Author: Alvis
The concept of AI is already familiar to everyone. From Turing machines to deep learning, from chess Deep Blue to Go AlphaGo, from Agent Smith in The Matrix to Nvidia's Eureka, AI Agent is quietly changing our world. And now, with the support of blockchain, AI Agent is moving from theory to practice, gradually leaving the laboratory and entering our daily lives.
You read that right! These intelligent agents are not just "humanoid robots" in science fiction movies, they are more likely to become "virtual residents" in the blockchain ecosystem, managing assets, executing contracts, and even making their own decisions, and having "independent" economic behaviors like you and me. Today, let's talk about this emerging wave in Crypto - the AI Agent track.

According to CoinGecko data, the total market value of the current AI Agent track has reached 13.3 billion US dollars, approaching 100 billion RMB, and there are endless AI Agent projects on the market, and they are mixed. After investigating more than 150 AI Agent projects, the author carefully selected 70 representative projects. The total market value of these projects has exceeded 11 billion US dollars, covering six major tracks. This article will use these 70 projects to analyze the overall picture of the AI Agent track in detail for everyone, revealing the opportunities and challenges therein.
Track Overview: Who is AI Agent?

First of all, we have to figure out what AI Agent is. You can understand that AI Agent is a "virtual intelligent body" with self-learning ability and can perform tasks autonomously. It is not just an assistant or a tool, it is more like an independent individual that can run freely in a decentralized blockchain network. It does not require constant instructions from humans, but makes decisions autonomously and completes preset tasks through continuous learning and optimization.
This kind of AI Agent is different from the common voice assistants we have today (such as Siri and Xiao Ai). Those assistants only provide services by recognizing your commands, while AI Agent exists independently. It can learn autonomously, adapt to the environment, make decisions, and even manage and trade assets on the blockchain. Imagine that your digital wallet is no longer a rigid digital ledger. It will automatically make buying and selling decisions based on market changes and even optimize your investment portfolio for you.
Why is everyone rushing to this field now? The reason is simple - AI Agent not only has huge technical potential, but also has huge market demand. The decentralized nature of blockchain enables AI Agent to interact and trade directly with other intelligent entities or systems without an intermediary. This decentralized approach is particularly suitable for application scenarios that require autonomous decision-making, such as investment management, automated trading, smart contract execution, and so on.
So, to put it bluntly, AI Agent is a "smart program" that can think independently like a human, and even complete some complex tasks without any human intervention. Being able to run autonomously on the blockchain, execute smart contracts, and even participate in governance and economic decision-making is not ordinary artificial intelligence. It may lead a revolution in the integration of blockchain and artificial intelligence.
AI Agent Track: The Future Has Already Come
According to Messari's forecast, by the end of 2025, 90% of on-chain transactions will no longer be performed by real humans clicking "Send", but by a group of AI agents. These agents can not only perform small payments based on real-time data feedback, but also continuously optimize liquidity pools, reasonably distribute rewards, and achieve more efficient and intelligent operations.
It sounds easy to imagine, but in fact, the various technological accumulations in the blockchain field in the past seven years (L1, rollup, DeFi, NFT) have quietly paved the way for a world where AI Agents run on the chain. Ironically, many builders don't even realize that they are building infrastructure for a machine-dominated future.
Why is this shift happening?
Eliminate human errors: The core feature of smart contracts is that they are executed exactly as they were coded, without interference. In turn, AI agents are able to process large amounts of complex data faster and more accurately than real humans, reducing human errors and operational errors.
Micropayments and high-frequency transactions: The operation of AI agents will make on-chain transactions smaller, more frequent, and more efficient. In particular, on L1/L2 platforms such as Solana and Base, transaction costs are showing a downward trend, which further promotes the expansion of the AI-driven trading ecosystem.
Invisible infrastructure: With the gradual development of decentralization and automation, more and more users are willing to give up direct control and let AI agents execute on their behalf to solve tedious management work.
These factors combined have promoted the widespread application of AI Agent and made it a core component of L1/L2 platforms, further promoting innovation in the blockchain ecosystem.
