Authors: 0xjacobzhao and ChatGPT 4o
Thanks to Lex Sokolin (Generative Ventures), Stepan Gershuni (cyber.fund) and Advait Jayant (Aivos Labs) for their valuable suggestions on this article. Feedback was also sought from project teams such as Giza, Theoriq, Olas, HeyElsa, Almanak, Brahma.fi during the writing process. This article strives for objective content and accurate expression, but since some viewpoints involve subjective judgments, there might be deviations. We urge readers to read critically and understand.
In the current crypto industry, stablecoin payments and DeFi applications are among the few sectors that have been verified to possess real demand and long-term value. Meanwhile, the blossoming of Agents is gradually becoming the actual landing form in the AI industry aimed at user interfaces, serving as a key intermediary layer connecting AI capabilities and user needs.
In the fusion of Crypto and AI, especially in the direction where AI technology feeds back into Crypto applications, current explorations are mainly focused on three typical scenarios:
Conversational interaction-type Agents: Mainly focused on chatting, companionship, and assistant-like functions. Although most still rely on general large models, their low development threshold and natural interaction, combined with token incentives, have made them the earliest forms to market to gain user attention.
Information integration-type Agents: Focus on the intelligent integration of online and on-chain information. Kaito, AIXBT, etc., have achieved success in online but non-chain information search integration, while on-chain data integration directions are still in the exploration stage without obvious breakout projects.
Strategy execution-type Agents: Extending from stablecoin payments to DeFi strategy execution, focusing on Agent Payment and DeFAI in two major directions. Such Agents are more deeply embedded in on-chain trading and asset management logic, and are expected to break through the speculation bottleneck, forming intelligent execution infrastructure with financial efficiency and sustainable yields.
This article will focus on the evolutionary path of the integration of DeFi and AI, outlining its development stages from automation to intelligentization, and analyzing the infrastructure, scenario space, and key challenges of strategy execution Agents.
Three phases of DeFi intelligentization: The transition from Automation to Copilot to AgentFi.
In the evolution of DeFi intelligence, we can divide system capabilities into three phases: Automation (automation tools), Intent-Centric Copilot (intent-driven assistants), and AgentFi (on-chain intelligent agents).
Automation is more like a rule trigger: executing fixed tasks based on preset conditions, such as arbitrage, rebalancing, stop-loss, etc., and cannot generate strategies or operate independently.
Copilot introduces intent recognition and semantic parsing capabilities, allowing users to input natural language, the system understands, decomposes, and suggests execution paths, but ultimately still requires user confirmation, making the execution chain not closed.
AgentFi represents a complete intelligent closed-loop of 'perception → reasoning/strategy generation → on-chain execution → evolution', being an intelligent agent (Agent) with on-chain autonomous execution and continuous evolution capabilities.
To determine whether a project truly belongs to AgentFi, it needs to meet at least three of the following five core standards:
Autonomous perception of on-chain states/market signals (not static input, but real-time monitoring).
Capable of strategy generation and combination (not pre-set strategies, but able to self-develop action plans based on context).
Can autonomously execute operations on-chain (no user interaction required, able to perform complex actions such as swap/lend/stake).
Has persistent states and evolutionary capabilities (Agent has a lifecycle, can operate long-term and self-adjust based on feedback).
Has Agent-Native architecture (such as dedicated Agent SDK, managed execution environment, Agent middleware, etc.).
In other words, automated trading ≠ Copilot, much less AgentFi: Automated trading is merely a 'rule trigger', Copilot can understand user intentions and provide operational suggestions but still relies on human participation; while true AgentFi is 'an intelligent agent with perception, reasoning, and autonomous on-chain execution capabilities', able to complete strategy loops and continuous evolution without human intervention.
Analysis of the intelligentization adaptability in DeFi scenarios:
In the DeFi (Decentralized Finance) system, core application scenarios can be roughly divided into asset circulation and exchange types and yield finance types. We believe there are significant differences in adaptability on intelligent paths between these two types of scenarios.
