🪙 Payment stablecoins

What a payment stablecoin is, and what it is not

A stablecoin is a token that circulates on a distributed ledger and whose issuer promises a fixed value in legal tender, almost always one dollar. In payments, the definition that matters is legal before it is technical. A payment stablecoin is a liability of an identifiable issuer, redeemable at par and backed by segregated reserves. All three conditions must hold at once. If one is missing, the token is not a payment stablecoin. Many assets that market aggregators label “stablecoins” meet none of them, and a PSP cannot accept them under the same procedure.

A token’s market capitalization measures the amount outstanding, not the volume of payments settled in it. A token can be worth several billion dollars without a single commercial payment settling in it. BFUSD serves as internal margin collateral on a derivatives exchange, FDUSD as a trading pair, and USDD as an ecosystem reserve. An aggregator’s ranking lumps together assets with very different functions, so it says nothing about how a token is used. Operational due diligence looks at three things: who the debtor is, what backs the claim, and how long redemption takes.

• The rail runs 24/7, holidays included: the one property that neither cards nor wholesale transfers fully match, and the reason behind most real-world use cases.
• Transactions are irreversible: no native chargeback and no way to cancel. Everything a card provides at no extra cost (disputes, implicit escrow, guarantees) has to be rebuilt by contract.
• Compliance travels with the token: a stablecoin that is not compliant in the payer’s jurisdiction cannot be accepted, however liquid it is globally.
• The real cost is not on the chain: network fees are negligible, but getting into and out of fiat (the on-ramp and off-ramp) carries all of the cost, all of the delay, and all of the risk.
• Converting to fiat is a banking business: the weak link in a stablecoin setup is almost always the conversion provider’s partner bank, not the blockchain.

Issuance, reserves, and redemption: how the peg holds

The peg of a fiat-backed stablecoin rests on arbitrage in the primary market, not on an algorithm. That market connects the issuer with the institutional clients it has onboarded, known as minters. A minter wires dollars and receives newly minted tokens one for one. In reverse, it returns tokens and receives dollars, and the returned tokens are burned. As long as this loop works, anyone with access to the issuer’s window arbitrages away any price gap on the secondary market. Trouble at the issuer’s bank, a suspended redemption, or a bank holiday weekend is enough to jam it. The peg then starts to float.

Major issuers publish attestations, and almost never audit reports. An attestation is an agreed-upon procedures report signed by an accounting firm. It records, as of a given date, the amount in circulation and the composition of the reserves. The result is a verified snapshot, with no opinion on the financial statements or on internal controls. Circle publishes a monthly attestation signed by Deloitte & Touche LLP, plus a weekly disclosure and a daily line-by-line breakdown of the portfolio. Tether produces a quarterly attestation signed by BDO. How granular and how frequent these disclosures are is a criterion for choosing between issuers.

• Where are the deposits? A dollar stablecoin whose cash sits with a single commercial bank carries that bank’s risk. That is exactly what happened in March 2023.
• What maturity? 90-day Treasury bills can be sold in one trading day; physical gold, bitcoin, or secured loans cannot. The asset mix determines how fast redemptions get paid in a crisis.
• How large are the excess reserves? Assets in excess of liabilities are the first-loss cushion. At Tether, they stood at $4.11 billion against $183.64 billion in liabilities as of June 30, 2026, or about 2% (Tether, Q2 2026).
• What redemption fees, and what contractual deadline? MiCA prohibits redemption fees above the cost of execution. Outside the EU, this is negotiated and spelled out in the issuer’s terms and conditions.
• Who can freeze my tokens? The USDT, USDC, and PYUSD contracts include a freeze function. It protects against fraud and creates operational exposure: the issuer, or an authority’s order, can freeze a balance without notice.

USDT, USDC, PYUSD, EURC: four business models

The four benchmark payment tokens play distinct roles and rarely compete for the same flows. USDT serves as a substitute dollar in economies short of dollars. USDC serves as a settlement asset for regulated institutions. PYUSD extends a consumer e-money account onto the blockchain. EURC reflects a European regulatory and sovereignty agenda, still at a modest scale. A payment architecture that ignores these differences will run short of liquidity in the token it picks.

MiCA and the GENIUS Act: two regimes, two philosophies

Two laws now shape the global market for payment stablecoins. The EU’s MiCA regulation, (EU) 2023/1114, has applied to its Titles III and IV since June 30, 2024. Title V, which covers the licensing of crypto-asset service providers, has applied since December 30, 2024. The US GENIUS Act was signed into law on July 18, 2025, and does not apply yet. It takes effect on the earlier of two dates: 18 months after enactment (January 18, 2027) or 120 days after federal banking regulators publish final rules. The EU therefore has about a two-year head start. A transatlantic architecture designed today falls under a regime that is already binding in Europe and still incomplete in the US.

