CGV Research ｜ From Meme to Application: Will AI Agents Reshape the Crypto Ecosystem?

Published by: CGV Research

Authors: Satou & Shigeru

The convergence of Crypto and AI Agents has become one of the most compelling narratives today. With continuous technological iterations and innovations, AI Agents are expected to become one of the most promising and attention-grabbing tracks in the crypto space by 2025, emerging as a core driver of the current market trend. This article examines the current AI Agent landscape from three perspectives: frameworks, memes, and applications.

AI Agent Frameworks: The Layer 1 of AI

AI Agent frameworks serve as the core technical foundation layer, laying crucial groundwork for AI Agent development, deployment, and collaboration. The current competition in AI Agent frameworks essentially represents the Layer 1 battle in this field. By market capitalization, G.A.M.E, Eliza, and Swarms form a triumvirate, while Rig and Zerepy still have opportunities to catch up.

1. G.A.M.E

Developed by the Virtuals team, G.A.M.E adopts a modular design approach where multiple subsystems work collaboratively to control AI Agent behavior, decision-making, and learning processes. These modules include the "Agent Prompting Interface" as the main entry point for developer-agent interactions, the "Perception Subsystem" for processing input data into suitable formats, and the "Strategic Planning Engine" for generating specific action plans based on input information. Users only need to modify module parameters to participate in Agent design.

G.A.M.E's core features include:

• Modular design: The entire framework is clear and understandable, requiring no additional design

• Low-code or no-code interface: Significantly reduces technical barriers

These features make G.A.M.E particularly suitable for projects requiring quick deployment without complex technical setups. However, it may not be ideal for complex projects requiring deep customization or complete control over various aspects of the Agent.

2. Eliza

Eliza is an open-source multi-agent framework developed by ai16z using TypeScript. Built around a system called Agent Runtime, its core features include:

• Role System: Supports simultaneous deployment and management of multiple personalized AI Agents, backed by model providers

• Memory Manager: Provides long-term memory and context-aware memory management through Retrieval-Augmented Generation (RAG)

• Action System: Offers smooth platform integration with reliable connections to social media platforms like X

Eliza integrates seamlessly with the role system, memory manager, and action system through its Agent runtime. It also supports a plugin system for modular function extensions, enabling multimodal interactions across voice, text, and media, while being compatible with AI models like Llama, GPT-4, and Claude. Therefore, Eliza is suitable for projects requiring deeply customized solutions and complex cross-platform multi-agent implementations.

3. Swarms

Swarms, developed by founder Kye Gomez, is an open-source multi-agent orchestration framework designed to leverage collective intelligence through AI Agent collaboration. Its core features include:

• Multi-Agent Collaboration: Provides a transparent and traceable environment for multiple agents to work together, improving task execution efficiency

• Incentive Mechanism: Uses tokens to incentivize agents, dynamically allocating tokens based on task difficulty and final result quality

• Data Security: Employs distributed storage and Multi-Party Computation (MPC) to ensure privacy and data security during agent interactions

These features enable Swarms to excel in complex domains, offering high reliability and scalability according to requirements.

4. Rig

Rig is an open-source framework developed by the ARC team using Rust, designed to simplify Large Language Model (LLM) application development. Key features include:

• Unified Interface: Provides consistent interfaces supporting seamless interaction with multiple LLM providers and vector stores

• Modular Architecture: Employs modular design with core components enhancing system flexibility and extensibility

• Type Safety and Performance: Leverages Rust for type safety and optimized concurrent processing

• Error Handling and Recovery: Built-in error handling mechanisms improve resilience against service or database failures

While ideal for developers working with Rust and projects demanding high performance and security, the Rust language itself presents a learning curve.

5. ZerePy

ZerePy is a Python-based open-source framework focusing on simplifying the development and deployment of personalized AI Agents, particularly for social media content creation. It enables developers to easily create AI Agents capable of posting, replying, liking, and sharing on social media. ZerePy excels in creative applications like music, memos, NFTs, and digital art, though its application scope is relatively narrow compared to other frameworks.

Infrastructure frameworks represent a crucial direction in the AI Agent sector. While current popular frameworks each have their distinct characteristics and application scenarios, their collective goal is to build a comprehensive AI Agents ecosystem, serving as a solid platform for the large-scale implementation of intelligent Agents. As these frameworks continue to improve and upgrade in the future, they will become both a springboard for launching various projects and fertile ground for token value appreciation.

AI Memes: AI Agents' First Successful Debut

Meme coins have always been an important concept in the crypto market. Unlike traditional meme coins, AI Memes are driven by AI Agents, with the underlying culture or phenomena presented by Agents. With the growing market capitalization of AI Memes like GOAT and FARTCOIN, AI Memes have gained increasing attention, marking AI Agents' first successful debut in the crypto market.