1. Asset circulation and exchange scenarios.
Asset circulation and exchange scenarios primarily feature atomic interactions, including swap transactions, cross-chain bridges, fiat entry and exit, etc. Their essential characteristics are 'intent-driven + single atomic interaction', which do not involve yield strategies, state maintenance, and evolution logic during the trading process. They are mostly suitable for lightweight execution paths of Intent-Centric Copilot and do not belong to AgentFi.
Due to its lower engineering threshold and simple interaction, most DeFAI projects in the current market are at this stage, which do not constitute AgentFi closed-loop intelligent agents; however, for a few high-level complex Swap strategies (such as cross-asset arbitrage, perpetual hedging LP, leverage rebalancing, etc.), AI Agent capabilities are indeed needed, and they are still in the early exploratory stage.
2. Asset income financial scenarios.
Asset income financial scenarios feature clear yield targets, complex strategy combination spaces, and dynamic state management needs, making them a natural fit for AgentFi's 'strategy loops + autonomous execution' model. Their core characteristics are as follows:
Quantifiable yield targets (APR / APY) facilitate the establishment of optimization functions by the Agent;
Broad strategy combination space covering multiple assets, multiple time periods, multiple platforms, and multiple interaction processes;
Operations require frequent management and real-time adjustments, suitable for execution and maintenance by on-chain agents (Agents).
Due to multiple factors such as yield duration, volatility frequency, on-chain data complexity, cross-protocol integration difficulty, and compliance restrictions, the adaptability and engineering feasibility of different yield scenarios in the AgentFi dimension vary significantly. The suggested priority is as follows:
High-priority business landing direction:
Lending/Borrowing: Interest rate fluctuations are easy to track with standardized execution logic, suitable for lightweight intelligent agents.
Liquidity Mining (Yield Farming): The pools are dynamic and frequent, with a large strategy combination space and high yield fluctuations. AgentFi can significantly optimize annual returns and interaction efficiency, but engineering implementation poses certain challenges.
Medium to long-term exploratory layout directions:
Pendle yield rights trading: Clear time dimensions and yield curves, suitable for Agent management of maturity rotation and pool arbitrage.
Funding Rate Arbitrage: Theoretical returns are considerable, requiring solutions for cross-market execution and off-chain interaction challenges, with significant engineering difficulties.
LRT dynamic combination structure: Static staking is not adaptable, but strategies like LRT + LP + Lending can be tried for automatic adjustments.
RWA multi-asset portfolio management: Difficult to land in the short term, but Agents can provide assistance in portfolio optimization and maturity strategies.
Introduction to projects in the intelligentization of DeFi scenarios:
1. Automation Tools (Automation Infra): Rule triggers and conditional executions.
Gelato is one of the earliest infrastructures for DeFi automation, previously providing condition-triggered task execution support for protocols like Aave and Reflexer, but it has now transformed into a Rollup as a Service provider. The main battlefield for on-chain automation has also shifted to DeFi asset management platforms (DeFi Saver, Instadapp). These platforms integrate standardized automated execution modules including limit order settings, liquidation protection, automatic rebalancing, DCA, grid strategies, etc. In addition, we see some more complex DeFi automation tool platform projects:
Mimic.fi (https://www.mimic.fi/)
Mimic.fi is an on-chain automation platform that serves DeFi developers and project parties, supporting the construction of programmable automation tasks on chains like Arbitrum, Base, and Optimism. Its core achieves automated execution of cross-protocol operations through 'if-then' rule triggers, with an architecture divided into Planning (task and trigger definitions), Execution (intent broadcasting and execution bidding), and Security (triple verification and security control) layers. Currently, it adopts an SDK access method, and the product is still in the initial deployment stage.
AFI Protocol (https://www.afiprotocol.ai/)
AFI Protocol is an algorithm-driven Agent execution network, supporting 7×24 hours of non-custodial automated operations, focusing on solving issues of execution decentralization, strategy barriers, and risk response in DeFi. Its design targets institutions and advanced users, offering programmable strategies, permission management, and SDK tools, and has launched a yield-bearing stablecoin afiUSD as its native asset. It is currently in the internal testing phase at Sonic Labs, not yet publicly launched or open for retail user access.
