MoonPay is expanding the role of artificial intelligence in decentralized finance by integrating Solana-based lending protocol Kamino into PayBox, its AI-focused crypto payment and asset management vault.
The new integration allows eligible users to lend crypto assets or borrow against collateral through natural-language conversations with AI assistants such as ChatGPT and Claude. Instead of navigating Kamino’s traditional interface, users can instruct an AI agent to supply tokens for yield or borrow USDC against deposited collateral.
The move represents another step toward agent-driven DeFi, where AI software can interact directly with decentralized protocols and execute financial transactions on behalf of users.
MoonPay Adds Kamino Lending to PayBox
Kamino is one of the largest lending protocols in the Solana ecosystem. According to DefiLlama data cited in the announcement, the protocol has approximately $1.3 billion in total value locked and around $1 billion in active loans.
Through the PayBox integration, users can give an AI assistant instructions such as supplying supported assets to Kamino to earn lending yield or using deposited crypto as collateral to borrow the dollar-pegged USDC stablecoin.
However, MoonPay is not launching a new lending product. The underlying loans and lending positions continue to operate through Kamino.
What changes is the interface.
Instead of manually interacting with multiple DeFi menus, users can initiate supported transactions using conversational instructions.
AI Agents Can Execute Transactions Within User-Defined Limits
MoonPay is also giving users different levels of control over how AI agents execute DeFi transactions.
Users can require passkey authorization before every transaction, providing an additional confirmation step before funds move. Alternatively, they can allow the AI agent to operate more independently within predefined spending, borrowing, or transaction limits.
That flexibility highlights one of the central questions surrounding agentic finance: how much control should users delegate to automated software?
AI agents could potentially simplify complicated DeFi workflows, but greater autonomy also increases the importance of transaction limits, permissions, and risk controls.
Agentic Crypto Moves Beyond Micropayments
Much of the early activity involving autonomous AI agents and cryptocurrency has focused on small, frequent payments.
Agents can use crypto to purchase computing resources, access datasets, pay for APIs, or use online services without requiring a human to manually approve every transaction.
Coinbase-developed x402, for example, has reportedly processed roughly $50 million across approximately 165 million payments.
MoonPay’s Kamino integration pushes the concept into a significantly different category.
Rather than simply paying for services, AI agents can now help manage lending positions and credit exposure. A single instruction could create a collateralized loan or contribute to a leveraged strategy that may require ongoing monitoring.
That makes agentic DeFi considerably more complex than automated micropayments.
AI-Managed DeFi Introduces New Financial Risks
Artificial intelligence remains an emerging technology, and AI systems can misunderstand instructions or execute actions incorrectly.
When those errors involve cryptocurrency, the financial consequences can be immediate.
Users borrowing through Kamino must also deal with the same liquidation risks that apply to conventional DeFi lending.
If the market value of deposited collateral falls far enough, a borrower’s position can reach Kamino’s liquidation threshold. Liquidators may then repay part of the outstanding debt and receive some of the collateral in return.
An AI interface does not remove those risks.
Users still need to understand collateral ratios, borrowing costs, market volatility, and liquidation conditions before opening a lending position.
The difference is that AI may increasingly become the layer through which those positions are created and managed.
PayBox Continues Expanding Into DeFi
MoonPay introduced PayBox in July with support for several crypto operations, including payments, token swaps, blockchain bridging, and access to yield opportunities through Aave.
The addition of Kamino therefore does not mark PayBox’s first entry into decentralized finance.
Instead, it expands the platform into one of Solana’s largest lending markets and strengthens MoonPay’s push toward AI-assisted capital management.
The integration also reflects a broader shift in the crypto industry. Agentic tools initially focused on executing payments, but developers are increasingly exploring whether AI agents can handle more sophisticated financial activities such as portfolio allocation, lending, borrowing, and yield management.
Geographic Restrictions Limit Access
The Kamino integration will not be available everywhere.
MoonPay said access is unavailable to users in the United States, United Kingdom, European Union, and Australia. Availability can also differ depending on the specific crypto asset and local jurisdiction.
Those restrictions underscore the regulatory complexity surrounding automated financial services, particularly when lending, stablecoins, and decentralized protocols are combined with AI-driven transaction execution.
AI Agents Could Become a New Interface for DeFi
MoonPay’s integration with Kamino illustrates how the user experience around decentralized finance could evolve.
DeFi has traditionally required users to understand wallets, lending dashboards, collateral settings, transaction approvals, and multiple blockchain interactions. AI agents have the potential to hide much of that complexity behind conversational commands.
Instead of manually navigating a lending protocol, users may increasingly be able to tell an AI assistant what financial outcome they want and allow software to coordinate the required on-chain transactions.
The Kamino integration is still subject to important limitations, particularly around AI reliability, liquidation risk, and regional access. Nevertheless, it offers an early example of how agentic crypto could move beyond payments and into active management of decentralized lending and credit positions.