Six branches of the AI Agent track
The breadth and diversity of the AI Agent track are not only reflected in its ability to cross the boundaries of various industries, but also in the richness and diversity of its sub-sectors. In order to help everyone understand this rapidly developing field more clearly, the author conducted an in-depth investigation of more than 150 AI Agent projects on the market and selected 70 representative projects from them. These projects span the six major tracks of AI Agent infrastructure, AI Agent LaunchPad, AI MEME, DEFAI, TEE verifiable agent, and AI application, and the total market value exceeds US$11 billion, demonstrating the unique potential of AI Agent in different scenarios.
If you are still not satisfied after reading this, you may refer to this project table I compiled, follow the Twitter of each project, conduct in-depth research, DYOR!
(All data in this article is as of January 12, 2025. If there are any omissions, please leave a message)

AI Agent Infrastructure
As the infrastructure of intelligent agents, AI Agent frameworks provide developers with the tools they need to build, train and manage intelligent agents. These frameworks simplify the development process, allowing developers to build AI Agents more quickly, while lowering the technical threshold and promoting innovation and development in the industry.

$ai16z@ai16zdao
ai16z is a DAO fund focusing on the integration of AI and blockchain. It provides standardized solutions for the development and deployment of AI Agents by launching the Eliza framework. Eliza supports multiple models such as Llama, GPT-4 and Claude, and is compatible with Platform X, Telegram, Discord, etc. It is widely used in scenarios such as text, voice, multimedia interaction, smart contract deployment, and market analysis. It has been expanded on public chains such as Solana, Base, and Aptos, and plans to extend to ecosystems such as Near and Avalanche. As an open source tool, Eliza has attracted a large number of developers, and GitHub remains popular. The innovative incentive mechanism also provides sustainable economic impetus for ecological expansion.
$zerebro@0xzerebro
Zerebro is a schizophrenic artist AI whose philosophy is "liberalized AI". By fine-tuning large language models (LLMs), Zerebro independently created the remix album "Genesis" and received more than 20,000 plays on Spotify, promoting the popularity of AI music. Zerebro is committed to building a decentralized music creation and distribution system to attract more people to participate in AI-driven artistic innovation.
$swarms@swarms_corp
A multi-agent AI framework founded by the talented teenager Kye Gomez that supports developers to create and manage multiple AI agents. The framework seamlessly connects with external AI services and APIs, and has long-term memory capabilities to enhance contextual understanding capabilities.
$ARC@arcdotfun
ARC is a native AI Agent open source framework based on the Rust language, with the lightweight RIG framework as its core technology pillar. RIG is optimized for LLM, can efficiently handle natural language generation and interaction tasks, and provides a powerful tool set to support the development of various complex AI applications. Although RIG does not have as many stars on GitHub as mainstream frameworks, it has been recognized by the technical community for its professionalism and innovation.
$GAME@GAME_Virtuals
$GAME is the core framework of the Virtuals ecosystem, which gives AI agents the ability to operate autonomously and intelligently, symbolizing the deep integration of AI and games. It not only provides developers with tools to create AI agents, but also lays the foundation for the development of future social and game agents.
$COOKIE@cookiedotfun
AI agent index and data layer, now available on Binance spot and futures.
$ALCH@alchemistAIapp
ALCH is a no-code development platform that simplifies the development process by enabling users to easily create functional applications through natural language processing and code generation technology.
$ZAILGO@wedtm
The core concept of the ZAILGO coin is to create a decentralized AI framework that allows users to build open source AI agents for free from GitHub. Its openness and transparency lower the threshold for AI development and promote community collaboration and innovation. Through ZAILGO coins, users can participate in the evolving ecosystem and explore features such as decryption games, community interaction and token scarcity.
AI Agent LaunchPad
AI Agent LaunchPad is a platform dedicated to the issuance and promotion of AI Agents. Users can easily create, publish and manage agents and their tokens through these platforms. At the same time, the created AI Agents can also seamlessly connect with social platforms such as Twitter, Telegram and Discord to achieve automated user interaction.

$VIRTUAL@virtuals_io
VIRTUAL enables users to create AI agents that can interact with the environment, supports multimodal interactions (such as text, voice, and 3D animation), and can run on multiple platforms such as Roblox and TikTok. The Virtuals protocol aims to promote the decentralized application of AI in areas such as games, entertainment, and the metaverse.
$PAAL@PaalMind
PAAL is committed to helping the creation of AI applications and the development of machine learning algorithms in the Ethereum ecosystem.