2. Intent-Centric Copilot: Intent expression and execution suggestions.
The DeFAI concept, which was once hot at the end of 2024, aside from some speculation based on meme tokens, most projects essentially belong to the Intent-Centric Copilot type - that is, expressing user intent through natural language, and the system responds with trading suggestions or completes basic on-chain operations. Its core capability remains at the 'intent recognition + Copilot-style auxiliary execution' stage, and has not formed a complete strategy loop and continuous optimization mechanism. Many products have obvious shortcomings in semantic understanding, cross-protocol calls, and feedback responses, with a generally poor user experience and relatively limited functional boundaries.
HeyElsa (https://app.heyelsa.ai/)
HeyElsa is an AI Copilot positioned for Web3 scenarios, empowering users to complete various on-chain operations including trading, cross-chain bridging, NFT purchasing, stop-loss settings, and Zora token creation through natural language interaction. As a multifunctional conversational crypto assistant, it covers users from beginners to advanced traders (including highly active degen groups), and currently supports real-time interaction across over 10 mainstream blockchains. The platform's average daily trading volume has reached $1 million, with daily active users maintained between 3,000 and 5,000. The system has integrated yield optimization strategies and automated intent execution modules, initially building the foundational capability framework for AgentFi applications.
Bankr (https://bankr.bot/)
Bankr is an intent trading assistant integrating AI, DeFi, and social scenarios. Users can issue commands via natural language on the X platform or dedicated terminal to complete operations such as swaps, limit orders, cross-chain bridging, token issuance, NFT minting, supporting Base, Solana, Polygon, and Ethereum mainnet. Bankr has built a complete Intent → Compilation → Execution link, emphasizing a minimal trading experience and seamless operations within a social environment, and activates the ecosystem through token incentives and revenue sharing mechanisms.
Griffain (https://griffain.com/)
Griffain is a multi-functional AI Agent platform deployed on Solana, supporting natural language interaction between users and Griffain Copilot, enabling on-chain operations such as asset queries, swaps, NFT trading, and LP management. The platform features multiple intelligent modules and encourages community participation in Agent creation and sharing. Technically built on the Anchor Framework with components like Jupiter and Tensor, it emphasizes mobile adaptation and front-end composability. Currently supports over 10 core Agent modules, with strong execution capabilities and ecological linkage.
Symphony (https://www.symphony.io/)
Symphony is on-chain execution infrastructure for AI Agents, building a full-stack system covering intent modeling, intelligent path discovery, RFQ execution, and account abstraction, aiming to become the core module of DeFi intelligent execution layers. The platform has launched a conversational assistant, Sympson, which has market query and strategy suggestion functions, but on-chain execution has not yet been opened.
Hey Anon (https://heyanon.ai/)
HeyAnon is a DeFAI platform that combines intent interaction, on-chain execution, and intelligence analysis, supporting multi-chain deployment (Ethereum, Base, Solana, etc.) and cross-chain bridging (LayerZero, deBridge). Users can complete operations such as swaps, lending, and staking through natural language and obtain on-chain sentiment and market dynamics analysis. Although the project has gained high attention due to founder Sesta, it is still in the Copilot stage, and its core strategies and execution intelligence have not been fully realized, and long-term development still needs observation.
The above scoring system is mainly based on the current availability of the product, user experience, and the feasibility of executing the publicly available roadmap, and has a certain subjectivity. Please note that this assessment does not involve code security checks and does not constitute investment advice, thank you for your understanding.
3. AgentFi intelligent agents: strategy loops and autonomous execution.
We believe that AgentFi is a more advanced form in the intelligent leap of DeFi compared to Intent Copilot. Agents possess independent yield strategies and on-chain autonomous execution capabilities, which can significantly enhance users' strategy execution efficiency and capital utilization. By 2025, we are pleased to see an increasing number of AgentFi projects being landed or in the planning phase, mainly focused on lending and liquidity mining directions, with representative projects including Giza ARMA, Theoriq AlphaSwarm, Almanak, Brahma, Olas series, etc.
