An analysis of stablecoin payment cards argues that the common pitch around them — cut out the card networks, save on fees, and pass those savings to merchants — does not hold up well in domestic card payments. The central point is simple: removing Visa or Mastercard does not remove the biggest cost in the stack.
The article says crypto credit card spending topped $759 million in July across 9 million transactions, nearly 2.5 times the level a year earlier. More than 90% of that volume still ran on Visa.
Where card fees actually go
The piece breaks down the economics of a merchant accepting a card payment. On a $100 rewards credit card purchase, the merchant may pay about 2.2% in merchant discount fees, or $2.20 in total.
That $2.20 is not mostly captured by Visa. According to the article, the fee pool is split across three parties:
- About $1.75 goes to the issuing bank, the institution that provided the card to the consumer. This is interchange and accounts for 70% to 80% of the merchant’s total fee burden.
- The merchant-side payment processor, or acquirer, takes about $0.30 to $0.70 in service fees.
- Visa or Mastercard collect only about $0.13 to $0.18 in assessment fees, roughly 7% to 9% of the merchant’s total cost.
On that view, taking Visa out of the chain would only remove the smallest slice of the fee stack.
Visa is not the lender in the system
The article says a common misunderstanding, including inside crypto, is that Visa takes the largest share of swipe fees. It does not. Visa does not extend credit to consumers, does not bear credit losses, chargebacks, or fraud losses, and does not directly move funds.
Its role is to operate the information network that turns on when a card is used, passing authorization messages between the merchant terminal and the issuing bank while setting the rules for the system. The issuer carries the heavier load: it extends credit, absorbs bad-debt risk, funds the gap between purchase and repayment, and uses interchange revenue to subsidize card rewards.
The article cites Visa’s 2025 operating figures to make the point. Visa processed $14.2 trillion in payment volume across 257.5 billion transactions, generated $40 billion in net revenue, and posted a net margin close to 50%. On average, it earned about $0.13 per transaction. The argument is that Visa’s economics come from massive scale and near-zero marginal cost, not from charging a large fee on each swipe.
Why stablecoin cards sit in a different model
The analysis then turns to stablecoin cards and says nearly all of them are debit products. Funds are already sitting in the user’s wallet in the form of USDC or USDT before the purchase happens. There is no credit period and no revolving interest income.
That creates a very different revenue model from credit cards. The article points to the Durbin Amendment, passed by the U.S. Congress in 2010, which caps debit interchange for banks with more than $10 billion in assets at $0.21 plus 0.05% per transaction. Most stablecoin card programs work with smaller partner banks or digital banks below that $10 billion threshold, which leaves them outside the cap. The piece describes this as a standard fintech partner-bank setup.
Even then, average interchange on exempt dual-message debit networks is only about $0.62 per transaction. The comparison is stark: a $100 rewards credit card purchase can generate $2.20 in total revenue, while a stablecoin debit card, even under the higher exempt rate, brings in only $0.62.
From that $0.62, the program still has to pay network fees, processor fees, partner bank costs, fraud losses, and operating expenses. Only what remains could be used to offer merchants any economic benefit. The article’s conclusion is that the margin available to run a stablecoin card business off debit interchange is extremely tight.
What stablecoins can actually replace
The article argues that the real value of stablecoins in payments is not in shaving off the network fee, but in replacing more back-end parts of the payment stack. It revisits an earlier framework that split cross-border payments into seven cost layers: acceptance, orchestration, licensing and compliance, custody, foreign exchange, issuing, and clearing and settlement.
Viewed through that lens, little changes for the merchant in a domestic card transaction. Acceptance looks the same. The merchant still needs a terminal, the customer still taps or swipes a card, and the merchant still pays a discount rate to the acquirer. The merchant may not even know whether the funding source behind the card is USDC, because the transaction appears no different from a standard Visa payment on the statement.
Orchestration also stays in place through Visa or Mastercard, and issuing still depends on a partner bank and a network BIN. The article says fintech firms were already using this structure before stablecoins existed, so there is no fundamental break here.
The real change comes in the back end, especially clearing and settlement. In the traditional model, settlement between issuers and card networks typically runs on a T+2 basis, with weekends and batch processing adding friction. To address that, providers such as Rain support end-of-day stablecoin settlement with Visa, while Mastercard has begun accepting stablecoins including USDC, PYUSD, and RLUSD for intraday settlement.
That compresses the old T+2 cycle toward near-real-time settlement and frees up working capital that issuers would otherwise have tied up.
Faster settlement does not mean lower merchant fees
Still, the article says the benefit from faster settlement accrues to issuers and card operators, not to merchants or consumers. Merchant discount rates do not fall just because settlement speeds up, and the checkout experience does not materially change for the customer.
In that framing, T+0 stablecoin settlement mainly means the card operator no longer has to pre-fund a two-day cycle. It is a treasury and working-capital improvement for the issuer side. At scale, that can be meaningful. It does not automatically create a better deal for merchants or cardholders.
The piece also says Visa has no reason to cut its fees because settlement float was never Visa’s cost to begin with. That burden sat with the issuer. If stablecoin settlement lowers issuer funding costs, Visa’s own cost base is unchanged, so merchant pricing would not be expected to move.
Visa and Mastercard are integrating stablecoins, not being displaced by them
The article notes that neither Visa nor Mastercard has resisted stablecoin settlement. Instead, both have been bringing it into their own networks. Visa has supported settlement in USDC on Ethereum and Solana since 2021, and the article says its annualized settlement volume now stands at $7 billion. Mastercard, for its part, acquired BVNK months ago to expand its stablecoin infrastructure.
It quotes the card networks as saying: 「Stablecoins will not disrupt the existing payments order; they will reinforce it.」
The article’s reading is that the networks are not being replaced. They are absorbing stablecoins as an upgrade to their settlement layer. Every stablecoin card that runs on Visa may claim to challenge Visa, but it still adds transaction volume to the network and continues to pay fees into it.
Cross-border logic does not map cleanly onto domestic spending
The article closes by drawing a line between two use cases. Stablecoins can reduce costs in cross-border payments by cutting through multiple correspondent layers and easing foreign exchange frictions. That argument, it says, is valid.
Domestic card spending is different. In that setting, the largest costs do not sit in settlement infrastructure. They sit in issuing, interchange, and merchant acquiring. The article argues that taking a stablecoin thesis that works in cross-border payments and applying it directly to domestic consumer card spending does not match the actual fee structure.
Its bottom line is that stablecoins can realistically replace clearing and settlement, and they can help with cross-border foreign exchange. What they do not replace is the biggest cost center in domestic card transactions.