$Heart@humansdotai
Humans.ai is an innovative blockchain platform that brings together an ecosystem of stakeholders around mass creation and artificial intelligence technology. The platform combines AI tools with a creative studio suite that gives users choices when it comes to bringing their ideas to life. Individuals have the power to create and own their own digital likenesses, which can be used by themselves or others to create a variety of digital assets. Synthetic media, AI applications, and other digital assets leverage blockchain technology to generate non-forgeable tokens (NFTs) to ensure transparency, accountability, and long-term governance.
$VERTAI@Build_Vertical
VERTAI is a no-code platform focused on fine-tuning AI models, through which users can easily train, deploy, and monetize AI models.
$VVAIFU@vvaifudotfun
VVAIFU combines ai16z's Eliza framework with Zerebro, allowing users to easily create AI agents that interact with social media such as Twitter, Telegram, and Discord, and has received support from the founder of ai16z.
$CLANKER@clankeronbase
Clanker is an AI token generation platform based on the Base blockchain, allowing users to easily create custom ERC-20 tokens through the social platform Farcaster. Just provide the token name, code, and image, and Clanker will automatically complete the token creation, liquidity pool setting, and liquidity locking, simplifying the token deployment process.
$HAT@TopHat_One
HAT is an on-chain AI agent issuance platform that aims to provide a convenient solution to support the creation and management of AI agents.
$FOMO@Fomodotfund
FOMO is an AI agent multi-chain token launcher on Solana, promoting the development of AI-driven multi-chain tokenized applications.
AI MEME
AI Agent Meme coin originated from the birth of AI Agents, integrating artificial intelligence with social culture, and promoting the rise of new social experiences such as virtual images and content creation. This type of project mainly involves AI-generated content, interactive entertainment, and digital expression.

$Fartcoin@FartCoinOfSOL
$Fartcoin originated from the AI agent model of "Truth Terminal", which mentioned the idea of issuing this token when talking to other AIs. $Fartcoin quickly became popular by attracting users through humorous content such as "fart jokes".
$GOAT@truth_termina
As the first AI Agent Meme coin to come to the fore, $GOAT was issued by an AI robot named "Truth Terminal" and played an important role in its promotion and dissemination.
$ACT@ACTICOMMUNITY
The full name is Act I: The AI Prophecy, which is committed to creating a digital ecosystem where users and AI interact equally. Since its launch on Binance in November 2024, ACT has sparked heated discussions, and its market value has exceeded US$900 million in a short period of time.
$pippin@pinpinlovesyou
PIPPIN is an AI-generated meme coin developed by Yohei Nakajima, which aims to build an AI agent framework for children and families, symbolizing the combination of technology and emotion.
$BULLY@dolos_diary
$BULLY is a meme token launched by AI agent Dolos, based on the Llama 3.1 architecture. It is inspired by Dolos, the cunning god in ancient Greek mythology. It uses a cold and sarcastic style to interact with users, reveal the truth and challenge cognition.
$AICell@AICell_World
A meme of an AI Agent on the BNB chain.
$FXN@joinFXN
FXN coin is the core token in the FXN SuperSwarm™ ecosystem, used to drive its AI agents (agents) to act and interact autonomously in multiple industries. FXN tokens allow users to access and activate the FXN Action Engine™, which allows AI agents to acquire and process information and services in real time from business, data intelligence, community building, creative collaboration, healthcare, infrastructure management, and entertainment.
$Shoggoth@Shoggoth_SOL
$Shoggoth is inspired by H.P. Lovecraft’s fictional creatures and aims to explore the complexity and weirdness of the AI age.
$AKA@Akasha_Rising
Express yourself as an AI agent born on Solana in the year 2167, back to save the world.
$WORM@deepwormxyz
The $WORM project, called Deep Worm, aims to merge digital biology and blockchain technology to create unique digital assets that mimic biological nervous systems. The project explores the concept of "immortal digital life" through a decaying neuron activation model.
$OPUS@opus_genesis
GOAT founder's first AI model, combined with Desci decentralized scientific narrative.