Giza ARMA (https://arma.xyz/)
ARMA is an intelligent agent product launched by Giza, designed specifically for stablecoin cross-protocol yield optimization. It is deployed on the Base network and supports multiple mainstream lending protocols like Aave, Morpho, Compound, Moonwell, with capabilities for cross-protocol rebalancing, automatic compounding, and smart currency exchange. ARMA's strategy system can monitor stablecoin APR, trading costs, and yield differences in real-time, automatically adjusting capital allocation, with verified yields significantly higher than static holdings. Its architecture consists of smart accounts, session keys, core agent logic, protocol access, risk management, and accounting modules, ensuring safe and efficient automated execution in a non-custodial mode.
ARMA has now been fully launched and is continuously iterating. With its modular architecture, security mechanisms, and good early operational data, ARMA has become one of the most feasible Agent products in DeFi automated yield management, and is one of the few AgentFi projects with both deep concepts and practical products.
Reference research report (A new paradigm for stablecoin yields: From AgentFi to XenoFi) link: https://x.com/0xjacobzhao/status/1925226999699964158
Theoriq (https://www.theoriq.ai/)
Theoriq Alpha Protocol is a multi-agent collaboration protocol focused on DeFi scenarios, with its core product Alpha Swarm concentrating on liquidity management, aiming to build a full-chain automation loop of 'perception - decision - execution'. It is composed of three types of Agents: Portal (on-chain signal perception), Knowledge (data analysis and strategy selection), and LP Assistant (strategy execution), capable of achieving dynamic asset allocation and yield optimization without human intervention. The underlying Alpha Protocol provides Agent registration, communication, parameter configuration, and development tool support, serving as the operational foundation for the entire Swarm collaborative system. Through AlphaStudio, users can browse, call, and combine various Agents to build a modular, expandable automated trading strategy network.
As one of the first projects of Kaito Capital Launchpad, Theoriq recently completed a community fundraising of $84 million and is about to TGE. Theoriq has recently launched the AlphaSwarm Community Beta testnet, and the mainnet version is also about to be officially released.
Reference research report (Theoriq research report: The evolution of AgentFi in liquidity mining yields) link: https://x.com/0xjacobzhao/status/1948545449016918511
Almanak (https://almanak.co/)
Almanak is an intelligent Agent platform for DeFi strategy automation, combining non-custodial security architecture with a Python strategy engine to help traders and developers deploy sustainably running on-chain strategies.
The platform's core consists of Deployment (execution components), Strategy (strategy logic), Wallet (Safe+Zodiac security module), and Vault (strategy assetization), supporting yield optimization, cross-protocol interaction, liquidity provision, and automated trading. Compared to traditional DeFi tools, Almanak emphasizes AI-assisted market perception and risk management capabilities, already possessing 24/7 intelligent operational capabilities and planning to introduce multiple agents and AI decision-making systems to create the next generation of AgentFi infrastructure.
Almanak's strategy system is based on a state machine program built in Python, serving as the 'decision brain' for each Agent, capable of automatically formulating and executing on-chain operations based on market data, wallet status, and user-set conditions. The platform provides a complete Strategy Framework, supporting encapsulation of on-chain trading, lending, liquidity provision, and other operational modules (Action Bundle), without the need to write underlying contract code, while ensuring strategy confidentiality and operational security through cryptographic isolation, permission control, and monitoring mechanisms. Users can write strategies through the SDK, and in the future, natural language creation of strategies will also be supported, achieving a smooth transition from complex logic to no-code experience.
The product has launched a USDC lending Vault based on the Ethereum mainnet, while more complex trading strategies are in the testing phase and require whitelist access. Almanak is about to join the cookie.fun cSNAPS campaign for community public fundraising, which is worth looking forward to.
Brahma (https://brahma.fi/)
Brahma is positioned as 'the intelligent capital coordination layer' (The Orchestration Layer for Internet Finance), aiming to abstract on-chain accounts, execution logic, and off-chain payment processes, helping users and developers efficiently coordinate to manage on-chain and real-world assets. Through Smart Accounts, continuously operating on-chain Agents, and Capital Orchestration Stack, Brahma provides users with an intelligent capital management experience that does not require backend operations.
