COMMENTARY: Why The Hybrid Model Will Win the Programmable Payments Race

Stablecoins excel at moving value instantly. They fall short on many other things: identity verification, dispute resolution, global acceptance—the very infrastructure card networks spent decades building.

Visa and Mastercard are now layering stablecoin settlement underneath their existing frameworks, keeping the checkout experience and merchant protections intact while moving actual funds on-chain. The programmability comes from blockchain. The reliability comes from proven rails.

During my time building reporting systems for small merchants  I watched sellers spend 20 to 50 hours per month trying to match what they earned with what landed in their bank accounts. Sales looked simple: customer pays, money moves. But behind the scenes there were holds, fees, refunds, and batching delays. The deposit rarely matched the register.

We built better statuses, labels, and timeline views, but we always knew we were treating symptoms. The real problem was upstream: traditional rails don’t settle instantly, so the gap between sale and deposit was baked into the system. 

In agentic commerce, where AI agents make purchases autonomously, this reconciliation problem compounds. Multiple agents transact across multiple merchants with various processors. Trust becomes nearly impossible to establish through reporting alone.

Programmability offers a structural fix. When an AI assistant purchases something, the merchant confirms delivery, and a stablecoin payment is executed automatically upon that confirmation. The rules are coded into the payment itself, visible in real time. Reconciliation stops being a separate problem and becomes part of the transaction.

Building entirely new rails carries substantial risk. Smart-contract bugs can freeze or lose funds, with no undo button. Backing assets can de-peg, causing all dependent contracts to miscalculate settlement values. Most event triggers occur off-chain, reported via APIs or oracles; incorrect signals result in incorrect payments. Dispute resolution, chargebacks, and AML compliance remain underdeveloped on pure blockchain rails. Public chains expose every transaction, potentially revealing business flows and purchase patterns to competitors.

Meanwhile, card rails handle billions of transactions daily. They work. Stablecoin rails have not achieved that scale or that trust. The pragmatic path forward blends both.

When you pay with a Visa or Mastercard, that does far more than just move money. Through issuer banks, the networks verify identities, detect fraud, manage compliance, and ensure merchants get paid when something goes wrong. That infrastructure took decades to build. Rather than compete with it, blockchain-based settlement can extend it.

Mastercard’s Multi-Token Network, launched in beta in 2023, settles transactions in regulated stablecoins such as USDC, PYUSD, and EURC while maintaining Mastercard’s rules and protections. Same checkout experience. Same dispute resolution. The final transfer of funds between institutions happens on-chain. In May 2024, Mastercard completed a live pilot with Standard Chartered using tokenized deposits for carbon-credit transactions within the Hong Kong Monetary Authority’s sandbox.

Visa is pursuing a similar model. Through its stablecoin settlement expansion, the company uses USDC on Solana and Ethereum to settle payouts between issuers and acquirers. Visa has already moved millions in USDC through live pilots with Worldpay and Nuvei. Merchants can now choose stablecoin settlement over traditional fiat for the card payments they accept.

Limits exist. Traditional rails were never designed for event-driven logic or conditional execution. In this bridge model, the rule does not travel with the money; event listening and payment capture remain off-chain. Over time, the hybrid could evolve toward something closer to full agentic commerce, where delivery confirmation or smart-contract triggers automatically initiate stablecoin settlements. That evolution will take years.

For now, programmability works best at the settlement layer. It boosts efficiency without breaking the protections that make card payments trusted worldwide. As AI agents grow more sophisticated, they will demand payments that carry their own context. Networks positioned to deliver programmability without abandoning proven infrastructure have the advantage.

—Divyarani Raghupatruni is senior director of product, data, and orchestration at Alacriti.