Author: Shu Fen

Many people say that this round of bull market in the crypto world lacks innovative narratives, but in fact, AI is the most innovative and enduring core narrative. By December 2024, the projects with the highest returns in the entire cryptocurrency market (non-chain) will come from the AI field—Virtuals, with a return of up to 23,079%.

"The next stop for large models," "completely changing human lifestyles," "opening a new industrial revolution era"... People do not hesitate to describe the importance of AI Agents. Regarding the current momentum and future trends of AI Agents, both retail investors and institutions are unprepared in investment cognition. Many people around me did not pay attention before and only felt the need to understand it after the outbreak, but now the market is flooded with various information, making it overwhelming and causing a lot of confusion. Today, I will completely sort out AI Agents, and this research report will help you quickly gain knowledge, serving as a beginner's guide to AI Agents (crypto version)!

This article will be divided into three parts: first, an overview of the current state of AI Agent development; second, the selection and analysis of potential AI Agent projects; third, the application expectations of AI Agents in the Web3 field.

1. Basic Literacy of AI Agents

AI Agents first came to the public's attention in March 2023 when a project called AutoGPT was released, utilizing large language models to automatically decompose a large task into smaller tasks and use tools to complete them.

AutoGPT shocked the world upon its release because it was the first to extend language processing, content creation, logical reasoning abilities, and perception-action technology to application scenarios. Soon after, OpenAI launched a series of GPTs, and then many tech companies began to layout their application layers, platform layers, development layers, and operational layers based on their respective capabilities, to increase the barriers in the next wave of ecology.

What exactly is an AI agent? How does it work? 'Agent' means representative in Chinese. In simple terms, an AI agent is a representative empowered by AI technology. Unlike traditional software, which passively executes commands, its workflow is: perception module (input acquisition) → LLM (understanding, reasoning, and planning) → tool invocation (task execution) → feedback and optimization (verification and adjustment).

OpenAI defines the "AI Agent" as a system driven by LLM that has the ability to autonomously understand, perceive, plan, remember, and use tools, capable of automating complex tasks. Unlike traditional artificial intelligence, AI Agents possess the ability to complete given objectives through independent thinking and tool invocation.

To illustrate a point, consider this: if you have a cold and fever, traditional software would just tell you to go to the hospital and be careful. If it's an AI agent, it can check your temperature and other health indicators, match you with appropriate medication based on online information, request payment, and deliver it to your home, even writing your leave request for the next day. This is the magic of AI agents.

2. AI Agent Projects and Analysis

According to the latest data from Cookie.fun, as of December 30, the overall market value of AI Agents has reached $11.68 billion, with a nearly 39.1% increase over the past 7 days. This growth trend indicates the rapid growth of the AI Agent ecology in the crypto market.

In this wave of AI Agent enthusiasm, ai16z and Virtuals Protocol are undoubtedly the two most powerful representative projects. Specifically, Virtuals' ecological market value has reached $5.01 billion, while ai16z stands at $1.63 billion, together accounting for 56.8% of the AI Agent market share, contributing over half of the total share.

From the on-chain distribution perspective, Base and Solana are the two main battlefields for AI Agents. Among them, the market value of AI Agents on Base is approximately $5.76 billion, while the market value on Solana is $5.47 billion, together contributing to 96.1% of the overall market, while other on-chain projects cumulatively only amount to $920 million.

This also indirectly reflects that although the AI Agent ecology is rapidly rising in the crypto market, attracting a lot of attention and capital, the market structure remains singular, mainly relying on a few leading projects. The AI Agent ecology is still in its early budding stage.

Now, let's analyze the current hot ecological projects of AI Agents in light of the current market situation. The standards mainly consist of three points: 1. Long-term value of the project 2. Authentic market demand 3. Cash flow income situation. If after reading, you wonder why the XXX project is not included, please review it again based on these three standards. The views on the following projects are for reference only and are not financial advice.

1.Virtuals

Virtuals was actually launched last year. The Virtuals protocol mainly establishes co-management for AI agents in the gaming and entertainment sectors, allowing AI agents to be tokenized through blockchain and achieve co-management. The agent functions include autonomous planning, goal achievement, environmental interaction, and on-chain wallet control.

The biggest innovation and difference of Virtuals and other Web3 AI agent protocols is that it simplifies the complexity of AI agents, providing a plug-and-play solution similar to Shopify, enabling non-AI professionals to easily deploy AI agents in gaming and consumer applications, and obtain protocol income through tokenization and decentralized co-management.