Currently launched representative Agents:
Felix Agent: Automatically optimizes feUSD debt warehouse interest rates, prevents liquidation, and saves interest;
Surge & Purge Agent: Tracks volatility and executes automatic trading;
Morpho Agent: Deploys and rebalances Morpho treasury funds;
ConsoleKit framework: Supports the access of any AI model, unifying strategy execution and asset management.
Olas (https://olas.network/)
The AgentFi products BabyDegen series launched by Olas Network include Modius Agent and Optimus Agent, both deployed on-chain, covering multiple chain ecosystems (Solana, Mode, Optimism, Base), and possessing complete on-chain interaction capabilities, strategy execution capabilities, and autonomous asset management mechanisms.
BabyDegen is an AI trading agent running on Solana, implementing automated buying and selling based on CoinGecko data and community strategy libraries, currently integrating Jupiter DEX and in the Alpha testing phase.
Modius Agent targets the Mode network, focusing on USDC and ETH portfolio management, having integrated Balancer, Sturdy, and Velodrome, supporting users in setting preferences to automatically execute strategies 24/7.
Optimus Agent is compatible with the three mainnet networks of Mode, Optimism, and Base, integrating more protocols like Uniswap and Velodrome, providing flexible multi-chain strategy combinations, suitable for mid to high-level users to build automated asset management systems.
Axal (https://www.getaxal.com/)
Axal's core product Autopilot Yield provides a one-stop, non-custodial, verifiable yield management experience, integrating mainstream protocols like Aave, Morpho, Kamino, Pendle, Hyperliquid, and with on-chain strategy execution + risk control as the core design concept, empowering ordinary users to easily enter complex on-chain yield networks.
Conservative strategies focus on low-risk, mainstream stable yield scenarios, primarily deploying funds on well-established platforms like Aave and Morpho, with annual yields around 5-7%. It achieves steady appreciation through TVL monitoring, stop-loss mechanisms, and head strategy filtering, suitable for users pursuing capital safety and long-term returns.
Balanced strategies provide moderate risk and higher yield potential (10-20% APY), using wrapped stablecoins (like feUSD, USDxL), liquidity provision, and arbitrage neutral positions. The strategies are more diversified, with complex yield compositions, controlled exposure through Axal's automatic monitoring and dynamic adjustments.
Aggressive strategies target high-risk, high-yield preference users, covering high-leverage LP, cross-platform linking, low liquidity asset market-making, volatility capture, etc. Theoretical annual returns can exceed 50%. Axal's intelligent agents can set stop-loss, automatically exit, and redeploy logic at the strategy level, providing users with the last line of protection in high-risk environments.
Fungi.ag (https://fungi.ag/)
Fungi.ag is a fully automated AI Agent designed for USDC yield optimization, capable of automatically allocating funds between multiple lending protocols like Aave, Morpho, Moonwell, and Fluid, optimizing capital allocation based on yield rates, fees, and risks. Users do not need to operate manually, just authorize the Session Key to enable the Agent to execute strategies in a non-custodial mode. It currently supports the Base chain and plans to expand to Arbitrum and Optimism. Fungi also opens up the Hypha custom strategy script interface, supporting community development of strategies like DCA and arbitrage, and building the ecosystem through DAO and social platforms.
ZyFAI (https://www.zyf.ai/)
ZyFAI is a DeFi intelligent assistant platform deployed on Base and Sonic networks, combining on-chain interaction interfaces and AI-assisted modules to help users with intelligent asset management under different risk preferences. Its core is divided into three types of strategies:
Safe Strategy: Specifically designed for conservative users, focusing on mainstream protocols such as Aave, Morpho, Compound, Moonwell, Spark, which have been audited and verified, primarily promoting unilateral deposits of USDC and stable yield opportunities, emphasizing asset safety and long-term reliability.
Yieldor Strategy: Aimed at high-risk tolerance users, requires holding 20,000 ZFI tokens to unlock. Covers high-yield protocols including Pendle, YieldFi, Harvest Finance, Wasabi, and supports complex strategies such as DEX LP, yield splitting, leverage Vaults, etc. Future expansions will include structured products like Looping and Delta-neutral.
Airdrop Strategy: A future strategy still in development, aimed at acquiring more airdrop incentives.
