Will agents adopt cards before stablecoins for payments?

Simon Taylor, writing on Fintech Brainfood, argues against the framing that cards and stablecoins are competitors for agent payments: "Not cards or stablecoins. Cards and stablecoins." He proposes instead that the two rails will be adopted in sequence as agent workflows grow more complex: "First, agents will use virtual cards... Then, agents will use cards that settle via stablecoins... Finally, wallets become a type of financial account for a new type of company." The question, he writes, "isn't which one wins. It's when each is the right tool for the specific job and at the right time."

Taylor's mechanism rests on a distinction between authorization and settlement: "Cards authorize the movement of money. Stablecoins move money." Virtual cards win the first stage because their controls are already mature — "Single-use. Budget caps. Merchant category restrictions. Lock it to a single merchant" — and because acceptance is universal. The second stage comes from a settlement problem: card-network payouts to merchants can take "up to 30 days" cross-border because "Visa doesn't move money. Banks do," and bank settlement is "slow and expensive." Card networks are moving to settle with stablecoins instead, which Taylor argues matters acutely for agents because an agent "buying expensive AI tokens" and scaling fast "could run out of money before your revenue arrives." The third stage is where stablecoins stop merely accelerating cards and start doing something cards cannot: letting a master agent create sub-wallets with fine-grained spending rules, across borders, without asking anyone's permission for a company built from agent workforces.

This sequencing sits alongside notes about why agents need payment infrastructure with controls and trust in the first place. Does naming a payments brand increase trust in AI transactions? documents the trust gap that Taylor's first stage implicitly answers: virtual cards work because a known, trusted payments brand sits behind the authorization step, not because the underlying rail is novel. Do consumers want to approve AI shopping purchases before they happen? shows a parallel demand for guardrails at the point of purchase, which is exactly the function Taylor assigns to card controls (single-use, budget caps, merchant locks) rather than to stablecoins. Taylor's argument reframes both surveys' trust and control concerns as reasons the card stage comes first, not as obstacles stablecoins must solve immediately.

The excerpt is a practitioner's prediction published on a fintech newsletter, not a study: it cites named developments (Circle's Nanopay, the x402 standard from Cloudflare and Coinbase, Visa and card-network moves to stablecoin settlement) as evidence the shift is underway, but it does not measure actual agent payment volumes, adoption timelines, or how much of Patrick Collison's claim that "agents will make orders of magnitude more payments than humans" has yet materialized. The implication Taylor draws — that payment infrastructure providers should build for cards now and stablecoin rails next, rather than choosing one — follows only as strongly as that unverified growth premise holds.