Visa spent decades as the toll road between cards and merchants. Its current pitch is something else: the layer that authorizes, secures and clears payments no matter what money settles them, including Visa stablecoins has no hand in issuing. The company’s leadership now talks openly about stablecoin and agentic payments as growth, which is a notable thing for a card network to say out loud about rails that bypass cards.
The move is visible in product. Visa stablecoins support settlement in USDC and EURC on Solana and Ethereum, so a transaction can run through Visa’s authorization and risk systems while the money moves as a stablecoin underneath. Through Intelligent Commerce, it is extending the same logic to AI agents, positioning itself as the party that verifies the agent, tokenizes the credential and stands behind the transaction, whatever rail carries the value.
From owning the rail to renting trust
This is a different business with a different moat. Visa’s old advantage was the network: the more merchants and cardholders on it, the more valuable it became, and the rails themselves were the asset. The new position is narrower and arguably more durable. Visa is selling trust services, fraud screening, identity, dispute resolution and settlement guarantees, and unbundling them from the card rail so they can sit on top of bank transfers, stablecoins or agent payments alike.
The logic is defensive and shrewd. If stablecoins and account-to-account rails erode card volume, a network that only earns when a card is swiped shrinks with them. A network that earns by underwriting trust on any rail keeps a seat at every transaction, even the ones that route around its cards. Visa is trying to make itself rail-agnostic before the rails it does not control get big enough to matter.
The part of the bet that can break
The vulnerability is that stablecoins were designed to need fewer intermediaries, not more. A USDC payment that settles on-chain in seconds, with finality and no chargeback, removes much of what Visa historically charged for. The buyer who wants irreversible, low-cost settlement is choosing a stablecoin precisely to avoid the dispute layer Visa is now selling as a service. For those flows, Visa’s trust bundle is a cost the transaction was structured to skip.
Which is why Visa’s real market is not crypto-native settlement at all. It is the large, cautious institutional and consumer flows that want a stablecoin’s speed but cannot accept its finality, the payments where someone still needs a number to call when a transfer goes wrong. There is genuine demand there, and Visa is better placed to serve it than any crypto-native firm, because it already runs the dispute and identity machinery at scale. The question is how big that middle is once both ends, pure card and pure stablecoin, pull customers toward themselves.
Scale is what makes the bet plausible. Visa already moves trillions a year and runs fraud, identity and dispute systems no crypto-native firm can match on day one, so unbundling those services onto new rails starts from a real advantage rather than a standing start. The harder part is what those services command in price once they are optional. On a card, the fee is baked into a rail the merchant cannot avoid. On a stablecoin payment, Visa’s trust layer becomes a line item a buyer can decline, and declining it is sometimes the entire reason the buyer chose the rail. Advantage in capability does not automatically convert to pricing power when the customer can route around you.
Visa is making the right move for a company that can see card rails plateauing. Becoming the settlement and trust layer across many rails is a more defensible position than defending card volume head-on. The open question is pricing power. As a toll road, Visa set the toll. As one trust provider among several on rails it does not own, it has to prove the service is worth paying for when the whole point of the new rails was to need it less. The repositioning is correct. The margins are the thing to watch.