• Redemption at par (Article 49): holders have a claim on the issuer, redeemable at any time and at face value, with no fee above the cost of execution.
• Ban on interest (Article 50): neither the issuer nor the service provider may pay interest or any benefit tied to how long the token is held. A European stablecoin therefore cannot be marketed as an investment.
• Investment of funds received (Article 54): at least 30% of funds in deposits with credit institutions (60% for a token designated as significant), the rest in highly liquid, low-risk financial instruments denominated in the same currency as the token.
• Authorization: only a credit institution or an e-money institution authorized in the EU may issue an e-money token, on the basis of a white paper notified to the competent authority.
• The usage cap (Articles 22, 23, and 58(3)): the reporting obligations and issuance restriction that apply to asset-referenced tokens are extended to e-money tokens denominated in a currency that is not an official currency of a member state. This provision directly targets dollar stablecoins used for payments in Europe.

A federal trust charter is a license granted by the Office of the Comptroller of the Currency (OCC), the federal bank regulator. It puts the holder under federal supervision. US issuers have been racing for federal trust charters since the GENIUS Act passed. Circle National Trust received final OCC approval on July 10, 2026, becoming the first stablecoin issuer with a federal trust bank charter. Bridge, Stripe’s infrastructure subsidiary, received conditional approval in February 2026. Anchorage Digital Bank, N.A. has issued USA₮ for Tether’s US business since January 2026. For a PSP, this means the issuer it contracts with becomes a supervised institution, subject to capital requirements, continuity obligations, and constraints on winding down.

The rest of the map: Hong Kong, Singapore, Japan, the UAE, and the UK

Several jurisdictions outside the scope of MiCA and the GENIUS Act have adopted their own stablecoin regimes. The corridors where stablecoins actually get used run through Asia and the Gulf, and those jurisdictions have legislated, sometimes ahead of the EU. An issuer or provider opening a corridor is subject to four legal regimes at once: those of the payer, the payee, the token issuer, and the conversion provider.

Real-world uses: cross-border B2B, remittances, treasury, cards

Stablecoins get real use where the correspondent banking chain is slow, expensive, closed on weekends, or missing. Within a single currency area, a token offers no advantage over an instant payment or a card, and it has not replaced domestic payments anywhere. Four use cases are documented and measurable today; the rest are still announcements.

Real adoption is very uneven, and it tracks the weakness of the local currency more than financial sophistication. Chainalysis’s Geography of Cryptocurrency report (2025) puts on-chain value received at $2.36 trillion in Asia-Pacific (up 69% year over year), with growth of 63% in Latin America and 52% in sub-Saharan Africa ($205 billion). In these regions, stablecoins work less as a payments innovation than as a substitute for a dollar account, used to save, pay for imports, and pay wages. USDT is the dominant token there, and it is the first one a PSP operating in those markets will encounter.

On-ramps and off-ramps: where the token turns back into money

An on-ramp converts legal tender into tokens; an off-ramp does the reverse. The two bracket every stablecoin payment in which at least one leg settles in legal tender. Moving value on-chain is instant and nearly free. The on-ramp and off-ramp account for all of the cost, the delay, the compliance risk, and the counterparty risk. A stablecoin architecture is therefore only as good as its ramps and the soundness of their partner banks, whatever chain it runs on.

• Bank risk comes first: the conversion provider depends on one or two partner banks. A de-risking decision can shut down a corridor overnight, and the blockchain has nothing to do with it.
• FX is not free: collecting in dollars while paying expenses in euros brings market risk into the company. When collections pile up over several days, this is the biggest source of potential loss, far larger than network fees.
• Liquidity is asymmetric: converting US$100,000 of USDT into Argentine pesos and into euros does not cost the same or take the same time. Always read an off-ramp quote by corridor and by amount.
• The weekend is the real test: it is the only time the “24/7” claim either holds or falls apart. If the off-ramp pays out only on business days, the rail is instant all the way to the last mile, and slow on the last mile.
• Consolidation is under way: Stripe bought Bridge for about $1.1 billion (the deal closed in February 2025), and on March 17, 2026, Mastercard announced its acquisition of BVNK for up to $1.8 billion, including $300 million in earn-outs, with closing expected before year-end. Choosing an orchestrator today means choosing its future owner.

