Loading prices…
🔥BULLISH

Hyundai Card tests AI agent payments with JPMorgan Kinexys

The proof of concept connects autonomous treasury decisions with permissioned Ethereum settlement, extending Hyundai's stablecoin remittance work into institutional cash management.

Hyundai Card tests AI agent payments with JPMorgan Kinexys
Hyundai Card tests AI agent payments with JPMorgan Kinexys
Hyundai Card tests AI agent payments with JPMorgan Kinexys
Hyundai Card tests AI agent payments with JPMorgan Kinexys

Hyundai Card completed a proof of concept with Kinexys, JPMorgan Chase's blockchain-based institutional payments platform, to test whether AI agents can automate cash management, foreign exchange and fund transfers. The simulated scenario involved hypothetical overseas subsidiaries, with Hyundai Card's AI agent monitoring balances, comparing foreign exchange quotes from multiple banks and requesting currency conversions when needed. JPMorgan developed the agents for both organizations and built identity verification, authorization, control and validation mechanisms into the transactions. Transfers within predefined limits were executed automatically, while larger instructions were escalated for human approval or rejected.

Why it matters

The experiment links two capabilities that financial institutions have been developing separately: autonomous decision-making and blockchain-based settlement. It also extends Hyundai's earlier stablecoin cross-border remittance pilot, which moved toward production readiness on the Avalanche blockchain, into a broader treasury workflow involving foreign exchange and liquidity management.

For JPMorgan, the project shows how Kinexys, formerly known as Onyx and operating since 2020, can support agent-to-agent payments between institutional clients. For Hyundai Card, it tests the application of its data science expertise in payments rather than limiting automation to consumer interfaces. The use of a permissioned Ethereum blockchain also provides a controlled environment for institutions that require strict controls over identity, transaction limits and approvals.

Market impact

The immediate result is a simulated proof of concept rather than a live payment network, so the development does not alter current transaction volumes. Its significance is institutional: Hyundai Card and JPMorgan demonstrated a working division of labor between a corporate treasury agent that determines what is needed and a bank agent that verifies, authorizes and executes the instruction.

That model could support more automated corporate payment operations if deployed beyond the test.

Related tokens
$ETH $AVAX

Frequently asked questions

  1. What did Hyundai Card test with JPMorgan's Kinexys platform?

    Hyundai Card tested AI agents that monitored subsidiary balances, compared foreign exchange quotes and requested currency conversions and fund transfers through Kinexys, JPMorgan's permissioned Ethereum blockchain platform.

  2. Were real funds transferred during the Hyundai Card proof of concept?

    No. The scenario used simulated transactions between hypothetical overseas subsidiaries to test the feasibility of automating treasury operations.

  3. How did the AI agents control and validate transactions?

    JPMorgan implemented identity verification, authorization, control and validation mechanisms. Instructions within predefined limits executed automatically, while others required human approval or were rejected.

  4. How does the Kinexys test extend Hyundai's earlier remittance work?

    Hyundai's earlier stablecoin remittance pilot focused on cross-border transfers on Avalanche. The Kinexys test broadened the use case to cash management, foreign exchange and fund transfers with AI agents.

  5. What does the test mean for institutional blockchain adoption?

    It demonstrated a controlled model in which a corporate agent identifies funding needs and a bank agent verifies and executes the transaction, combining autonomous decision-making with permissioned blockchain infrastructure.

Source attribution
Aggregated from CoinDesk · Verified · Last refreshed 10m ago
Open original →