The most common analytical error in stablecoin commentary is pricing a card transaction as if the merchant fee bought a money movement. It does not. It buys a bundle, and until the bundle is priced honestly, every "cards cost 2 to 3 percent, stablecoins cost cents" comparison is comparing a package holiday to a bus ticket.
What the four-party model actually sells
In the four-party model, the cardholder's bank (the issuer) and the merchant's bank (the acquirer) transact under rules the network writes and enforces. When a consumer pays 100 dollars at a merchant, the acquirer passes the issuer an interchange fee, the network charges both sides scheme fees, and the merchant receives the remainder. Interchange is the large component, and it is not network revenue; it flows to the issuer and funds things the merchant is, in effect, buying on the cardholder's behalf.
- Credit and float. On credit cards, the issuer finances the purchase interest-free until the statement date and takes the underwriting risk. A stablecoin transfer is prepaid money; it extends nobody credit.
- Fraud liability shifting. Under network rules, a fraudulent card-present transaction is generally the issuer's loss, and a disputed card-not-present transaction can be charged back to the merchant under defined rules with defined evidence standards. The consumer is largely made whole either way. An onchain transfer is final; the fraud loss sits wherever it lands.
- Dispute rights. Chargeback rules give the cardholder a non-judicial process for goods not delivered or not as described. Settlement finality, the property tokenised rails are proudest of, is precisely the absence of this.
- Acceptance. A card works at tens of millions of merchants because 2-sided network effects were built over roughly 6 decades. Acceptance is an asset with a replacement cost, not a default state.
What the numbers look like
Where regulators have intervened, they capped interchange, which is a decent indication of where they think the economics concentrate. The EU's Interchange Fee Regulation caps consumer card interchange at 0.2% for debit and 0.3% for credit (Regulation 2015/751). US credit interchange is uncapped and typically runs around 1.7% to 2.2% depending on card type and merchant category, per commonly cited industry compilations of the networks' published schedules. Scheme fees, the networks' own take, are smaller; worked industry examples put them in the region of 0.15% to 0.2% of transaction value, and the UK Payment Systems Regulator's market review of March 2025, "Market review of card scheme and processing fees", found Mastercard and Visa scheme and processing fees to acquirers had risen at least 25% since 2017 without effective competitive constraint. Singapore and Hong Kong publish no regulated interchange schedule; typical all-in card processing costs cited to merchants there run roughly 1% to 4%, on industry guides rather than a regulator series.
Two things follow. First, the network's own margin is a fraction of the merchant fee; most of the cost the merchant pays funds the issuer's bundle. Displacing the network while leaving issuers and their economics in place saves the merchant much less than the headline suggests. Second, where interchange is already capped near zero, as in EU debit, the displacement prize at point of sale is small, and what remains is the acceptance and dispute infrastructure that a bare transfer does not replicate.
The honest displacement case for cards
At point of sale, the case is narrow. A consumer paying with a stablecoin wallet gives up credit, gives up dispute rights, and gains nothing they can feel, since domestic instant payments in most of Asia are already free to them. Merchant savings are real but bounded by the bundle arithmetic above, and merchants have learned from account-to-account (A2A) schemes that cheaper acceptance does not by itself move consumers.
The case strengthens exactly where the bundle stops earning its keep. A corporate paying an overseas supplier does not want credit from the payment instrument, cannot charge anything back in practice, and needs no acceptance network for a known counterparty. In those flows, detailed in Parts 3 and 5, the card bundle is dead weight and the correspondent stack is the relevant, and weaker, incumbent. That asymmetry, bundle-lite flows versus bundle-heavy flows, is the first and cleanest cut on the displacement map in Part 1.
One caution the other way. The bundle is not immutable; it is funded by interchange, and interchange is a regulatory variable. Every jurisdiction that has capped it has compressed the rewards-and-credit subsidy that keeps consumers loyal to cards. A world of spreading interchange regulation is a world where the bundle thins from the inside, which would do more for account-to-account and wallet rails at point of sale than any blockchain property will.