How stablecoin cards work: from wallet funding to merchant settlement

Scale is real; maturity is not

Visa reported more than 130 stablecoin-linked card programs across more than 50 countries as of early 2026. It also reported approximately $5.2 billion of 2025 volume, up 319% year over year. The denominator is the more useful fact: Visa says that volume represented about 0.04% of its $14.2 trillion total global volume.

The evidence supports a repeatable launch model and rapid early growth. It does not support the claim that stablecoin cards are already a mainstream payment rail. A serious market map should preserve both observations instead of turning a small but growing category into an inevitability story.

Sources: [1] Visa: Stablecoin-linked cards and money movement

The visible card is the last layer

A typical program combines a card network, a licensed issuer or BIN sponsor, a program manager, a processor, identity and fraud systems, a wallet or custodian, a conversion venue and one or more settlement accounts. The consumer may know only the brand printed on the card. The program's country coverage and continuity depend on every contracted layer.

The funding model is equally important. Some programs reserve or convert a stablecoin balance at authorization; others maintain fiat program liquidity and rebalance later. Custodial products can control the conversion directly, while self-custodial products require permissions or smart-contract logic to make funds available at the moment of purchase.

Sources: [1] Visa: Stablecoin-linked cards and money movement · [2] Mastercard: Crypto Card Program

A card payment has more than one completion point

A wallet debit, an issuer authorization and a merchant payout answer different questions. Visa describes both conversion to fiat before network settlement and issuer settlement using supported stablecoins. In neither model does the cardholder's asset choice establish that the merchant receives that same token.

Refunds add another leg. Gnosis Pay documents a separate merchant-initiated credit through the card network. Our conclusion is operational: keep the purchase and refund references linked, rather than treating a successful onchain payment as the end of customer support.

Sources: [1] Visa: Stablecoin-linked cards and money movement · [4] Gnosis Pay: Transaction refunds

Dependency risk is product risk

A card may stop onboarding in one country without the global brand disappearing. A sponsor bank can change risk appetite; a wallet can lose support for an asset; a liquidity provider can widen spreads; a network tokenization integration can fail while the physical card remains usable. These are not back-office details when they change whether a customer can spend.

Network programs increasingly package partner introductions and settlement options to reduce launch effort. That improves access to infrastructure but does not collapse accountability into a single vendor. The product owner still needs contractual service levels, incident escalation, safeguarding logic and an exit plan for critical dependencies.

What a useful card database must reveal

A marketing page can answer whether a provider says it supports Apple Pay or USDC. It rarely proves fee totals, issuer identity, ATM conditions, residency restrictions or the exact entity holding customer assets. Those fields need a source, observation date and confidence status at the level of the specific program.

Comparison should prioritize availability, legal issuer, custody, stablecoin funding path, total conversion and FX cost, cash access, dispute handling and closure rights. Cashback belongs lower in the hierarchy because a reward rate cannot compensate for an unavailable or structurally opaque program.