The word "backed" gets used loosely with stablecoins. You'll see "100% backed" on marketing materials, in congressional testimony, and in audit summaries — but what that means in practice depends entirely on the reserve model, who's verifying it, and what assets actually sit in the vault.
There are three mechanically distinct approaches to stablecoin reserve management. They're not variations of the same idea. They have different failure modes, different regulatory footprints, and different degrees of verifiability. Conflating them — or assuming "stablecoin" implies any particular backing structure — leads to real analytical errors.
The Three Reserve Models
Fiat-collateralized stablecoins hold traditional financial assets to match the tokens in circulation. For every $1 of token outstanding, the issuer holds roughly $1 in some combination of cash, short-term treasury bills, money market funds, or similar instruments. USDT (Tether) and USDC (Circle) operate this way, though their specific reserve compositions differ significantly.
The mechanics: users deposit dollars, the issuer mints tokens, and the deposited dollars flow into the reserve portfolio. When users redeem, tokens are burned and dollars are returned. The token doesn't represent a claim on any particular asset — it represents a claim on the issuer's balance sheet.
Where the complexity lives: not all reserves are equally liquid or safe. The difference between holding 3-month US Treasury bills versus commercial paper versus bank deposits versus crypto assets is substantial. Early Tether reserve disclosures revealed heavy commercial paper holdings — less liquid and carrying credit risk. Under regulatory and market pressure, both Tether and Circle shifted toward shorter-duration, higher-quality assets. USDC now holds primarily cash and short-term US Treasuries. Tether's reserve composition has improved but still includes some non-cash equivalents.
Verification happens through attestations — a third-party accounting firm checks that stated assets exist at a point in time. This isn't a full audit. It doesn't verify internal controls, historical accuracy, or the quality of counterparties. Full reserve audits are the standard being pushed toward by regulators, but as of early 2026, neither major issuer has completed one.
Crypto-collateralized stablecoins use on-chain assets as backing, held in smart contracts rather than banks. DAI, issued by MakerDAO (now Sky), is the canonical example. You deposit ETH or other approved assets, and you can borrow DAI up to some percentage of your collateral value — typically requiring 150% or more collateral relative to the borrowed amount.
The over-collateralization is designed to absorb price volatility. If ETH drops 30%, a 150% collateral ratio means the position is still fully backed. If collateral value falls below the minimum ratio, automated liquidation kicks in: the smart contract sells collateral to repay the debt and maintain system solvency. All of this happens on-chain and is publicly verifiable — no trust in an issuer required.
The constraint is reflexivity. In a severe market downturn, collateral values fall at the same time liquidation pressure spikes, and the liquidation selling can itself push prices down further. March 2020 tested this directly — a flash crash exposed gaps in DAI's liquidation engine, with some vaults cleared with zero bids due to network congestion, creating undercollateralized bad debt. The system survived but required protocol changes.
Algorithmic stablecoins attempt to maintain a peg without holding direct reserves. Supply expansion and contraction mechanics — often combined with a secondary token designed to absorb volatility — are meant to keep the price at target.
Terra's UST was the largest attempt at this model. The mechanism: when UST traded above $1, users could mint UST by burning LUNA, expanding supply and pressing price down. When UST traded below $1, users could burn UST to mint LUNA, contracting supply and supporting the price. The problem was that the incentive structure depended on confidence in the whole system remaining intact. A sufficiently large withdrawal of confidence broke the mechanism and triggered a reflexive death spiral. UST depegged in May 2022 and both UST and LUNA effectively went to zero in days.
That event has made "algorithmic stablecoin" roughly synonymous with failed design among market participants and regulators, at least in the pure unbacked form. Hybrid models with partial reserves still exist, but the fully algorithmic approach has lost credibility.
Where the Regulatory Pressure Is Focused
Fiat-backed stablecoin issuers are the primary regulatory target. The questions being asked: What exactly is in the reserves? How often is it verified? Who has custody of the assets? What happens if the issuer becomes insolvent?
US legislative frameworks around stablecoin regulation have centered on reserve quality requirements — specifically requiring that reserves consist of high-quality liquid assets like cash and short-term US government securities, not commercial paper or crypto assets. This would regulate which backing assets are permissible, not just whether reserves exist at all.
Banking relationships are a soft but real constraint. Stablecoin issuers holding fiat reserves need banking partners. When Silvergate and Signature Bank collapsed in March 2023, Circle disclosed that $3.3 billion of USDC reserves were held at Silicon Valley Bank, which also failed that weekend. USDC depegged briefly to around $0.87 before the FDIC backstop was confirmed. The event showed that risk in fiat-backed stablecoins can transmit through traditional banking channels, not just crypto-native mechanisms. A stablecoin can be fully backed on paper and still face a run if users don't trust the issuer's ability to access those assets.
What Would Confirm the Reserve System Is Strengthening
Formal reserve audits replacing point-in-time attestations, published quarterly or monthly. Regulatory frameworks mandating reserve composition disclosure with independent verification. Reserve asset migration toward T-bills and cash and away from money market funds with any credit exposure. Banking relationship diversification with explicit concentration disclosures.
What Would Break or Invalidate It
A major fiat-backed issuer losing banking access would effectively suspend redemptions — users couldn't recover dollars even if reserves technically existed somewhere. Large rapid redemption pressure on a primarily T-bill portfolio would require selling assets into markets, creating short-term gaps between token value and asset value.
For crypto-backed stablecoins: a synchronized crash across all approved collateral types faster than liquidation mechanisms can clear. Smart contract exploits targeting the oracle or liquidation infrastructure. Governance decisions that lower collateral requirements faster than market conditions can support.
For any reserve model: a jurisdiction making it illegal to operate, combined with banking partners exiting.
Timing
Now: Reserve composition for fiat-backed stablecoins matters actively. Regulatory frameworks are being finalized in the US and EU, and the reserve quality standards being set will determine which issuers remain viable under the new rules. Circle's pursuit of a bank charter and the status of the US stablecoin bill are live developments worth tracking.
Next: Mandatory audit standards rather than attestations are likely. The outcome of US stablecoin legislation will determine whether non-bank issuers can operate under a federal framework or face a patchwork of state regimes.
Later: Cross-chain reserve verification is an unsolved coordination problem. As stablecoins deploy across more networks, matching reserve assets to token supply across all chains requires infrastructure and governance that doesn't yet exist in standardized form.
Boundary Statement
This post describes how the three reserve models work mechanically and where the constraints live. It doesn't assess any specific stablecoin's current safety or recommend one reserve model over another. Reserve structures change — the USDT composition of 2019 differs significantly from 2026. Verify issuer disclosures and attestations directly for current status.
The mechanism described is stable. Whether any particular implementation of it holds under stress is a separate question.
