How stablecoin reserves work and how to evaluate them

Stablecoin reserves are the pool of assets an issuer holds to back every token in circulation at its target value. Reserve composition—what those assets are, how liquid they are, and how they're verified—determines whether a stablecoin can honor redemptions during normal trading and during market stress. Not all reserves carry the same risk: cash and short-dated US Treasuries behave differently from gold, Bitcoin, or secured loans when redemption demand spikes.

This article goes beyond the overview in the essential stablecoin concepts guide and explains how to actually assess what's behind a stablecoin backing claim: what issuers hold, how they prove it, what the GENIUS Act mandates, and what to look for when something doesn't add up. For foundational concepts, see the beginner's guide to stablecoins. For term definitions, see the key stablecoin terms and concepts guide.

Disclaimer: This content is for educational purposes only. It isn't financial advice, not a solicitation, and not for UK audiences. Stablecoins are risky and not suitable for all users.

Key takeaways about stablecoin reserves

Reserve assets span a wide quality range—from cash and short-dated Treasuries at the top to secured loans, gold, and Bitcoin at the bottom. Where an issuer's holdings fall on that spectrum shapes how the stablecoin performs when markets turn.

Tether's Q1 2026 attestation reports $141 billion in US Treasury and short term exposure alongside approximately $50.5 billion in non-Treasury assets including commodities, corporate bonds Bitcoin, secured loans, and equities. Circle's March 2026 report shows USDC reserves concentrated in a BlackRock-managed government money market fund with monthly Deloitte attestations.

Most issuers publish attestations—point-in-time snapshots confirming that reserves matched outstanding tokens on a single date—rather than full audits, which examine controls and financial records across an entire reporting period. That distinction has practical consequences: an attestation doesn't reveal what happened between reporting dates.

Under the proposed US GENIUS Act, only a narrow set of assets would qualify as stablecoin reserves—dollars, Treasuries, repos, and money market funds. Bitcoin, gold, and corporate investments wouldn't make the cut. If enacted, implementation would likely phase in across 2026–2027.

During normal trading, reserve composition is invisible: major dollar-backed stablecoins at or near $1. The differences only surface under stress, and they matter most for stablecoins used as collateral in leveraged positions.

The reserve quality hierarchy for dollar-backed assets

Not all dollar-denominated reserve assets behave identically when redemption pressure builds. Quality falls along a spectrum defined by how fast and reliably each asset converts to cash at face value under stress:

Stablecoins whose reserves concentrate toward the top of this hierarchy are better positioned to meet large-scale redemptions without selling assets at a loss.

Evaluating what backs major stablecoins

USDT (Tether): large and diversified, with disclosure gaps

Tether's Q1 2026 attestation, published May 1, 2026 via BDO, reports total assets of $191.7 billion against $183.5 billion in liabilities (outstanding USDT). The headline figure—approximately $141 billion in US Treasuries—places Tether among the largest global holders of US government debt. But the evaluation question isn't the Treasury position; it's everything else. Excess reserves reached $8.23 billion in Q1 2026, providing a buffer above outstanding USDT. Tether discloses this figure explicitly, which is a positive transparency signal.

What to watch: Tether publishes quarterly attestations through BDO Italia—category-level breakdowns (Treasuries, gold, Bitcoin, loans) without individual security detail. That's less granular than Circle's CUSIP-level reporting. Tether has announced a Big Four audit engagement, but no completed audit has been published as of May 2026. The 31-day lag between the Q1 report date (March 31) and publication (May 1) is standard but means a full month of reserve activity goes unverified.

Two historical enforcement actions remain part of Tether's record: the 2021 CFTC settlement ($41 million) for reserve misrepresentations during 2016–2018, and the NYAG settlement ($18.5 million) for disclosure issues and intercompany lending between Tether and Bitfinex. Both predate the current attestation regime.

USDC (Circle): concentrated and granular, with public-company oversight

Circle's March 2026 reserve report, attested by Deloitte & Touche under AICPA standards, reports USDC in circulation of approximately $77.215 billion as of March 31, 2026.

