AI Agents Could Be the Future of Decentralized Finance (DeFi)

In the past 18 months, Artificial Intelligence (AI) has made a significant leap from novelty to substrate in the crypto world, particularly in Decentralized Finance (DeFi). AI agents are now rebalancing portfolios, summarizing governance, and executing on-chain strategies in real time.

Academic work is surveying the roles of autonomous AI agents in DeFi, focusing on their potential impact on market microstructure and governance design. DeFi, with its decentralized nature, provides a perfect playground for autonomous agents to trade, rebalance, scan risks, arbitrage, and manage market-making (MM) operations, all while using on-chain and forum data for governance.

One such AI wallet companion is AskGina.ai, which can analyze holdings and execute on-chain transactions directly from chat. The emerging end-state is an autonomous agent system where users state an intent, and the agent composes the stack, executes, and transacts without human intervention.

However, the rise of these AI agents brings new challenges, particularly in terms of identity & permissions. Wallet-bound permissions, cryptographic attestations, and policy guards around agent autonomy are essential to ensure secure and controlled operations.

The programmable data layer is another crucial component, offering cross-chain ingestion, normalized schemas, real-time replication/mirroring, and deterministic APIs consumable by agents. Defi App, for instance, offers native intent-based swaps via solver/relayers, routing across multiple aggregators/DEXs.

Latency-aware compute is necessary for triggers related to price/volatility/MEV risk, agent policy evaluation, and pre-trade checks. However, the current state of Web3 is splintered by chains, schemas, and sources, making it difficult for agents to access reliable, low-latency data.

The intent-centric pattern is catching on, with users typing goals and agents handling the plumbing. This offers a scalable solution compared to the status quo. Igris.bot, for example, focuses on destination-based swaps, determining the portfolio source, route, and fees between chains based on the specified outcome.

Chat-based DeFi platforms are emerging, allowing users to interact with decentralized finance applications through natural language interfaces. AI crypto co-pilots like Elsa aim to route, bridge, swap, lend across chains with MPC-secured wallets and safety rails.

Despite the rapid growth of AI-agent projects, no specific agents have established themselves in the DeFi market over the last 18 months, nor is there information available on the exact amount of funds they have raised between January 1, 2023, and June 26, 2025.

AI and crypto are complementary, with blockchains supplying ownership, credible commitments, and identity, which primitive AI systems lack. Vitalik Buterin, co-founder of Ethereum, suggests the most viable role for AI in crypto could be as a player in crypto games, which maps cleanly to markets.

Kuvi.ai positions itself as Agentic Finance, offering an end-to-end intent pipeline and cross-domain ambition (finance, identity, gaming). As of June 26, 2025, AI-agent projects had raised $1.39B year-to-date, outpacing 2024's run-rate.

NVIDIA's Jensen Huang views AI as the start of a new industrial revolution, implying new user layers and automation patterns in finance. As we move forward, it's clear that AI will continue to play a significant role in shaping the future of DeFi.