Stripe, Visa and Bridge split stablecoin checkout into two merchant fee stacks
Stablecoin cards still preserve much of the card-fee stack. Direct stablecoin checkout is where fee compression, chargeback changes, cross-border savings, incumbent M&A, network-branded digital dollars and agentic commerce start to matter for merchant underwriting.
a16z crypto counted $759 million of crypto-card spend in July, with stablecoins doing most of the work and the merchant seeing an ordinary card payment. That is the current state. Stablecoins are growing at the back of the payment chain, not replacing the checkout habit at the front of it. For a merchant statement, the difference is not philosophical. A stablecoin-funded card can still arrive through Visa or Mastercard, with interchange, network assessments, acquirer markup, gateway fees, risk controls and chargeback operations attached. A direct stablecoin checkout can bypass much of that stack, but only when the merchant accepts the token or a processor converts it outside the card rails. The operator read The merchant does not save money because the consumer funds a card with USDC. The merchant saves money when the card network, interchange layer and legacy acquirer economics are no longer the route. That is why the same stablecoin dollar can be expensive in one checkout flow and cheap in another. Scenario A is card-front, blockchain-back. The customer taps a card or uses a normal wallet button. The issuer, wallet or card program may use stablecoins for funding or settlement, but the merchant processor still sees a card transaction. A Main Street merchant can still pay a blended 2.5% to 3.5% because the visible acceptance product is the same one processors already price. Visa is already proving that model. Its U.S. USDC settlement launch lets issuer and acquirer partners settle with Visa in USDC, with seven-day funds movement and weekend treasury advantages, while consumers and merchants can keep the card front end. That improves back-office treasury before it reprices the merchant discount rate. Scenario B is direct stablecoin checkout. Stripe stablecoin payments can let a merchant accept USDC directly and pay a processor-led stablecoin fee instead of the full card stack. Stripe has positioned that fee around 1.5% in the Shopify rollout, which means the real comparison is not crypto versus cards. It is card-rail stablecoin versus direct-rail stablecoin. The hidden savings sit beyond the sticker fee. A direct stablecoin payment cannot be reversed through the unilateral card-chargeback path, so merchants can reduce one of the most expensive fraud and dispute buckets. That does not make fraud work disappear: merchants still need refunds, customer-support rules, wallet-screening, sanctions controls and delivery evidence. It does change who can pull funds back after settlement. The second savings pocket is cross-border bloat. Card acceptance, FX spread, correspondent banking delays and payout timing all stack up when a merchant sells across borders or pays overseas suppliers. Stablecoin rails can compress those costs and remove weekend float because settlement can run intraday, every day. For a cash-tight SMB, faster usable funds can matter more than the headline processing rate. Three shifts that follow from the fee split First, the incumbents counter-attack by buying and issuing infrastructure. Stripe completed its Bridge acquisition after agreeing to a $1.1 billion deal. Mastercard then agreed to buy BVNK for up to $1.8 billion, according to Reuters coverage carried by Investing.com . These are not side bets. They are attempts to own compliance, wallet connectivity, minting relationships, off-ramp coverage and enterprise distribution before Circle, Tether or a new payment chain captures the transaction layer. The next move is network-branded money. Visa and Mastercard do not need to start by putting a speculative token in consumers' hands. They can embed regulated digital dollars, sponsored issuer tokens or consortium stablecoins inside their existing rails, then expose the stablecoin leg only to banks, acquirers, wallets and large merchants. A plausible three-year path is not every cashier seeing a new coin. It is a merchant statement where a Visa or Mastercard digital-dollar settlement option sits behind the same acceptance mark. There is also a yield-arbitrage lane. The GENIUS Act framework restricts payment-stablecoin issuers from paying interest to holders, but the consumer incentive can move to a separate rewards account, cash-management product or third-party yield product. The card app starts looking like a high-yield balance that pays at checkout through a stablecoin-equivalent settlement leg. That preserves the legal separation while competing with bank deposits for idle customer cash. Second, the fee war moves from percentages to software pricing. If settlement clears on Solana, Base or payment-specialized chains such as Stripe-backed Tempo-style infrastructure, the marginal cost of moving the payment can fall toward fractions of a cent. Processors will not voluntarily give up margin, but the basis for margin changes. The toll moves from moving money to fraud tooling, treasury management, instant conversion, reconciliation, tax handling and accounting integrations. That means the old 2.5% to 3.5% merchant card fee is not likely to vanish in one step. It compresses first for high-volume platforms, cross-border merchants, B2B invoices, supplier payouts and categories with heavy chargeback losses. Later it becomes harder to defend percentage pricing for ordinary money movement when a competing processor can offer subscription pricing, micro-basis-point routing or bundled treasury software. Third, direct stablecoin growth comes from agentic commerce before it comes from human checkout. Consumers change payment habits slowly. Software agents do not care about card rewards, plastic cards or familiar wallet buttons. They care about programmable authorization, instant settlement, low unit cost, spending limits, identity rules and machine-readable receipts. That is where machines paying machines become the distribution wedge: ad agents buying impressions, procurement agents paying approved vendors, logistics systems releasing carrier payments, AI operators paying API vendors, marketplaces settling creators, and embedded-finance platforms routing thousands of small payouts. A card network is overbuilt for those payments. A stablecoin rail with policy controls is closer to the shape of the transaction. The underwriting file has to identify the rail For MCA and SMB credit, the main mistake is treating all stablecoin revenue as one signal. Card-front stablecoin spend still produces ordinary processor evidence. Direct stablecoin checkout can land in a Stripe balance, bank payout, custodial wallet or external wallet before it touches the debit account. Those flows can create real revenue with weaker bank-statement visibility. The processor statement also stops being enough when payment processors become treasury routers. Underwriters will need rail-level sales volume, refund and dispute policy, wallet settlement reports, conversion timing, reserves, chargeback-equivalent claims and which account receives final usable cash. A merchant with lower processing costs can still be harder to debit if the funds land outside the monitored account. Opinion The cheaper rail can still make the file noisier The stablecoin winner for merchants is not the token with the loudest brand. It is the route that cuts interchange, reduces dispute drag, settles usable funds faster and still leaves enough evidence for capital partners to trust the receivable. The funder that only checks bank deposits will miss both sides of the shift: real margin improvement and real collection risk. Plutus covers this beat daily at plutusai.net .