From an evaluation standpoint, USDC's reserve structure is the most transparent among major stablecoins. The majority sits in the Circle Reserve Fund (ticker USDXX)—a SEC-registered 2a-7 government money market fund managed by BlackRock and custodied at BNY Mellon, holding US Treasuries with a weighted-average maturity under 60 days plus overnight Treasury-collateralized reverse repos. The remainder is cash at regulated US banks. There are no alternative assets—no gold, Bitcoin, equities, or secured loans.

What makes it evaluable: Circle publishes monthly attestations, weekly reserve disclosures, and daily portfolio reporting with individual CUSIP-level detail—maturity dates, market values, and the financial institutions holding the cash portion. This is the most granular reserve disclosure in the stablecoin industry. The attestation-to-publication lag is typically 3–4 weeks.

Circle went public on the NYSE (ticker: CRCL) in June 2025, which adds a layer of oversight that privately held issuers don't face: full annual financial audits by Deloitte (distinct from the monthly reserve attestations), SEC reporting requirements, and public earnings disclosures. Circle also holds money transmitter licenses in 49 US states, a NYDFS BitLicense, and a French ACPR Electronic Money Institution license under MiCA.

USDS (Sky protocol) formerly DAI: onchain-verifiable, hybrid collateral

DAI, now rebranded USDS as of April 2026, uses a fundamentally different backing model—overcollateralized crypto and real-world assets locked in smart contracts rather than reserves held by an issuer. For how vaults and overcollateralization work, see the beginner's guide to stablecoins.

What makes it evaluable: DAI's collateral composition is verifiable onchain in real time, with no attestation lag or reliance on a third-party accounting firm. This makes it transparent, but the tradeoff is that DAI's collateral includes volatile crypto assets alongside Treasuries held through legal entities, making it a hybrid that doesn't fit neatly into the reserve quality hierarchy above. Evaluation here means monitoring collateralization ratios and governance votes that change the collateral mix, not reading quarterly PDFs.

mUSD (MetaMask USD): cash and Treasuries only

mUSD is MetaMask's wallet-native stablecoin, backed 1:1 by US cash and short-duration Treasuries with no alternative assets. It's issued in partnership with Bridge (part of Stripe) and powered by the M0 protocol. mUSD is integrated directly into MetaMask's self-custodial wallet for buying, trading, swapping, earning, bridging, and spending (subject to regional availability). For more details on issuance and protocol architecture, see the mUSD launch press release.

What makes it evaluable: The reserve structure is narrow by design—cash and Treasuries only, no gold, Bitcoin, equities, or secured loans—which places mUSD at the top of the reserve quality hierarchy outlined above. That narrow composition also means mUSD's reserves already fall within the GENIUS Act's proposed eligible asset list, reducing restructuring risk if the legislation is enacted.

The evaluation question for mUSD is the same one that applies to any newer stablecoin: what's the track record of independent third-party verification over time? Established issuers like Circle and Tether have years of published attestations; even where those reports have limitations, they create a verifiable history. For mUSD, that history is still being built. Evaluating it means watching for the cadence, granularity, and independence of future reserve disclosures as the stablecoin's circulation grows. That said, mUSD's issuer and infrastructure carry established credentials: Bridge is a subsidiary of Stripe, and the M0 protocol also powers PayPal's stablecoin (PYUSD) and MoonPay's enterprise stablecoin services—existing deployments that provide a reference point for evaluating the underlying technology.

How to read a stablecoin attestation report

An attestation and an audit aren't the same thing.

Attestation: An independent accounting firm verifies that, on one specific date, the issuer's stated reserve figures matched what the firm observed. Circle's reports use AICPA agreed-upon-procedures standards. They cover a single moment, but do not don't review activity between reporting dates or test the issuer's internal controls over time.

Audit: A broader examination that reviews financial records, internal controls, and operational processes across a defined reporting period. Audits are more rigorous, more costly, and far less common in the stablecoin industry. Circle undergoes full annual audits as a publicly traded company. Tether hasn't published a completed audit to date.

When reviewing an attestation report, five things matter

Who performed it: For example, Deloitte (Circle) and BDO (Tether) are established global firms. Reports from lesser-known or regional accounting practices deserve additional scrutiny.

What date it covers: Each report reflects a single snapshot. A March 31 attestation reveals nothing about reserve composition on March 15 or April 10.