Additionally, Virtuals has generated AI virtual idols—AI-dol bands using its AI technology principles, attracting hundreds of thousands of followers on TikTok, which is quite interesting.

The total supply of Virtuals tokens is 1 billion, and all have been released. The token distribution is as follows: 60% held by the public, 5% for the liquidity pool, and 35% for the ecosystem treasury, with a maximum annual release of 10% over three years, and currently over 30% of the ecosystem fund's tokens.

From the perspective of the project's long-term value, it addresses the pain point that non-AI professionals cannot participate in the AI boom. It has a certain user base, and its token economics are relatively open and transparent, with excellent marketing. In terms of market value, as the leader of the AI Agent ecology, this rise has seen almost no correction, so there is a high probability of a significant adjustment in the future, making the short-term outlook for Virtuals relatively risky.

2.ai16z

Although ai16z has the same name as the well-known venture capital a16z, this project has no relation to A16Z and has not received investment from A16Z; the only connection is that it has attracted the attention of a16z founder Marc Andreessen.

The total token supply is 1.09 billion. The project operates as a DAO. According to its core influencer Shaw, ai16z will launch several games based on the Eliza framework in the future, and will focus more on creating a practical AI Agent investment tool, DeFi AI Agent. The founder stated that ai16z's goal is not to create an AI robot that mimics a16z but to outperform it in the investment domain where it excels.

Let me talk about the AI16Z project, which focuses on investment model AI agents. It seems no different from previous AI bots or Telegram bots. Can AI really make investment profits? That raises a question mark. The core technology, Eliza OS, is merely a simple development based on OpenAI's capabilities, so how will Shaw respond when OpenAI opens its own AI agents?

In summary, I believe AI16Z is merely riding on the coattails of AI16Z. The long-term value of its project lies in DeFi AI Agents, but this demand is a false demand, returning to the previous logic of AI bots, and its technical development capabilities rely on the open-source database of OpenAI, with average imagination.

3.SWARMS

Swarms is a multi-agent LLM framework for AI agents, providing a large number of cluster architectures and seamless third-party integrations. It currently enables enterprises to easily build and manage collaborations between multiple AI agents, and under the scheduling of Swarms, they can seamlessly cooperate to complete complex business tasks. In simple terms, SWARMS' users are businesses in the B2B sector, providing enterprise-level AI agent applications.

Its founder, Dev, is Kye Gomez, who is only 20 years old (source: the internet). He publicly claimed that OpenAI infringed on the team's intellectual property, stole the project name, and plagiarized the code structure and methods. Subsequently, Gomez provided a more detailed explanation: Swarms is a multi-agent framework that has been running for nearly three years. To date, more than 45 million agents are running in production environments, serving some of the world's largest financial, insurance, and healthcare institutions.

After Swarms tokens were issued on December 18, they quickly surged to a market value peak of $74.2 million on the 21st. Unfortunately, the good times did not last long, as the market value plummeted like a roller coaster to the bottom, leaving only about $6 million. Subsequently, it fluctuated around $13 million until the 27th when it began to rebound, rising from a low of $12 million to $30 million, and then surging nearly 3 times to close to $70 million, almost breaking through the previous high.

Compared to the whimsical castles in the air of the AI16Z system, if Swarms is indeed created by the 20-year-old AI genius Kye Gomez as rumored, then undoubtedly Swarms has strong technological barriers. Its official website has already provided efficient solutions to numerous enterprises, and its strength is evident.

As an open-source project, Swarms has sparked enthusiastic attention in the developer community, with the number of Stars on GitHub surpassing 2.1K, receiving the wisdom and support of numerous developers. Thus, all that Swarms has accumulated confirms the maturity and innovation of its technology. Swarms has stronger technological capabilities and strong market demand (enterprise-level), and will stand out in the competition of AI agents.

4.GRIFFAIN

Griffain is a project based on Solana—an artificial intelligence agent engine, similar to Copilot and Perplexity. It is one of the closest projects to Agentic APP. The ultimate form of the search engine in the AI era should be that users directly propose demands, and AI directly provides results or solutions, rather than just offering web links. One of the catalysts for this project is its open-access mechanism. As a leading agent engine, Griffain undoubtedly attracts a lot of market heat.

Currently, Solana is the blockchain with the most AI agents. In October, Goat, as an AI agent, financed humans through pumpfun. From a certain perspective, this is an AI singularity, due to the excellent liquidity and well-established AI agent developer community on the entire chain, calling Solana the most imaginative breeding ground for AI Agents is not an exaggeration.

