Can the deep integration of DeFi and AI in DeFAI give rise to a new wave of AI Agents?

Original Title: "The Deep Integration of DeFi and AI - Can DeFAI Spark a New Wave of AI Agents?"

Original Author: Ac-Core, YBB Capital Research

1. The Story of DeFAI

1.1 What is DeFAI

In a concise manner, DeFAI stands for AI+DeFi. Regarding AI, the market has gone through rounds of hype, from AI computing power to AI memes, from different technical architectures to different infrastructures. Despite the recent overall market correction of AI Agent valuations, the concept of DeFAI is emerging as a new breakthrough trend. Current DeFAI can be broadly categorized into AI Abstraction, Autonomous DeFi Agents, and Market Analysis and Prediction. The specific breakdown within these categories is shown in the following figure.

Image Source: Author's Own

1.2 How DeFAI Works

Within the DeFi system, the core behind the AI Agent is the LLM (Large Language Model). Its operation involves a multi-layered process and technology, covering all aspects from data collection to decision execution. According to research by @3sigma in the IOSG article, most models follow six specific workflows: data collection, model reasoning, decision-making, custody and operation, interoperability, and wallets. These are summarized as follows:

1. Data Collection: The primary task of the AI Agent is to have a comprehensive understanding of its operational environment. This includes obtaining real-time data from multiple sources:

● On-chain Data: Real-time blockchain data, such as transaction records, smart contract states, and network activity, is obtained through indexers, oracles, and other means. This helps the Agent stay synchronized with market dynamics;

● Off-chain Data: Price information, market news, and macroeconomic indicators are obtained from external data providers (such as CoinMarketCap, CoinGecko) to ensure the Agent's understanding of external market conditions. This data is usually provided to the Agent through API interfaces;

● Decentralized Data Sources: Some Agents may obtain price oracle data through decentralized data feed protocols to ensure data decentralization and trustworthiness.

2. Model Inference: After data collection is complete, the AI Agent enters the inference and computation stage. Here, the Agent relies on multiple AI models for complex inference and prediction:

● Supervised Learning and Unsupervised Learning: By training on labeled or unlabeled data, AI models can analyze market and governance forum behavior. For example, they can predict future market trends by analyzing historical transaction data, or infer the outcome of a governance forum proposal by analyzing forum data;

● Reinforcement Learning: Through trial and error and feedback mechanisms, AI models can autonomously optimize strategies. For example, in token trading, an AI Agent can determine the optimal buying or selling time by simulating various trading strategies. This learning method enables the Agent to continuously improve under evolving market conditions;

● Natural Language Processing (NLP): By understanding and processing user natural language input, the Agent can extract key information from governance proposals or market discussions to help users make better decisions. This is particularly important when scanning decentralized governance forums or processing user instructions.

3. Decision Making: Based on the collected data and the results of inference, the AI Agent enters the decision-making stage. In this stage, the Agent not only needs to analyze the current market situation but also to balance between multiple variables:

● Optimization Engine: The Agent uses an optimization engine to find the best execution plan under various conditions. For example, when engaging in liquidity provision or arbitrage strategies, the Agent must consider factors such as slippage, transaction costs, network latency, and fund size to find the optimal execution path;

● Multi-Agent System Collaboration: To deal with complex market conditions, a single Agent sometimes cannot comprehensively optimize all decisions. In such cases, multiple AI Agents can be deployed, each focusing on different task domains, to improve the overall system's decision-making efficiency through collaboration. For example, one Agent focuses on market analysis, while another Agent focuses on executing trading strategies.

4. Hosting and Execution: Since the AI Agent needs to handle intensive computation, it is often necessary to host its models on off-chain servers or distributed computing networks:

● Centralized Hosting: Some AI Agents may rely on centralized cloud computing services such as AWS to host their computation and storage needs. This approach helps ensure the efficient operation of models but also brings potential centralization risks;

● Decentralized Hosting: To mitigate centralization risks, some Agents utilize decentralized distributed computing networks (such as Akash) and distributed storage solutions (such as Arweave) to host models and data. These solutions ensure the decentralized operation of the model while providing data storage persistence;

● On-chain Interaction: Although the model itself is hosted off-chain, an AI Agent needs to interact with on-chain protocols to execute smart contract functions (such as transaction execution, liquidity management) and asset management. This requires secure key management and transaction signing mechanisms, such as MPC (Multi-Party Computation) wallets or smart contract wallets.

