Oobit, a crypto wallet startup backed by Tether, has launched a Visa-supported virtual card designed for AI agents. The card lets autonomous software spend USDT online on behalf of businesses, converting a stablecoin balance into card-network purchasing power. The launch matters because it takes a different route to agent commerce than the protocol-first approaches that dominate the conversation: instead of asking merchants to adopt new payment endpoints, it tunnels agent spending through rails that already accept cards everywhere.
What the Product Actually Does
The core mechanic is a virtual Visa card attached to an Oobit wallet. An AI agent operating for a business receives spending authority over that card and uses it for online purchases — subscriptions, API access, inventory, services — denominated in USDT. At the point of transaction the card behaves like any other Visa credential: the merchant sees a card payment, not a stablecoin transfer. Settlement behind the scenes converts USDT into the fiat the card network requires. In effect, Oobit has made agent spending legible to the entire card acceptance universe without any merchant-side integration. That is the entire value proposition, and it is not a trivial one.
Why This Route Differs From x402-Style Rails
Protocol-level agent payments — x402 being the most prominent — invert this model. There, the agent pays over an open HTTP-native protocol in stablecoins, and the merchant must implement the receiving endpoint. The reward is lower-cost, programmable, native-crypto settlement; the cost is adoption friction on the supply side. Oobit’s card approach has zero merchant friction but reintroduces the card networks as intermediaries, with their fees, chargeback logic, and authorization rules. It also concentrates spending-policy enforcement inside Oobit’s card controls rather than in a public protocol. Both bets can coexist: x402 suits machine-to-machine payments between crypto-native services, while card intermediation suits agents buying from ordinary web merchants. Oobit has effectively chosen the second market.
Tether’s Positioning
Oobit is Tether-backed, and the product is USDT-denominated. That is a deliberate placement. USDT dominates offshore and emerging-market stablecoin usage, and agent spending by businesses — procurement, tooling, ad spend — is a natural corporate use case where dollar-pegged balances already sit in treasuries. Context matters here: industry reporting now counts 165 million-plus autonomous payments processed via x402, and AWS, Coinbase, Stripe, Mastercard, and Visa are all building agent payment capability. Tether does not have an equivalent open protocol in market at comparable scale, so backing a card intermediary that channels USDT into Visa acceptance is a pragmatic way to keep USDT in the settlement path of agent commerce without winning a protocol war.
Open Questions
The model’s weak points are the classic ones. Authorization — proving a human permitted the agent’s spend — remains unsolved industry-wide, and card rules were written for humans holding plastic, not software with delegated authority. Spending limits, per-merchant controls, and audit trails will live inside Oobit’s product rather than a standard, which raises lock-in and compliance questions when regulators eventually examine agent-initiated card transactions. Fees are another unknown: card-intermediated stablecoin spending typically costs meaningfully more than on-chain transfer. And whether card networks remain comfortable with fully autonomous cardholders, rather than humans clicking approve, is a policy question Visa will answer eventually — perhaps restrictively.