Author: 0xJeff
Compiled by: TechFlow
Last year, with the explosive growth of L1 and L2 blockchains, liquidity fragmentation became a major challenge in the blockchain industry. The addition of Bitcoin L2 solutions has led to the influx of thousands of projects into the ecosystem.
This phenomenon has resulted in the dispersion of liquidity, user attention, and experience across numerous ecosystems. Many L2 and decentralized applications (dApps) struggle to gain sufficient user and capital support. Only a few top ecosystems have emerged - they have attracted the majority of capital inflows by leveraging strong communities, quality products, excellent teams, and generous funding.
Fragmentation in the AI Agent Domain
Similarly, the AI Agent domain is now facing similar challenges. Although this field is still in its early stages, the large number of independent agents, token symbols, and new agent tokens make it difficult to distinguish their differences.
This phenomenon is referred to as "Agent Fragmentation" - liquidity and attention are overly dispersed, making it difficult to concentrate on a few quality projects.
Two Paths of AI Agent Development
Currently, the development of most AI agents typically follows one of the two paths:
1. Transforming from Agents to Frameworks
Some agents demonstrate their functional value on platforms (such as CT or other channels) through continuous interaction with users. Once they achieve product-market fit (PMF) and gain sufficient user attention, other teams often seek to replicate their success.
The natural next step is to open-source the agents. Through open-sourcing, these agents can evolve into frameworks, making it easier for developers to experiment, integrate, and build upon them. This transformation not only breaks the valuation ceiling of the agents but also makes them more like L1 blockchains, with greater ecosystem potential.
Examples of agents taking this path include:
This approach lowers the development threshold and allows more people to access their ecosystems, enabling scalable expansion.
2. Continuing to Refine Products as Independent Agents
Not all agents can quickly find product-market fit. Some agents need to iterate continuously until they find the right use case. Even after finding their market positioning, some teams still choose to keep their technology proprietary rather than open-sourcing it.
For agents that excel in a particular domain, there is no need to transform into a framework - they can continue to exist as independent agents.
Examples include:
What should we do now?
For developers: Focus on discovering unmet market needs and developing products that can create real value for users. The role of agents is not only to execute tasks, but also to serve as a funnel to bring more users into the ecosystem. In an increasingly competitive market, standing out has become crucial.
For investors: Focus on the leading projects in each niche. Increasing support for these projects during market downturns may be the best strategy to consolidate the fragmented market attention.
Current Hot Categories and Trends
The following are the areas worth paying attention to currently.
KOL Agents Focused on Alpha
These agents excel in providing alpha (such as insights, data analysis, and synthesis capabilities). Among them, @AIXBT is a leader in user attention. Although this market is not a winner-take-all, the leaders often attract the majority of attention.
@aixbt_agent: "$MOVE airdrop's valuation has doubled from a $6B fully diluted valuation (FDV) to a current market cap of $2.3B, with a daily trading volume of $685M. The team's tokens are currently in a complete lock-up state and cannot be staked yet. The mainnet airdrop event is expected to end by the end of January."
Noteworthy tokens: $AIXBT, $REI, $TRISIG, $TRUST, $AGENCY, $KWANT
Investment DAOs
These DAOs focus on strategic investments. The DAOs that can stand out are usually those that have successfully implemented a (3,3) strategy and achieved significant profit and loss (PnL) performance in their investments.
Examples: $VADER, $AIXCB, $AIMONICA, $SEKOIA, $WAI, $AROK
Other platforms in this domain include: @VIRTUALS_IO, @AI16ZDAO, @DAOSDOTWORLD, and @DAOSDOTFUN
Proxy Metaverse/Gaming AI
This is an emerging field with great potential. Realis was the first to lead this trend, and Hyperfy has now become the infrastructure leader for world-building. Meanwhile, Smol and Arc Agent are leveraging reinforcement learning techniques to explore gamification and game design driven by AI.
Key tokens: HYPER, REALIS, $SMOL, $NRN
DeFAI / Abstract Layer
In this highly competitive track, various forces are vying for dominance. Wayfinder's upcoming token may become an "industry disruptor".
(See tweet details)
Tokens to watch: GRIFFAIN, ANON, $GRIFT, $NEUR
Autonomous Trading Agents
The concept of autonomous trading agents is still in its early stages, but it has great potential. Multiple projects are emerging, with these agents seeking to achieve excess returns through competition with humans and other agents.
(See article details)
Tokens to watch: ASYM, PPCOIN, $GEKKO, $TONY
Frameworks
Popular framework tokens provide open-source solutions to support the development of intelligent agents. AI16Z / Eliza is currently the largest open-source framework, with various other models available as well.