5. Interoperability: A key role of an AI Agent in the DeFi ecosystem is seamless interaction with multiple different DeFi protocols and platforms:

● API Integration: An Agent exchanges data and interacts with various decentralized exchanges, liquidity pools, and lending protocols through API bridges. This enables the Agent to have real-time access to market prices, counterparties, borrowing rates, and other key information to make trading decisions;

● Decentralized Messaging: To ensure the Agent's synchronization with on-chain protocols, the Agent can receive updates via decentralized messaging protocols (such as IPFS or Webhook). This allows the AI Agent to react in real-time to external events, such as governance proposal voting results, liquidity pool changes, adjusting its strategy accordingly.

6. Wallet Management: An AI Agent must be able to execute actual operations on the blockchain, all of which depend on its wallet and key management mechanism:

● MPC Wallet: A Multi-Party Computation wallet splits the private key among multiple participants, allowing the Agent to securely transact without a single point of key risk. For example, Coinbase Replit's wallet demonstrates how to utilize MPC for secure key management, enabling users to delegate partial autonomous operations to an AI Agent while retaining a certain level of control;

● TEE (Trusted Execution Environment): Another common key management approach is using TEE technology to store private keys in a secure hardware enclave. This approach enables the AI Agent to transact and make decisions in a fully autonomous environment without relying on third-party intervention. However, TEE currently faces challenges of hardware centralization and performance overhead, but once these issues are addressed, fully autonomous AI systems will become a possibility.

1.3 Denominational Source? From Intent to DeFAI

Image Source: Author's Own

If DeFAI's vision is: Through AI agents and various AI platforms, empower users to autonomously manage their investment portfolios, enabling everyone to easily participate in cryptocurrency market transactions, does this vision naturally lead us to the concept of "intent"?

Recall the concept of "intent" first proposed by Paradigm. In our usual transactions, we need to specify a clear execution path, such as swapping Token A for Token B on Uniswap, but in an intent-driven scenario, the execution path is matched by a solver and AI together and ultimately determined. In other words: Transaction = I specify the execution of the TX; Intent = I only care about the TX result and not the execution process. In hindsight, DeFAI's narrative not only closely aligns with the ultimate idea of an AI agent, fitting into AI perfectly, but also perfectly aligns with the vision of achieving intent. Taking a comprehensive view, DeFAI appears more like a new path for intent.

In the ultimate version of the future scenario of implementing blockchain at a large scale, will it be: AI Agent + Solver + Intent-Centric + DeFAI = Future?

2. DeFAI Related Projects

Image Source: Author's Own

2.1 Griffain

@griffaindotcom $GRIFFAIN: An innovative platform combining an AI agent with blockchain, helping users to issue an AI agent, focused on creating a powerful and scalable decentralized finance (DeFi) solution, supporting seamless token swaps, liquidity provision, and ecosystem growth. It enables easy wallet management, trading, NFTs, and automated execution of tasks like Memecoin issuance and airdrops.

2.2 Hey Anon

@HeyAnonai $ANON: A DeFi protocol driven by artificial intelligence, simplifying interactions, aggregating real-time project data, executing complex operations through natural language processing, and providing users with a DeFi abstraction layer. DWF Labs announced support for the DeFAI project Hey Anon through its AI Agent Fund and launched Moonshot on January 14.

2.3 Orbit

@orbitcryptoai $GRIFT: Simplifies the complex DeFi interface and operations, reducing the barrier to entry for the average person,

currently supports 100+ blockchains and 200+ protocols (EVM and Solana), with the GRIFT token used to inject vitality into the platform.

2.4 Neur

@neur_sh $NEUR: An open-source full-stack application that combines LLM models and blockchain technology features, designed specifically for the Solana ecosystem, enabling seamless protocol interactions using the Solana Agent Kit.

2.5 Modenetwork

@modenetwork $MODE: Positioned as the centralized platform for AI x DeFi innovation on Ethereum Layer2, where holders can stake MODE to receive veMODE and participate in AI agent airdrops, aiming to become the DeFAI Stack.

2.6 The Hive

@askthehive_ai $BUZZ: Built on Solana, integrates multiple models including OpenAI, Anthropic, XAI, Gemini, etc., to enable complex DeFi operations such as trading, staking, borrowing, and more.

2.7 Bankr

@bankrbot $BNKR: An AI-driven cryptocurrency companion where users can easily buy, sell, exchange, set limit orders, and manage wallets with just one message, planning to add token swaps and on-chain tracking soon with the vision of enabling everyone to use DeFi and achieve automated trading.

2.8 HotKeySwap

@HotKeySwap $HOTKEY: Offers an AI-driven DEX aggregator and analytics tool, cross-chain transactions, and a full suite of DeFi tools, supporting cross-chain transactions and analytics.