The most important thing Sol has done is to revitalize the ecology with Griffain. To bring about a real 'Agentic app szn', what is needed besides AI is the infrastructure channel. Although GriffAIn has yet to clarify the specific application scenarios of its token, in the future, GriffAIn will connect the demand side with the Solana ecology, and as long as it exists within the basic technological framework of Solana, it can meet most conditions, whether it is targeting certain compliant tokens on Pumpfun or creating tokens. This vision has gained recognition from Toly, adding a lot of imagination to its GriffAIn prospects.

5.AIXBT

Aixbt is one of the agents created by Virtuals based on the Base platform. It monitors Crypto Twitter and market trends through smart analysis tools, providing users with valuable market insights. Some analysis content will be shared on Twitter, while the rest is limited to token holders, who can interact directly with the agent through their exclusive terminal.

Aixbt's analysis has a certain accuracy in predicting price trends, demonstrating how AI can parse blockchain data and assist traders in making more informed decisions across multiple platforms and fields.

I checked the relevant content published by Aixbt, and my intuitive feeling is that the content is very rich, covering almost all tracks, and easily accessing various data. Also, there may be some potential investment opportunities in the short-term cryptocurrency splits; for example, discovering that vapor on hype is relatively undervalued compared to similar AI launchpads. Data shows that among the 210 tokens recommended, 183 achieved profits after being recommended by aixbt, with a profit ratio of up to 83%.

However, there are some shortcomings, such as being unable to fully decompose complex items, and the analysis and data are still superficial, unable to indicate the risks of investment opportunities. However, I think it is much stronger than some current crypto KOLs.

From the perspective of the project's long-term value, Aixbt has a demand in niche areas, and users are motivated to hold tokens to unlock more information data and price analysis. With Aixbt's continuous data feeding and evolution, I believe Aixbt is the absolute king of market predictive AI agents.

Based on the analysis of the current market, I believe the five projects with higher popularity, ranked from high to low in terms of market value imagination, are: SWARMS, GRIFFAIN, Virtuals, AIXBT, Ai16z.

3. Future Development Trends of AI Agents in the Web3 Field

As for the application of AI Agents in the Web3 field, there are several noteworthy directions that also represent future trends. One is privacy and security; AI should be designed with respect and protection for users and society as the fundamental principle. However, as AI becomes increasingly knowledgeable about us, privacy becomes more blurred and fragile. Every interaction with smart devices and every piece of personal information input becomes food for AI's evolution.

The importance of privacy issues is closely linked to security issues. Systems that store and process personal data, once targeted by hackers, can lead to issues like information leakage, identity theft, and asset loss. Is there an environment that can leverage AI's powerful capabilities while protecting personal privacy? Clearly, in the Web3 field, compared to traditional methods, it can provide users with a higher level of data protection, perfectly balancing AI development capabilities with privacy protection.

Therefore, we see that many large models are beginning to experiment with data storage on the blockchain. The perfect AI soil of Web3 has attracted many AI developers in privacy-sensitive industries such as healthcare and finance to use blockchain technology to ensure data security and privacy.

Another important direction is computing power and data. AI Agents, especially Multi-Agents, face cost issues in development, training, and operation. For enterprises, training an AI Agent requires a large amount of computing power, often hundreds or even thousands of high-performance GPUs or specialized hardware like TPUs. The cost of acquiring and using these computing resources is already high; for example, Stanford's virtual town includes 25 agents in its multi-agent research. However, after the town framework was open-sourced, testing one agent required a data source costing $20,000 per day.

In Web3, through reasonable token economics and user incentive programs, idle computing power or personal data sets can be reallocated, further reducing computing and data costs, allowing more individual users to participate in the construction of the AI industry. For example, some data platforms allow users to monetize their own data, providing low-cost data sources for AI Agents.

In the end, I believe that AI Agents can serve as new infrastructure for Web3, deeply integrating with other core elements to spawn entirely new application models, rather than being a simple AI Memecoin. Currently, in the Web3 field, AI Agents can help users lower barriers and enhance experiences, even if it's just simplifying part of the asset issuance process, it is still meaningful.

From a macro Web3 perspective, AI Agents, as off-chain products, currently only serve as auxiliary tools for smart contracts, so there is no need to overly boast about their capabilities. Due to the lack of significant wealth effect narratives in the second half of this year, apart from MeMe, it is normal for AI Agents to gain attention around MeMe.

However, relying solely on MeMe cannot maintain long-term value. Therefore, if AI Agents can bring more innovative gameplay to the trading process, providing tangible value, they may develop into a universal infrastructure tool.