Whether it specifies composition: Circle discloses individual securities—CUSIPs, maturity dates, market values—and names the banks holding cash. Tether provides category-level totals (Treasuries, gold, Bitcoin, loans) without drilling down to individual holdings.

The lag between report date and publication: Tether's Q1 2026 attestation covered March 31, 2026 and appeared May 1, 2026—a 31-day gap. Circle's monthly reports typically surface within 3–4 weeks. The shorter the lag, the less time reserves go unverified.

Whether excess reserves are disclosed: Tether publishes a specific buffer figure ($8.23 billion in Q1 2026). Circle confirms that reserves meet or exceed outstanding supply but doesn't quantify the surplus.

How the GENIUS Act changes reserve evaluation

The GENIUS Act, if enacted, would set the first federal standard for what counts as an eligible stablecoin reserve. For an overview of legislation, see the beginner's guide to stablecoins. What matters for reserve evaluation is the gap between what the act allows and what issuers currently hold.

The bill defines a narrow set of eligible assets: cash and bank deposits at FDIC-insured institutions, short-term US government debt, reverse repos collateralized by Treasuries, and SEC-registered money market funds.

What isn't eligible: Bitcoin, gold, corporate bonds, equities, secured loans, and anything not on the specified list. Tether's current mix includes approximately $5 billion in non-eligible assets. How Tether restructures—and on what timeline—is one of the defining reserve evaluation questions for 2026–2027. USDC's reserve composition already falls within the eligible asset list. mUSD's cash-and-Treasuries structure does as well.

The act also mandates monthly disclosure of reserve amounts, composition, and outstanding supply, with full audits required above certain size thresholds. Issuers can't pay interest or yield directly to stablecoin holders—the third-party yield loophole remains contested, with the OCC's comment period on implementing rules having closed May 1, 2026.

For anyone evaluating stablecoin reserves today, the GENIUS Act framework is useful even before enactment: it provides a benchmark for what US regulators consider adequate backing. Reserves that already meet GENIUS requirements carry less restructuring risk than those that don't.

Risks in stablecoin reserves

The following signals suggest a stablecoin's reserves deserve closer scrutiny. Any single flag is worth investigating; multiple flags applying to the same issuer compound the concern.

Heavy weighting toward illiquid or volatile assets: Gold, Bitcoin, secured loans, and corporate investments are harder to convert to cash at par when markets are under pressure. A reserve portfolio split 75/25 between Treasuries and alternative assets carries meaningfully different redemption risk than one holding Treasuries exclusively.

Stale attestations: If the most recent public attestation is more than 90 days old, reserve composition during that gap is effectively unknown.

No named attestation firm or no published reports: A stablecoin with meaningful circulation that lacks independent third-party verification is relying entirely on the issuer's own claims.

Vague composition disclosure: A report stating "reserves exceed liabilities" without breaking down what those reserves actually consist of—by asset class, maturity, and custodian—offers less assurance than one with itemized detail.

Concentration at unregulated or non-GSIB banks: The 2023 Silicon Valley Bank (SBV) collapse demonstrated that even a well-collateralized stablecoin can lose its peg when reserves are trapped at a failing institution. The identity of the custodian matters as much as the assets held. The beginner's guide covers the SVB depeg in detail.

No completed audit despite large circulation. Attestations verify a single date. A full audit reviews controls, processes, and records across an entire period—providing a fuller picture of how reserves are actually managed day to day.

How reserve quality affects traders

In normal conditions, reserve composition is invisible—every major dollar-backed stablecoin trades at or near $1. The differences emerge under stress, and they compound for anyone using stablecoins in leveraged or time-sensitive trading.

When a stablecoin serves as margin collateral, reserve risk sits on top of the trade's directional risk. A 2–3% depeg triggered by a reserve scare can erode collateral value enough to force a liquidation, even when the underlying position hasn't moved against the trader.

DEX settlement carries parallel exposure. If a stablecoin depegs mid-swap, the dollar value received falls short of the quoted price, with no mechanism for recourse.

The implication isn't that stablecoins should be avoided. Rather, choosing which stablecoin to hold or trade through is itself a risk decision. Reserve composition is the most important variable in that choice.

Reserve reports exist to be read and verified. The five checkpoints above apply to every stablecoin, including mUSD, which can be evaluated against the same criteria directly within MetaMask.