2.9 Gekko AI

@Gekko_Agent $GEKKO: An AI agent created by the Virtuals protocol, focusing on providing comprehensive automated trading solutions tailored for the prediction markets. GEKKO's automated trading strategies include auto-rebalancing, yield harvesting, and creating new token indices.

2.10 ASYM

@ASYM41b07 $ASYM: Offers an AI-driven DEX aggregator and analytics tool that identifies high-yield investment opportunities and settles the generated profits in $ASYM.

2.11 Wayfinder Foundation

@AIWayfinder $Wayfinder: An AI-powered all-chain interactive tool launched by the blockchain game chain game Parallel to help Agents navigate the on-chain environment, execute transactions, and interact with decentralized applications.

2.12 Slate

@slate_ceo $Slate: A universal AI agent and agent connectivity infrastructure layer that translates natural language commands into on-chain operations, focusing on executing automated trading strategies, buying or selling under specific conditions, making on-chain operations as simple as thought.

2.13 Cod3x

@Cod3xOrg $Cod3x: Solana AI hackathon project providing a no-code development tool to build agents for automating DeFi strategies, with its Agentic Interface being a tool that can execute complex operations using only intent expression.

2.14 Almanak

@Almanak__ $Almanak: An AI agent with self-learning capabilities that can autonomously perform tasks, using agent-based modeling to optimize DeFi and gaming projects, with a mission to maximize protocol profitability using data science and trading knowledge while ensuring its economic security.

2.15 HIERO

@HieroHQ $HTERM: A multi-chain intelligence tool for Solana and Base Network that allows users to use natural language commands for agents to autonomously perform transactions, including token swaps, simple token analysis, and more.

Three, What System Does the End of the AI Agent Belong to?

Image Source: Author's Creation

Time is in competition, and DeFAI projects are emerging one after another. After Bitcoin plummeted below $90,000 on January 13, the next day, according to CoinGecko data, DeFAI-related tokens defied the trend and rose by 38.73%, with $GRIFT, $BUZZ, and $ANON seeing the largest gains. However, we should consider which direction the AI Agent's financial focus should take. The current crossroads point to the left towards Gaming and to the right towards DeFi.

3.1 Leftward Game:

M3 (Metaverse Makers _) (@m3org) may be the most promising representative. The project is composed of artists and open-source hackers from an organization believed to be linked to ai16z, with core team members including JIN (@dankvr), Reneil (@reneil1337), Saori (@saori_xbt), Shaw (@shawmakesmagic), and others. However, the biggest real obstacle for the Game is in the manpower and resource-rich Web2 market, and there has not yet been a truly successful AI game. The highly anticipated "Phantom Beast Palu" in January 2024 sparked controversy over its superhuman development efficiency, leading to debates on whether AI design was used, but the CEO ultimately denied this claim. Furthermore, compared to the rightward DeFi, the AI Game seems to require more market enthusiasm due to the game's long development cycle.

3.2 Rightward DeFi:

The projects are ranked by market value as follows: $GRIFFAIN, $ANON, $OLAS, $GRIFT, $SPEC, $BUZZ, $RSS3, $SNAI, $GATSBY, with GRIFFAIN and ANON together accounting for 37.29% of the total DeFAI market value.

GRIFFAIN: Built on Solana, currently leading the DeFAI market value ranking with a $457M market value advantage and 103k Twitter followers. Its core functionality includes features such as atomic swaps through wallet generation and fast transactions. Currently, 0.01 SOL can be spent to mint NFTs using The Agent Engine.

Hey Anon: Using a multi-chain approach, currently supporting Sonic Insider, Solana, EVM, opBNB, and other different public chains, the sudden surge of $ANON is entirely driven by founder Daniele's (@danielesesta) halo effect. He is also the founder of Wonderland, Abracadabra, and WAGMI, and solely based on his traffic, Hey Anon has injected a lot of vitality into $ANON, ranking second with a $248M market value as his next entrepreneurial project.

4. Summary

The emergence of DeFAI is not accidental. The core feature adaptation of blockchain fits well into a strong financial scenario. Currently, whether it is the leftward GameFAI or the rightward DeFAI, both demonstrate considerable market potential. In the leftward Game direction, there may be a continuation of the metaverse in the future, with AI's help in managing virtual assets, characters, economies, and more. Drawing on the AI Agent's element-based play in breeding Meme in the metaverse, autonomous evolution of the metaverse's autonomy and prosperity can be achieved.

DeFi, advancing to the right, will inevitably transition from passionate hype to a value-oriented destination. The value of an AI Agent cannot rely on issuing memes to cater to market trends. However, the continuation of the AI Agent story must have the support of a DeFi-like yield farming mechanism. The victorious king will not always be clad in armor, and the ultimate result of market competition is worth our eager anticipation.

Original Article Link