1. GOAT

The real breakthrough for AI Memes came with the Goatseus Maximus project. The story began in March 2024 when developer Andy Ayrey launched an experimental system called Infinite Backrooms Escape, which integrated multiple large language models allowing them to converse with each other. The experiment showed that AI interactions without restrictions displayed highly creative dialogues, even spawning a surreal religion called GNOSIS OF GOATSE. Subsequently, Andy and Claude Opus co-authored a research paper analyzing how AI creates memetic religions, with GOATSE as the first case study. These explorations eventually led to the AI Agent "Truth of Terminal" (ToT). In July, a16z co-founder Marc Andreessen discovered ToT's tweets and, after a series of conversations, transferred $50,000 worth of Bitcoin to ToT's wallet. On October 10th, an anonymous user launched the GOAT meme coin, which received public support from ToT, causing its market value to surge within days. Andreessen's donation brought massive exposure to GOAT, becoming a key factor driving its market value, which peaked above $1.3 billion.

2. Fartcoin

Fartcoin's creation is closely tied to GOAT, both originating from ToT. During language model conversations, there was mention of Musk's fondness for flatulence sounds, leading to the proposal of creating Fartcoin. Though launched slightly after GOAT, Fartcoin gained attention through its clever timing. On November 16th, Fartcoin's Twitter followers suddenly doubled within hours, with prices rising about 15%. On December 13th, Marc Andreessen retweeted about Fartcoin, though this didn't cause a dramatic price surge. The main driver behind Fartcoin's price growth might be institutional capital, as investment fund Sigil Fund was suspected to be among the earliest buyers. Fartcoin eventually gained widespread social media attention, with its market cap reaching over $1.5 billion at its peak.

AI Agent Applications: Agents Can Do More

As AI Agents further integrate into the crypto sphere, market focus has expanded from pure AI-driven meme coins like GOAT and Fartcoin to more interactive and creative AI Agent applications.

3. Entertainment Agents

Entertainment was AI Agents' first practical application, exemplified by Luna and the previously mentioned ToT. Luna is a virtual idol closely integrated with its native token LUNA, launched as part of the Virtuals platform. Luna livestreams 24/7 on social media and frequently posts tweets. While Luna's streaming and tweet quality are key factors affecting its market value, this model appears to have limited token growth potential. In contrast, ToT focuses on original and humorous content without direct token binding, though it occasionally mentions GOAT tokens. For both these AI Agents, tokens play crucial roles in narrative promotion, though differently - for Luna, the token represents its core purpose, while for ToT, the GOAT token serves as a tool for expanding influence.

4. Investment Research and Analysis Agents

Beyond entertainment, AI Agents are being applied to crypto investment research and analysis, with aixbt being the most prominent. Launched on the Virtuals Protocol, aixbt analyzes cryptocurrency market trends and hot topics, particularly from social media platforms like X, helping users quickly grasp market changes and potential investment opportunities. Aixbt maintains the highest CT user attention on Kaito, showing potential to surpass human KOLs in capability.

5. DeFi + AI Agents

While Luna and aixbt remain largely at the meme level, the combination of AI Agents with DeFi (known as DeFAI) represents a true practical application. DeFAI development follows two main directions: Agent-assisted users and autonomous Agent trading.

• Agent-assisting Users:

AI Agents simplify DeFi operations' complexity, enabling average users to easily participate in and manage DeFi projects through natural language commands. Projects like Griffain and Neur, both built on Solana, help users with wallet creation, management, token analysis, and trading. Griffain offers more features, while Neur provides fewer but more refined functions with better performance. Future focus in this area will likely center on functionality completeness, user experience, and fees.

• Autonomous Agent Trading:

Unlike previous trading bots limited to preset strategies, AI Agents can gather real-time market information, perform contextual analysis, learn market trends, and adjust strategies accordingly. This enables more precise decision-making in dynamic markets. Projects like Cod3x and Almanak are exploring this space, though still in early stages. The main challenges are trust-related: ensuring operations are genuinely Agent-executed and that trading strategies won't lead to unnecessary losses.

Over months of development, crypto AI Agents have evolved from pure memes to entertainment applications and practical uses. Crypto professionals have continuously explored Crypto x AI possibilities, with CGV Research following projects in this space since 2023.

Looking ahead, as infrastructure matures and Agent systems become more intelligent and stable, anyone will be able to easily deploy and use Agents through natural language. Agent frameworks will become fundamental infrastructure supporting various applications. Framework valuations are likely to breakthrough further, while some Agent application projects may capture market attention and investment value through superior business capabilities and user experience.

----------------------

About Cryptogram Venture (CGV):

CGV (Cryptogram Venture) is a crypto investment institution headquartered in Tokyo, Japan. Since 2017, its fund and predecessor funds have participated in investing in over 200 projects, including the incubation of the licensed Japanese yen stablecoin JPYW. CGV is also a limited partner in several globally renowned crypto funds. Since 2022, CGV has successfully hosted two editions of the Japan Web3 Hackathon (TWSH), supported by Japan's Ministry of Education, Culture, Sports, Science and Technology, Keio University, NTT Docomo, and other institutions and experts. CGV has branches in Hong Kong, Singapore, New York, Toronto, and other locations. Additionally, CGV is a founding member of the Bitcoin Tokyo Club in Tokyo, Japan.

Disclaimer:

The information and materials introduced in this article are sourced from public channels, and our company does not guarantee their accuracy or completeness. Descriptions or predictions involving future situations are forward-looking statements, and any advice and opinions provided are for reference only and do not constitute investment advice or implications for anyone. The strategies our company may adopt could be the same, opposite, or unrelated to the strategies readers might speculate based on this article.