A depeg hurts a payment flow only while you hold the stablecoin. What USDC's 2023 depeg and UST's collapse teach, and the limits and triggers to set.
A stablecoin depeg only hurts a payment company while it holds the stablecoin at a floating price. A payout locked by a quote and converted within minutes carries almost no depeg risk. A treasury balance, a pre-funded float, or unconverted customer receivables carry all of it. Set limits on those holdings, not on payments.
This article is general information, not legal or investment advice.
Key takeaways
Definition. A stablecoin depeg is a period when a stablecoin trades on the open market away from its target value, usually one US dollar.
A payment flow holds a stablecoin for minutes, while a treasury holds it for weeks, so the same price move costs them very different amounts.
Think of exposure as balance multiplied by time held. A payout that locks a rate, pulls the stablecoin, and pays out in local currency is exposed only between the lock and the conversion. A company that keeps a stablecoin float to fund payouts, or lets merchant receivables sit in stablecoins until month-end, is exposed for the whole holding period.
Here is where stablecoin exposure actually sits in a typical payments business:
| Where the stablecoin sits | Typical holding time | Depeg exposure | Who carries it |
|---|---|---|---|
| Quote window before a payout executes | Minutes | Low, if the rate is locked | Usually the provider, inside the quote |
| On-chain transfer to the payout provider | Seconds to minutes | Low | Sender until the transfer confirms |
| Pre-funded balance at a provider | Days to weeks | High | You |
| Customer receivables held in stablecoins | Days to a month | High | You or your customer, by contract |
| Treasury reserve held in stablecoins | Weeks to months | Highest | You |
The design question is simple: which rows can you shrink? The no pre-funding model removes the third row entirely.
Two events show the two main failure types: a fiat-backed stablecoin with a reserve problem, and an algorithmic stablecoin with no real reserve.
USDC, March 2023 (reserve concentration). On March 10, Silicon Valley Bank failed and Circle disclosed that $3.3 billion of USDC reserves, about 8 percent, were held there. The Federal Reserve's December 2025 note on the episode records:
UST, May 2022 (algorithmic design). TerraUSD tried to hold its peg through an exchange mechanism with a sister token, LUNA, not through cash reserves. The SEC's complaint states that in May 2022 UST depegged and its price, along with its sister tokens, "plummeted to close to zero." It never recovered.
The lesson for payments teams: a reserve-backed stablecoin can trade at a discount for a weekend and recover. An unbacked one can go to zero. Payment flows should only touch the first kind. USDC vs USDT for payments compares the two largest issuers' track records.
A fiat-backed stablecoin's redemption depends on the banking system, which closes on weekends, while blockchains and exchanges keep trading.
That gap is the mechanism. During the week, an arbitrageur who buys USDC at 97 cents can redeem it with the issuer at one dollar, which pulls the price back up. When redemption pauses, that pull disappears, and the secondary market price floats freely until banks reopen. The SVB failure landed on a Friday, and the trough came on Saturday.
For a payment company, this means the riskiest time to hold a stablecoin balance is Friday evening through Monday morning, which is also when local bank rails are thinnest. Plan for both at once.
A depeg spreads through backing and liquidity links: a stablecoin that holds another stablecoin as collateral, or trades mainly against it, inherits its stress.
In March 2023, DAI fell because a large share of its backing was USDC held in stability modules. USDP fell too. Meanwhile USDT rose above one dollar because traders sold USDC for it.
So diversifying across two stablecoins helps only if they don't share reserves, banks, or collateral. Before you count a second token as a hedge, check what backs it.
Set limits on holdings, not on payment volume, and decide the rules before an event.
Depeg playbook (the failure modes to plan for)
The right controls depend on how long each use case holds stablecoins.
| Use case | Typical holding time | Main depeg exposure | Controls that matter most |
|---|---|---|---|
| Payroll to Latin America | Minutes per run | Funding balance before the run | Fund just in time, locked quotes, token fallback |
| Merchant collections in the US | Hours to days | Receivables held before conversion | Holding-time limit, auto-conversion, customer terms |
| Treasury holding | Weeks to months | Full balance | Concentration cap, price triggers, issuer redemption access |
| Remittance | Minutes | Quote window | Locked quotes, requote discipline |
| B2B supplier payments | Minutes to hours | Pre-funded float | No pre-funding, concentration cap |
Read the table from the right: the further a use case sits from "minutes," the more of the playbook it needs.
Regulation reduces the reserve-quality risk behind past depegs, but it does not stop market prices from moving during a panic.
The GENIUS Act requires permitted US issuers to hold 1:1 reserves in cash and short-term Treasuries and to publish monthly reserve reports. Its core provisions take effect on January 18, 2027. The GENIUS Act guide has the full timeline. In the EU, MiCA already requires e-money tokens to be fully reserved and redeemable at par.
Better reserves make a 2023-style bank concentration less likely. They don't make a weekend gap between redemption and trading go away. Confirm how any regulation applies to your business with qualified counsel.
BlindPay converts USDC and USDT into local currency over Pix, SPEI, ACH, RTP, SEPA, and SWIFT (POBO/COBO), and payouts execute against a quote that locks the rate and fee for 5 minutes. There is no pre-funded balance to hold, so the exposure window in a payout is the minutes between the quote and the execution.
Token support varies by chain: USDT payouts run on Ethereum, Polygon, Solana, Tron, and Tempo, while Base, Arbitrum, Arc, and Stellar carry USDC only. The supported chains reference has the full matrix, which is what you need to plan a token fallback. For the wider map of who holds what in each kind of provider, see types of stablecoin APIs.
List every place your business holds a stablecoin for longer than a few minutes, and put a size and a time limit on each one. Then write the weekend version of the playbook: who decides, which token you fall back to, and which payouts keep running. Do it this week, not during the next event.
Stablecoin payments are as safe as the issuer, the network, the provider, and your own controls. The seven risks to check, with real incidents and fixes.
Five stablecoin APIs compared for cross-border payments: primary use case, pre-funding requirement, payout regions, and developer experience, plus how to choose by buyer scenario.
Ten stablecoin APIs compared for 2026: BlindPay, Circle, Bridge, BVNK, Fireblocks, Crossmint, Zero Hash, Conduit, Sphere, and Borderless, across rails, custody, pricing, and compliance.