Margin calculations for perpetual futures are usually shown in dollars, so stablecoin collateral can feel like cash sitting in an account. It is not quite that. A stablecoin is a token whose value depends on an issuer, its reserves, a redemption route, the token contract’s administrative controls and the market price at which other people will trade it. The venue then decides how much of that token counts toward your margin. Each layer can change the headroom between your account value and a liquidation threshold without any change in the positions you hold.
Collateral is a claim, not cash
A stablecoin’s target price is a design goal, not a guarantee. Reserve-backed tokens depend on what the issuer holds and on the terms under which holders, or a narrower set of approved customers, can redeem. The guide to what a stablecoin actually is explains why a peg is not a promise. Reserve reports also have limits: an attestation describes a particular date and scope, and the explainer on reading reserve attestations and audits shows what such a disclosure does and does not establish.
For a trader, the practical point is that “$1,000 of collateral” is really a quantity of a specific token, on a specific network, held under a specific venue’s custody and rules.
Five layers that can move the number
| Layer | What can change | Where to check |
|---|---|---|
| Issuer and reserves | Confidence in backing, reserve composition | Issuer disclosures and their dates |
| Redemption | Who can redeem, minimums, fees, delays | Current redemption terms |
| Administrative control | Freezes, pauses, blocked addresses | Token contract powers and issuer policy |
| Market price | Trading below target during stress | Several liquid markets, not one quote |
| Venue valuation | How the venue counts the token as margin | The venue’s current margin documentation |
A depeg is the market-price layer becoming visible. The explainer on why stablecoins lose their peg separates price, liquidity, collateral and redemption stress, which often arrive together. Administrative control is separate: some tokens let an issuer freeze or block transfers, so holding the keys does not guarantee every transfer will succeed. See stablecoin freezes and issuer control for how those powers coexist with ordinary blockchain transfers.
How a venue counts the collateral
The margin rule decides whether these layers affect an open position. Hyperliquid’s margining documentation offers one documented example. Cross margin is the default and shares collateral across cross positions; isolated margin constrains collateral to one position. Cross positions are liquidated when account value, including unrealized profit and loss, falls below the maintenance margin times total open notional, and maintenance margin is set to half the initial margin at maximum leverage. The page refers to depositing USDC to add margin to cross positions and notes that some markets can use different collateral. It does not describe a valuation haircut for the collateral itself, so check the venue’s current documentation for how each collateral asset is counted rather than assuming any rule.
Hyperliquid’s liquidation documentation says liquidations use the mark price, that the first step sends market orders to the book, and that collateral remaining after that process stays with the trader. If equity falls further and book liquidation fails, a backstop process takes over and the maintenance margin is not returned. These are one venue’s rules on the date checked; other venues can value, haircut or restrict stablecoin collateral differently. The companion guides on cross versus isolated margin and how crypto liquidation works cover the mechanics in more detail.
A hypothetical example of shrinking headroom
This example uses invented rules for a fictional venue. It is not Hyperliquid’s rule or any other venue’s.
A fictional cross account holds 1,000 units of a stablecoin and two perpetual positions of $2,000 notional each, the same setup as the ledger diagram above. Assume a flat 5% maintenance requirement on $4,000 of open notional, or $200. The fictional venue values collateral at its market price and, when the token trades below $0.98, applies a 10% valuation haircut.
| Hypothetical state | Collateral counted | Position P&L | Account value | Headroom over $200 |
|---|---|---|---|---|
| Token at $1.00 | $1,000 | $0 | $1,000 | $800 |
| Token at $0.96 | $960 | $0 | $960 | $760 |
| $0.96 with 10% haircut | $864 | $0 | $864 | $664 |
| Same, plus a $40 mark loss | $864 | −$40 | $824 | $624 |
Nothing about the positions changed in the first three rows, yet headroom fell by $136, about 17%. Add the $40 mark loss from the diagram and headroom is $624, compared with $760 for the same loss with the token at $1.00 and no haircut. Stress in a stablecoin often coincides with fast markets, so the collateral and the positions can deteriorate together. The arithmetic ignores fees, funding, tiered maintenance and the venue’s own mark-price rules, and it is a teaching illustration, not a forecast.
The reverse case matters too. If a venue keeps valuing a depegged token at its target, margin headroom may look unchanged while the exit value is not: withdrawing and selling or redeeming the token can still realize the discount.
Freeze and access risk
A depeg changes what collateral is worth. A freeze or pause can change whether you can move it. An issuer action against an address, a contract pause, or a venue’s own deposit and withdrawal controls can each delay a top-up or a withdrawal, which matters most when margin is tight. Holding collateral in more than one token can reduce dependence on a single issuer, but it adds conversion costs and its own rules. The framework for comparing USDT and USDC shows how to compare issuer disclosures, networks and redemption terms without treating either as a recommendation.
A collateral checklist before sizing a position
Record the exact token and network, its issuer and current reserve disclosure date, redemption terms that apply to you, known administrative powers, the venue’s collateral valuation and any haircut, the margin mode, the maintenance rule, and the withdrawal path. Then size the position with collateral stress included, not only price stress. The guide to leverage versus position size shows why exposure, not the leverage label, drives the outcome.
SuperQuant’s research engine uses an educational isolated-margin model with a flat maintenance proxy and synthetic hourly data. It does not model collateral depegs or valuation haircuts, does not place live orders and does not verify live execution.
Keep learning
The stablecoins topic hub gathers the issuer, reserve and freeze guides in one place, and collateral and liquidation in crypto borrowing explains how a position can close without your approval. For the perpetual side, continue with the perpetual futures primer. This is educational research context, not investment advice.