Why can an AI agent's crypto wallet set its own spending rules, while its virtual card can't?
How do stablecoin wallets differ architecturally from virtual card programmability?
This explores where the spending rules actually sit when an AI agent pays with a stablecoin wallet versus a virtual card, and what each design allows an agent to do on its own.
This explores where the spending rules actually sit when an AI agent pays with a stablecoin wallet versus a virtual card, and what that lets an agent do without asking anyone. The short answer from the corpus is that a virtual card's programmability belongs to the card issuer, while a stablecoin wallet's programmability belongs to whoever holds the wallet. That one difference explains most of the rest. With a virtual card, an issuer creates each card through the card network and attaches controls such as spending limits or merchant restrictions. With stablecoins, a 'master' agent can create new wallets itself, write custom spending rules into them, and hand them to sub-agents with no intermediary approving the step. Fintech commentator Simon Taylor argues that cards can't match this at machine speed Can stablecoins let agents create spending authority without approval?.
That doesn't make wallets simply better. Taylor's larger argument is that these are tools for different jobs, adopted in sequence. Virtual cards come first because their controls are mature and almost every merchant already accepts them. Stablecoin-settled cards come next, keeping the card on the front end while fixing slow cross-border settlement behind it. Native wallets come last, once agents start building teams of sub-agents that need to grant each other spending authority Will agents adopt cards before stablecoins for payments?. Each stage is driven by more complex agent structures, not by one technology beating another. There are early signs the wallet end is becoming real use rather than experiment. On the x402 payment rails, the median time from funding an agent wallet to its first spend dropped from 73 days to 18 minutes, and repeat funding rose to 89% Are AI agent wallets becoming genuinely operational instead of experimental?.
The more interesting question is what you lose when the agent holds the rule-making power. An IMF analysis argues that agent payments should split into three layers. The AI's probabilistic decision to buy something sits on top, and deterministic authorization and settlement sit underneath, because payment infrastructure needs predictability and legal enforceability that AI reasoning can't guarantee Should AI agents make payment decisions independently?. Cards get this separation by default, since the issuer and network are an outside authorization layer the agent can't rewrite. Sub-wallets pull authorization closer to the agents, so the separation has to be designed in on purpose. A useful parallel comes from agent security research. Progent uses a formal solver, not the LLM, to check whether each proposed permission change narrows or expands the agent's authority, and it blocks silent expansions Can an SMT solver prevent LLM policy drift during agent execution?. A sub-wallet tree raises the same question: can a child wallet ever have more spending authority than its parent?
Neither design settles who the agent is. Card credentials stored in a vault protect the consumer's secrets, but they don't tell a merchant that an agent is shopping or on whose behalf Can credential vaulting alone solve merchant-facing agent identification?. That's why the protocol landscape is layered. Visa's TAP and Mastercard's Web Bot Auth handle identity, while ACP, AP2 and UCP handle the application, governance and discovery layers Do agentic commerce protocols compete or complement each other?. Whether money moves through a card or a wallet, identity is a separate layer that has to be stacked on top.
A caveat on depth: the corpus frames this mainly through one commentator's three-stage model plus some usage data. It doesn't go into smart-contract mechanics, chargebacks and dispute handling, or how wallet spending rules are actually enforced. Treat it as a map of the trade-offs rather than a technical comparison.
Sources 7 notes
Taylor argues stablecoin architecture enables a master agent to instantiate new wallets with custom spending rules and delegate them to sub-agents without intermediary approval, a capability virtual cards cannot match at machine speed.
Fintech commentator Simon Taylor argues agent payments follow a three-stage sequence driven by increasing complexity: virtual cards leverage mature controls and universal acceptance, stablecoin-settled cards solve cross-border settlement delays, and native wallets enable multi-agent permission-less structures. The progression reflects different tools for different jobs rather than competition.
Median time from wallet funding to first spend collapsed from 73 days to 18 minutes, and repeat-funding rates climbed from 71% to 89%, signaling agentic micropayments transitioned to habitual use.
An IMF analysis argues that agent-mediated payments must split probabilistic AI decision-making (Layer 1) from deterministic authorization and settlement (Layers 2–3) because core payment infrastructure requires predictability and legal enforceability that probabilistic reasoning cannot guarantee.
Progent uses an SMT solver to verify each LLM-proposed policy update as either a safe narrowing or a risky expansion, automatically applying safe changes while blocking silent privilege escalation even when the LLM is manipulated by adversarial input.
Show all 7 sources
Secure credential storage is a unilateral consumer-security control that prevents agent exposure of secrets, but it cannot substitute for a visible, retailer-legible self-identification protocol that both the account holder and merchant must agree to separately.
Boboev argues ACP, AP2, UCP, TAP, and Web Bot Auth each solve different layers—application, governance, discovery, and identity—creating a capability ladder merchants stack rather than choose between, with each layer absorbing different levels of fraud risk.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- How Agentic AI Will Reshape Payments
- Payments in the Age of AI Agents
- The agentic commerce opportunity: How AI agents are ushering in a new era for consumers and merchants
- North America: Inside the Growth of Agentic Commerce
- Visa Introduces Trusted Agent Protocol: An Ecosystem-Led Framework for AI Commerce
- Agents will use cards first. Then stablecoins.
- Infrastructure for AI Agents
- Measuring AI agent autonomy in practice