Depegs: what breaks, why, and what it costs

A depeg is a price gap on the secondary market between a token and its stated peg, caused by doubts about whether it can be redeemed at par. Only rarely does it signal an issuer default. There are three recurring causes. The first is impaired reserves, through bank or credit risk. The second is a closed or slow redemption channel, an operational risk. The third is a lack of backing from the start, a design risk. For a payments professional, depeg exposure is measured over the time window during which collected funds wait to be converted, not over the life of an investment position.

The gap seen in March 2023 comes down to the difference between the primary market and the secondary market. Circle’s direct clients kept their right to redeem at $1.00. Secondary holders, with no access to the issuer’s window, were selling at $0.87. The price gap therefore measured the value of access to redemption, not the value of the reserves. For a PSP, depeg exposure grows with the number of intermediaries between collection and the issuer.

• Automatic conversion on receipt: the simplest and most effective hedge. The token only passes through, and exposure drops to a few minutes.
• An exposure cap per issuer: treat each stablecoin like a bank counterparty, with an exposure limit and periodic review. Corporate treasury rules apply unchanged.
• A split across at least two tokens on critical corridors: a depeg is an idiosyncratic event, and diversification handles it better than reserve analysis.
• Continuous monitoring of the secondary price: a sustained gap of more than a few dozen basis points on deep markets is a signal to stop accepting the token, not a topic for debate.
• A token exit procedure written in advance: the BUSD precedent shows a regulator can halt a token in a day. Work out how long a switch to another asset takes before the event, not during it.

What a PSP must check before accepting stablecoins

Accepting a stablecoin adds four new items to the risk map: an issuer, a chain, a conversion provider, and a legal regime. It goes well beyond adding a payment method to the product catalog. Each of the four goes through the same process as onboarding a new partner bank, with a written file and periodic review.

• The issuer and its license: which legal entity issues the token, under what status, in which country? A token that is compliant in one jurisdiction is not necessarily compliant elsewhere. Check compliance in both the payer’s and the payee’s country.
• The redemption right: at par? At any time? With what fees and what contractual deadline? Above all: is it my right, or that of an intermediary several links up the chain?
• With multi-issuance, which entity redeems, and from which reserves? The issue raised by the European Systemic Risk Board in September 2025 is still unresolved.
• The reserves: composition, custodian, deposit banks, average maturity, excess reserves. Attestation or audit? How often, by which firm, and in how much detail?
• The freeze function: who can freeze a balance, on what grounds, and with what recourse? On April 23, 2026, Tether froze more than $344 million at two Tron addresses in coordination with OFAC and US authorities.
• The chain: time to finality, fee levels and volatility, history of incidents and congestion, and whether the contract is upgradable and who administers it.
• The conversion provider: license (CASP under MiCA in the EU, an equivalent license elsewhere), partner banks, corridor coverage, commitments on timing and price.
• Flow compliance: the travel rule (Regulation (EU) 2023/1113 since December 30, 2024), sanctions screening, handling of self-hosted addresses, and the procedure when funds come from a flagged address.
• Accounting and tax treatment: classification of the token, valuation method, FX differences, and supporting documents the statutory auditor will accept.
• The exit: how long does it take to stop accepting a token, switch to another, and clear the balances? Write it down before go-live.

chain            : ethereum            settlement network (each network has its own finality)
token_contract   : 0xA0b8...eB48        contract address: the ONLY proof of the asset received
                                        (a token with the same name at another address
                                         is a different asset: a classic trap)
tx_hash          : 0x7f3c...9ad1        unique transaction identifier
block_number     : 23481902             block height at inclusion
confirmations    : 64                   internal policy: credit the merchant at N confirmations
from_address     : 0x51d2...c07a        payer address (screen before crediting)
to_address       : 0x9ee4...41bf        deposit address dedicated to the merchant or invoice
amount_raw       : 125500000            amount in the contract's base units
decimals         : 6                    USDC and USDT: 6 decimals -> 125.50 USD
                                        (assuming 18 decimals throws the amount off
                                         by a factor of 10^12; never hard-code it)
reference        : INV-2026-004182      business reference: matched off-chain; the chain
                                        carries no standardized payment description
fx_rate_applied  : 0.9214 USD->EUR      rate applied by the off-ramp, timestamped
settlement_leg   : SEPA / D+1           fiat exit leg: this is what sets the real delay

The competitive position of stablecoins is still debated. In retail, they do not compete with cards. They compete with correspondent banking, which is why the card networks are investing in this infrastructure instead of fighting it. Visa already settles on nine blockchains, Mastercard bought its infrastructure, and Stripe bought its own before launching the chain to go with it. The next battleground is the guarantee layer: rebuilding, on top of an irreversible rail, the protections payments have offered for half a century. Those protections are disputes, proof of delivery, and intermediary liability. Issuing the token has become a commodity business, and value is moving up to these higher layers.