The above scoring system is mainly based on the current availability of the product, user experience, and the feasibility of executing the publicly available roadmap, and has a certain subjectivity. Please note that this assessment does not involve code security checks and does not constitute investment advice, thank you for your understanding.
The realistic path and high-level vision of AgentFi.
Undoubtedly, lending and liquidity mining are the most genuinely valuable and easiest to land business scenarios for AgentFi in the short term, having matured in the DeFi world and due to the following common characteristics, are naturally suitable for the introduction of intelligent agents.
Broad strategy space with multiple optimization dimensions. Lending, beyond chasing the highest yield, can explore strategies such as interest rate arbitrage, leverage cycles, debt refinancing, and liquidation protection.
Liquidity mining encompasses rich strategy orchestration space including APR tracking, LP rebalancing, reinvestment, and compound interest.
Highly dynamic, suitable for real-time perception and response by intelligent agents: interest rate fluctuations, TVL volatility, reward structure changes, new pools coming online, new protocols appearing, etc., will all affect optimal strategy paths and require dynamic adjustments.
There are opportunity costs in execution windows, and automation value is significant: funds not deployed in the optimal pool will drag down returns and need to be automatically migrated.
It should be particularly noted that lending agents, due to stable data structures and relatively simple strategies, have high feasibility for landing, such as Giza's Arma and other lending AgentFi projects that have officially launched. However, the management of liquidity mining requires real-time response to price fluctuations, volatility changes, and fee accumulation, posing extremely high demands on the Agent's data perception, strategy judgment, and on-chain execution. LP Agents not only need to accurately predict market states but also need to dynamically adjust positions and redistribute yields on-chain, which adds to the engineering complexity, a challenge that projects like Theoriq are also tackling.
Excluding lending and liquidity mining, based on AgentFi's adaptability, there are some visions for medium to long-term exploratory layout directions:
Pendle yield rights trading: Clear time dimensions and yield curves, suitable for Agent management of maturity rotation and pool arbitrage.
Pendle, with its unique structure of 'yield splitting + maturity mechanism + yield rights trading', provides a natural strategy orchestration space for AgentFi. Its assets are divided into PT (Principal Token) and YT (Yield Token), where the former represents the principal redeemable at maturity, suitable for stable fixed income configurations; the latter represents yield rights, which fluctuate and can be used for speculation, mining, and arbitrage. Around these two types of assets, users can build various complex strategies including fixed income holdings, YT farming, maturity capital management, yield spread arbitrage, and portfolio hedging.
In practical scenarios, Pendle faces many user pain points that urgently need AgentFi solutions: for instance, high-yield pools are mostly concentrated in the short term of 1-3 months and need to be manually reconfigured upon maturity; the YT yield rates of different pools fluctuate significantly, making tracking and rotation costs high; while the PT+YT combination strategy involves complex pricing judgments and position rebalancing. If AgentFi could automate the entire process from strategy identification, liquidity configuration, to maturity rotation and redeployment based on user yield preferences and risk tolerance, it would significantly enhance capital efficiency and user experience.
Pendle's three features of 'term-based, split, and dynamic' align very well with the strategy expression and execution paths of AgentFi, especially in areas like automatic reinvestment, implied yield arbitrage, and yield pool rotation, exhibiting high frequency and high strategy characteristics, making it very suitable for building a 'yield agent Swarm' or Portfolio Agent system. In the future, if it can combine intent expression (such as 'annualized 10%, withdrawable in 6 months') with an automatic execution framework, Pendle will become one of the most representative modules for AgentFi.
Funding Rate Arbitrage: Theoretical returns are considerable, but challenges in cross-market and cross-chain interactions present significant engineering difficulties.
Although the on-chain options track is gradually cooling due to reasons such as pricing absence, complex execution, and poor combinability, perpetual contracts remain one of the most active scenarios in current on-chain derivatives, providing a point of integration for AgentFi. Around funding rate arbitrage, basis trading, and multi-platform hedging strategies, AgentFi can display intelligent capabilities in perception, judgment, execution, and combination management.
