer publishes.
What proof of reserves does not tell you
It is worth being precise about the limits here, because it is easy to read "proof of reserves" as a stronger guarantee than it actually is.
A snapshot attestation can be gamed by borrowing assets briefly to look fully funded for the one day being checked, then returning them afterward. This is sometimes called window dressing, and it is one reason continuous on-chain checks are considered a meaningful improvement over a single date stamped report. Neither method, though, proves the issuer has no other liabilities. A reserve check confirms assets exist against the tokens in circulation; it does not audit the issuer's broader balance sheet, debts, or legal structure the way a full financial audit would. And a smart contract reading an oracle feed is still only as honest as the process feeding it data in the first place, since the oracle reports what it was told to check, not an independent investigation of where the custodian's assets physically are.
None of this replaces ordinary caution. Holding a wrapped or custodied asset always means trusting an issuer to some degree, which is exactly the trade off our guide to self custody covers from the other direction. Proof of reserves narrows that trust gap and gives you something concrete to check, but it is one input into your own judgment about an asset, not a substitute for it. If an issuer publishes nothing verifiable at all, that absence is itself useful information before you decide how much of that asset to hold.
Sources: Chainlink Proof of Reserve, BitGo: Chainlink Brings Onchain Proof of Reserve to WBTC, Circle: New Levels of Detail in the Monthly USDC Attestation, Cointelegraph: Coinbase to add proof of reserves to cbBTC