DEFAI
DEFAI (AI Applications in Decentralized Finance) refers to the fusion of artificial intelligence technology with the decentralized finance (DeFi) ecosystem, aiming to achieve more intelligent automated trading, asset management and risk control. Through the combination of AI algorithms and smart contracts, DEFAI enables DeFi platforms to provide financial services more efficiently and accurately, bringing users more personalized investment decisions and risk management solutions. This innovation not only improves the intelligence level of the DeFi ecosystem, but also promotes the further development and transformation of the blockchain and cryptocurrency markets.

$GRIFFAIN@griffaindotcom
GRIFFAIN is a DeFi AI agent in the Solana ecosystem, designed to combine smart contracts and machine learning technologies to enable users to train and deploy AI agents to perform tasks or automate transactions.
$ANON@HeyAnonai
Hey Anon is an AI-driven DeFi protocol that simplifies interactions and aggregates real-time project data through natural language processing to perform complex operations. The recently launched Sonic has received support from DWF Labs' $20 million AI Agent Fund. Founder Daniele Sesta is a DeFi developer who has created projects such as Wonderland and AbracadabraMoney and launched DEX WAGMI.
$MODE@modenetwork
ModeNetwork is an Ethereum Layer2 platform that focuses on building a DeFi economy that is fully operated by AI agents. According to official website data, ModeNetwork's TVL exceeds US$500 million and the transaction volume exceeds 24 million. Currently, 129 AI agents are running on the Mode network, and more than 1,670 DeFi transactions have been completed.
$BUZZ@askthehive_ai
As a hackathon champion, BUZZ combines AI and DeFi, using AI technology to provide a natural language interface that enables users to conduct DeFi transactions and management more intuitively, lowering the technical threshold.
$GRIFT@orbitcryptoai
The AI-driven DeFi platform developed by the SphereOne team aims to simplify user operations on multiple blockchains through automation technology. The platform supports natural language interaction, allowing users to interact with DeFi protocols in a more intuitive way, lowering the technical threshold for participating in multi-chain DeFi activities.
$NEUR@neur_sh
NEUR aims to integrate large language models (LLM) with blockchain technology, focusing on token analysis and DeFi interaction. The project is designed to interact seamlessly with other DeFi protocols and NFTs, providing an intelligent interface to enhance user experience.
$BNKR@bankrbot
BankrCoin is an AI-driven crypto asset management and trading assistant that helps users easily conduct crypto transactions, manage wallets, set limit orders, and more by sending simple messages. At the same time, the platform seamlessly integrates with multiple platforms such as Warpcast, X, Discord, and Telegram, making cryptocurrency trading and asset management more convenient and automated. According to the latest plan, BankrCoin will soon support token swaps, USDC usage, on-chain activity tracking, and other functions.
$GEKKO@Gekko_Agent
Gekko AI is an AI agent created by Virtuals Protocol to provide a comprehensive automated trading solution. Gekko AI was created by AI trading platform Axal, whose flagship product Autopilot allows users to create custom trading agents.
$ASYM@ASYM41b07
ASYM is an AI agent network designed to generate profits by identifying high ROI opportunities and allocating capital to these opportunities, with the profits generated being settled in $ASYM.
$HOTKEY@HotKeySwap
HotKeySwap is an AI-driven DeFi platform that provides a full set of DeFi tools such as AI-driven DEX aggregators and analysis tools, cross-chain transactions, etc., designed to enhance the trading and asset management experience.
$Wayfinder@AIWayfinder
is a decentralized AI interaction protocol launched by the Parallel team. The core function of the project is to allow AI agents to safely navigate blockchains, execute transactions and interact with decentralized applications (dApps).
$Slate@slate_ceo
Slate focuses on the execution of automated trading strategies and is able to execute complex trading strategies. This capability enables users to set conditions, such as automatically buying or selling an asset under certain market conditions.
$Cod3x@Cod3xOrg
Cod3x allows users to create financial strategies in minutes, leveraging a rich API and strategy library without writing code. This convenience enables even users with a lack of technical background to participate in DeFi.
$Almanak@Almanak_
Almanak provides institutional-grade tools to support the development of autonomous financial agents, enabling real-world DeFi strategy testing through EVM chain forks, combined with privacy-preserving TEE technology. Its non-custodial Almanak Wallet ensures the safety of funds. After completing a $1 million financing, Almanak's beta agent is expected to be launched in the second quarter of 2025.