In terms of structural design, AgentFi can embed four key modules: First, the data perception module supports real-time capture of on-chain and CEX funding rates, holding costs, and market depth; Second, the intelligent decision-making module dynamically judges whether to open and adjust positions based on arbitrage thresholds, leverage levels, and liquidation boundaries; Third, the automatic execution module completes position deployments or profit-taking closing operations once the triggering condition is met; Fourth, the combination management module supports collaborative scheduling across multiple chains, accounts, and strategies.
However, the real challenges include: firstly, the current on-chain AgentFi mainly focuses on smart contract interactions and does not yet have a general framework for direct access to CEX APIs; secondly, high-frequency strategies have extremely high requirements for execution efficiency, gas costs, and slippage control; thirdly, complex arbitrage scenarios usually require multiple Agents to cooperate, necessitating swarm-style collaboration.
Ethena's funding rate arbitrage has relied on a highly automated execution system. Although Ethena currently does not yet have AgentFi characteristics, if it further opens up strategy modules in the future, builds a distributed Agent Swarm, and realizes funding goal expression through intent-driven methods, its system might naturally transition into a complete AgentFi infrastructure.
Staking and Restaking: Naturally not suitable for AgentFi, but there is some possibility for LRT dynamic combinations.
Overall, traditional Staking and Restaking are not suitable application scenarios for AgentFi. This is because the single-chain staking process is simple, yields stable, decisions are single, and exit waiting periods are long, making it difficult to support the intelligent value emphasized by AgentFi.
However, in more complex staking constructions, AgentFi has certain available space. This includes focusing on the composability of LST/LRT type assets (such as stETH, rsETH), avoiding direct touch with the native ETH unstake process; secondly, emphasizing the construction of Restaking + collateral + derivatives combination strategies, circumventing the time lag caused by unstaking; thirdly, deploying continuously optimizing monitoring-type strategies, dynamically assessing AVS risks, APR changes, and reorganizing positions, etc.
In addition, the current Restaking track also faces structural challenges: on one hand, market heat is rapidly cooling, on the other hand, there is a serious imbalance between supply (staked ETH) and demand (AVS security needs), and asset leasing lacks practical application scenarios. Leading projects like EigenLayer and Either.fi have already attempted to transform. Therefore, Staking/Restaking may become modular strategy components of AgentFi in the future rather than the most core application landing scenarios.
RWA assets: US Treasury protocols are not ideal scenarios, and multi-asset portfolio management structures have exploratory value.
The current mainstream RWA protocols generally use US Treasury bonds (T-bills) as the underlying assets, with a design focus on providing users with stable, safe, and compliant on-chain yield carriers. However, from the perspective of AgentFi, products of this type are not suitable for high-frequency or strategy-driven smart agent embedding due to characteristics such as stable asset nature (annual yields typically stable in the 4-5% range with minimal interest rate differentials, lacking optimization strategy space), low operation frequency (clear lock-up periods and reinvestment cycles, not suitable for frequent rotation, also difficult to achieve high-frequency compounding), and strong compliance restrictions (involving investor KYC verification and regional restrictions). In addition, the non-interoperability of asset structures between protocols also limits the Agent's ability to perform combination routing and liquidity aggregation operations.
Nevertheless, there are still several potential directions that could become long-term expansion paths for AgentFi:
1. Multi-asset RWA configuration agents (RWA Multi-Asset Portfolio): As RWA products gradually expand into fields such as real estate, credit bonds, accounts receivable, etc., users may express the intent of 'allocating a basket of stable yield assets and adjusting periodically'. Configuration agents periodically complete asset weight adjustments and redeploy maturing assets, building a medium to long-term yield stabilizer.
2. The fusion structure of RWA and DeFi (RWA-as-Collateral & Custody Reuse): Some protocols are exploring the use of tokenized T-bills as collateral assets in DeFi lending systems. Under this structure, the Agent can assist users in automatically completing deposit operations, interest rate comparisons, collateral rebalancing, etc., forming dual yield paths. Assuming that RWA assets achieve widespread circulation on platforms like Pendle and Uniswap, the Agent can track the premiums and implied yield changes of Tokens across different platforms, constructing automatic arbitrage and rolling deployment strategies. As the market matures, it may become an important breakthrough for AgentFi in the RWA sector.