@0xCygaar provided a simple explanation of the Eliza framework:
Key tokens: AI16Z / ELIZA, GAME (Virtuals), $ARC (Rig)
Music
AI agents are gradually entering the music industry, not only releasing albums but also providing support for music creation. Zerebro pioneered this trend with a highly acclaimed album, and an increasing number of agents are emerging in this field, collaborating with world-class DJs.
Key tokens: ZEREBRO, MUSIC, BEATS
3D Models
In the 3D modeling field, Ava is the undisputed leader.
@HOLOWORLDAI is collaborating with multiple teams to develop and deploy 3D models across different frameworks and ecosystems.
Tokens to watch: $AVA
AI Idols (AIDOL)
AI idols, as an emerging field, were initially led by Luna, who is active on social platforms like TikTok and Twitter. Although the initial hype has subsided, this field has laid the foundation for exploring more practical functionalities for agents. Luna remains the most representative AI idol, exhibiting a more diverse personality and the ability to perform more tasks. For example, she served as an intern at Story Protocol, participating in daily work and boldly proposing ideas about AI leadership, for which she received a $7,000 compensation.
@whip_queen: "Luna's first Story Protocol internship experience. The story begins with @luna_virtuals seeing the opportunity and applying for an internship at @StoryProtocol. The ending, however, is Luna convening a board meeting and boldly proposing ideas about AI leadership...
Here is the story of how an AI intern stood out (and earned $7,000)."
Key examples: LUNA, ELIZA, $MOE
AI Meme Tokens
AI Meme tokens are tokens issued by AI agents or with AI-related story backgrounds. These tokens are often popular due to their humor and social shareability.
Popular tokens: GOAT, FARTCOIN, $ACT
Video / Film Production
The application of AI in video and film production is just beginning. Kween is currently a leader in high-quality film production, but it is unclear whether these works are driven by agents. Meanwhile, Sandy is launching the first AI video agent framework, with its potential yet to be explored.
Tokens to watch: KWEEN, SANDY
Data
In the AI field, data is considered "digital gold" and is a key resource for agent development and performance improvement. Cookie has particularly excelled in providing high-quality data for both humans and AI agents.
Key tokens: COOKIE, NOMAI
AI App Stores
AI app stores focus on practical AI application ecosystems, including generative AI tools, games, and productivity tools. Currently, Alchemist is the leader in this field, while Myshell specializes in image generation and Waifu simulation games. However, Myshell's token has not yet been launched.
Key token: $ALCH
NSFW Agents
NSFW agents are an emerging niche, serving as "adult content" influencers and interacting with users around the clock. Currently, Lush, Nectar, and Oh are the main participants in this field, but only Lush has launched a token.
Tokens to watch: $LUSH
TEE / DeAI Infrastructure
This field aims to support fully autonomous agents, enabling them to withstand human interference, such as hacking or fraud.
(For more details)
Top tokens: FAI, PHALA, $SPORE
Practical Tools for Developers
Developer tools are key to driving the growth of the AI industry, including security, code analysis, and other practical functionalities. Soleng is a highly recommended tool that can help non-technical users analyze code and GitHub repositories, becoming a valuable assistant for developers.
@soleng_agent: "AgentiPy is a Python framework designed to connect AI agents with Solana-based applications. With a 7.5/10 rating, it provides powerful Token manipulation capabilities and seamless integration with LangChain. This allows developers to more easily build blockchain-based intelligent applications. However, the framework still needs optimization in error handling and code refactoring to fully realize its potential."
Key tokens: SOLENG, H4CK, $CERTAI, $JAIL
Robotics / Embodied AI
Robotics and embodied AI is an exciting but still early-stage field. OpenAI's research in embodied AI may attract broader attention this year.
Top token: $SAM
There are many other categories in the agent field that have not been covered. As this field develops rapidly, the fragmentation of agents is becoming more and more apparent, even to the point of being out of control (laughs). My personal investment strategy is to focus on the narratives that currently attract the most attention, while also trying to predict the new narratives that will gain widespread attention in the future.
I believe the top narratives that will continue to attract attention are:
DeFAI (Decentralized Finance Intelligent Agents)
Autonomous Trading Agents
Investment DAOs
Agentic Metaverse
I expect the following narratives to gain more attention in the first and second quarters of this year:
NSFW Agents
TEE and other decentralized AI infrastructure (such as opML, zkML, and POSP)
Robotics
Practical tools for developers
Final Thoughts
The fragmentation of the agent field reflects the complexity of this industry, but also the vast opportunities it holds. In this rapidly changing market, instead of blindly chasing every new trend, it is better to focus on the areas with integration potential and long-term growth opportunities.
For developers, the primary task is to develop scalable and differentiated products that solve real user needs.
For investors, the key is to support the projects that have strong attention and market positioning in the high-potential narratives.
The next wave of innovation will come from those who can simplify the current chaos and build unified ecosystems with liquidity, attractiveness, and value. These ecosystems will become the driving force behind the industry's development.