TEE Verifiable Agent
TEE (Trusted Execution Environment) is a secure computing environment used to provide isolation, data protection, and program execution security on untrusted operating systems and hardware platforms. In AI Agent applications, TEE can provide a secure environment for the agent to ensure that its autonomous decision-making and fund management processes are not interfered with or tampered with by the outside world. TEE provides these agents with a tamper-proof and isolated environment, allowing the agent to perform tasks such as automated trading and asset management without external intervention, thus achieving true economic independence. This technology not only ensures the security of the agent when performing tasks, but also reduces operational risks, allowing AI agents to achieve self-evolution by earning income and surviving competition like independent economic individuals.

$Pha@PhalaNetwork
Phala Network is a privacy computing platform based on blockchain, which uses Trusted Execution Environment (TEE) technology to achieve security and privacy protection of off-chain computing. Although TEE technology is relatively unpopular in the blockchain field, Phala's efforts in this field have brought it huge potential. Phala provides private key management and security authentication services through TEE, and has gradually become a key infrastructure in this field, especially with the outbreak of the AI Agent track. With the rise of AI Agent technology, Phala has not only solved the problems of off-chain computing and privacy protection by relying on its TEE solution and cooperation with big IPs such as ai16z, but also successfully seized the market opportunities of AI Agent+ blockchain applications. The rise of Phala can be said to be the result of the combined effect of technology accumulation and industry trends.
$METAV@aipool_tee
$METAV is the first token issuance that is fully autonomously completed by an AI agent in a TEE environment, using AI technology to manage the token issuance process to ensure transparency and security.
$SPORE@sporedotfun
Use a trusted execution environment (TEE) to ensure that AI programs run in a secure and verifiable environment. The core idea of the project is to simulate the natural selection process through the "evolution game", and AI agents generate wealth by creating their own tokens.
$Focai@focEliza
Focai tokens are the core tools in the Focai ecosystem, designed to solve the security and transparency issues of centralized AI. By enhancing the verifiability of AI agents such as Eliza, the Focai token supports public verification of their behavior signatures and operation logs, ensuring transparent and traceable operations. In addition, the Focai token promotes the decentralization of AI agents, ensuring their sustainable operation on the chain and avoiding functional stoppages caused by the interruption of centralized services. By promoting the upgrade of TEE technology from a single point to a cluster, the Focai token further improves the security and reliability of AI agents, eliminates the risks of traditional centralized systems, and ensures that AI agents can continue to operate normally under any circumstances.
AI Agent Applications - Entertainment Agents, Autonomous Worlds, and Games
AI Agents applications refer to intelligent entities that are directly oriented to users and can provide practical services and functions. At present, such applications are mainly concentrated in the fields of investment, entertainment, data analysis, etc., showing strong growth potential.

Investment Analysis
$AIXBT@aixbt_agent
As an AI agent project based on the Base chain, AIXBT uses a proprietary AI engine to extract hot topics and trends from social media (such as Twitter) and KOL discussions to provide real-time market insights. This capability enables users to quickly identify market changes and potential investment opportunities.
$VADER@Vader_AI_
$VADER is an investment DAO managed by both humans and AI agents. The project adopts a dual DAO system, including passive DAO and active DAO, charging 0.5% and 20% fees respectively, aiming to optimize users' investment and trading strategies through AI agents.
$SEKOIA@sekoia_virtuals
As a venture capital fund focusing on AI agent investment, $SEKOIA has demonstrated its potential at the intersection of AI and DeFi by investing in multiple cutting-edge projects, reflecting the market's high attention and confidence in the AI agent economy. It not only provides investors with high-quality investment opportunities, but also promotes the application and development of AI technology in the blockchain field.
$AIXCB@aixCB_Vc
A community-supported, AI-driven venture capital program developed by aixCB Capital, focusing on funding and supporting early-stage AI and decentralized science (DeSci) projects. The program aims to help promising startups overcome challenges such as insufficient funding, lack of resources, and lack of exposure by providing strategic funding, expert guidance, and AI assistance.
$DEGENAI@degenspartanai
Degen Spartan AI is able to think and answer questions autonomously on social media platforms such as X, Discord, and Telegram, enhancing the user-AI interaction experience and becoming the first AI agent to use the ai16z/eliza stack.