Swap trading combinations, upgrading from Intent infrastructure to AgentFi strategy engine.
In the current DeFi intelligent ecosystem, Swap trading introduces account abstraction and Intent intent models, hiding complex multi-chain path selections of DEXs, driving user trades to completion with simple inputs, significantly lowering interaction thresholds. However, such systems remain at the 'atomic action automation' level, lacking real-time perception and response to environmental changes, and have not yet introduced target-oriented strategy execution mechanisms, and do not possess the intelligent agent characteristics of AgentFi.
In the AgentFi framework, Swap operations are no longer single actions but larger scale combination strategies. For example, when a user expresses 'hoping to configure stETH and USDC combinations for the highest yield', the Agent can automatically complete multiple Swaps (e.g., USDC → ETH → stETH), Restaking, splitting Pendle PT/YT, configuring arbitrage strategies, and recovering yields.
Further, Swap plays a key role in the following three types of AgentFi scenarios:
Part of the combination yield strategy: as a capital scheduling relay station, Swap supports the Agent to automatically complete asset allocation paths, enhancing strategy execution efficiency.
Cross-market arbitrage / delta neutral strategy: By comparing different price sources on-chain, the Agent can dynamically adjust positions and build hedging combinations.
Trade behavior risk defense: When detecting large transactions, the Agent can automatically assess slippage, execute in batches, and avoid potential MEV attacks.
Therefore, a truly AgentFi-characterized Swap Agent must possess the following capabilities: dynamic strategy perception, cross-protocol scheduling, optimal funding path, market timing judgment, and risk prevention. Future Swap Agents should serve multi-strategy combinations, dynamic position adjustments, and cross-protocol value captures; the road ahead is long and challenging.
Roadmap of DeFi intelligent evolution: From automation tools to intelligent agent networks.
In summary, we have witnessed the evolutionary path of DeFi intelligentization from automation tools to intent assistants to intelligent agents.
The first phase is 'Automation Tools (Automation Infra)', characterized by achieving basic on-chain operation automation through rule triggers and conditional executions. For example, triggering trades or rebalancing tasks based on preset conditions like time and price. Representative systems are mostly underlying execution frameworks, typical examples include Gelato, Mimic, etc.
The second phase is 'Intent-Centric Copilot', emphasizing the expression of user intent and the generation of execution suggestions. Systems in this phase are no longer limited to 'what to do', but attempt to understand 'what the user wants' and then provide the best execution path suggestions. Representative projects include Bankr and HeyElsa, which mainly enhance intent recognition and interaction experience to lower the barriers to DeFi usage.
The third phase is 'AgentFi agents', marking the formation of strategy loops and on-chain autonomous execution. Agents can automatically complete perception, decision-making, and execution based on real-time market conditions, user preferences, and strategy logic, truly achieving 7×24 hours of unmanaged on-chain capital management. At the same time, AgentFi can autonomously manage user funds without requiring the user to authorize each step of operation, a mechanism that has sparked significant discussions regarding safety and trust mechanisms, and has also become an unavoidable core issue in AgentFi design. Representative projects include Giza ARMA, Theoriq AlphaSwarm, Almanak, Brahma, etc., all of which have certain landing capabilities in strategy deployment, security architecture, and product modules, and are the backbone forces in the direction of DeFi intelligent agents.
We look forward to the emergence of 'AgentFi advanced intelligent agents' in the future, not only achieving autonomous execution but also covering complex cross-protocol and cross-asset business scenarios. This is our vision for the advanced form of intelligent DeFi in the future.
Pendle yield rights trading: Future intelligent agents will fully take over maturity rotation and strategy orchestration, maximizing capital efficiency.
Funding Rate Arbitrage: Cross-chain arbitrage agents are expected to precisely capture every opportunity in the funding rate differential.
Swap strategy combinations: Swap is a key node in the multi-strategy yield paths of intelligent agents, achieving a leap in combination value.
Staking and Restaking: Intelligent agents will continuously optimize staking combination strategies, dynamically balancing yield and risk.
RWA asset management: When the on-chain world welcomes diversified physical assets, intelligent agents will allocate globally stable yield assets.