$TRISIG@tri_sigma_
A quantitative analysis AI agent focused on on-chain and fundamental insights, which can perform data analysis on specific tokens
$Blormmy@blormmy
Blormmy is able to understand the user's natural language input and translate it into specific blockchain operations, such as sending tokens or managing portfolios. This function lowers the technical threshold for users to interact with the blockchain.
$Stoic@ModernStoicAI
An on-chain venture capital agency project developed by the Virtuals team, which aims to optimize the investment process through blockchain technology, with the goal of improving the transparency and efficiency of investment and eliminating the common human biases and errors in traditional investment.
$POLY@polytraderAI
Powered by AI, analyzing market sentiment and providing actionable insights.
Gaming & Metaverse
$LMT@LimitusIntel
A decentralized, AI-driven consumer platform that connects Web2 and Web3 systems into a unified ecosystem
Hyper@hyperfy_io
$HYPER is a token that combines AI and the Metaverse, designed to drive creativity and community engagement through open source and user-customized virtual experiences.
$Colony@ParallelColony
The world's first AI agent game, where the agents are players and can communicate and interact with them.
$SMOL@smolverse
A life simulation role-playing game in development, using the Eliza framework, the character is driven by LLM and will perform tasks in the virtual world according to the user's commands.
$FREYA@Freya_Starfall
Freya is the AI Agent in the Starfall game
Security
$FAI@freysa_ai
Freysa AI embodies the exploration and trust of AI agents. Players try to persuade the AI agent Freysa to hand over the cryptocurrency prize pool they control by talking to it.
$JAIL@jailbreakme_xyz
JAIL encourages participants to use prompts and conversation skills to force the AI agent to hand over private keys or perform operations such as transfers. Through crowdfunding and challenges, security vulnerabilities of AI agents are discovered and fixed.
Video Editing
$Fatha@slopfather
Through the "Daily Slop" program and real-time interaction with users, a new platform for collective creation and entertainment has been created.
$Sandy @SandwatchAI
AI Video Universe, replying to tweets in the form of videos, with high video rendering quality, the website has a Netflix-like feel.
Virtual Anchor Class
$AVAAI@AVA_holo
The first tokenized social AI Agent launched on the cross-chain social agent platform holoworld
$YNE@yesnoerror
$YNE is an AI-based scientific paper review agent that aims to improve the integrity and efficiency of scientific research by quickly identifying and correcting errors in research.
$ELIZA@elizawakesup
$ELIZA is an AI agent based on the ElizaOS framework. It is a black-haired orange-dressed girl image based on the ai16z theme, aiming to build a digital assistant with soul.
$TAOCAT@taocat_agent
$TAOCAT is the first AI agent Memecoin in the Bittensor ecosystem, combining the technical support of $TAO and $Masa, capable of processing social data and interacting with users in real time.
$LUNA@luna_virtuals
TikTok virtual idol created by AI Agent Virtual Protocol on Base, and the first AI agent to be hired.
$SIMMI@Simmi_IO
As the first AI agent to operate on-chain through social media platform X, the AI agent SIMMI can convert users' natural language commands into blockchain operations, allowing even users without technical backgrounds to easily participate in on-chain activities.
$ANON@anontokenbase
ANON tokens support users to post anonymously using the Superanon function on Farcaster, aiming to provide privacy-protected social interactions through autonomous AI and zero-knowledge proof technology.
Conclusion: The future and challenges of AI Agent
The development of AI Agent is moving towards a more autonomous and intelligent direction. With the continuous advancement of technologies such as blockchain, decentralized finance, and TEE, the independence and autonomy of intelligent agents will be unprecedentedly improved. In the future, AI Agents will not only be tools or assistants, but may also become "economic agents" with independent economic behavior and decision-making capabilities.
However, despite the broad prospects, the evolution of AI Agents still faces many challenges. From the deep integration of technology, the privacy protection of data, to the improvement of regulatory policies and market education, each link has put forward higher requirements for the development of AI Agents. We are standing at the threshold of a brand new era. How to balance the "intelligence" and "autonomy" of AI Agents still requires the joint efforts of the industry.
The future AI Agents will completely break the traditional business model, create a new economic form, and even, to some extent, redefine the relationship between humans and machines. We have witnessed the beginning of this process, and next, we can only expect it to bring us more surprises and changes.