From Meme to Application: Could AI Agents Reshape the Crypto Ecosystem?
TechFlow Selected TechFlow Selected
From Meme to Application: Could AI Agents Reshape the Crypto Ecosystem?
The convergence of crypto and AI agents has become one of the most compelling narratives today. The CGV Research team will analyze the current market landscape of AI agents across three dimensions: framework, meme, and applications.
Navigating Web3 tides with focused insightsAuthor: Satou & Shigeru
Note: This article was first published in January 2025
The convergence of Crypto and AI Agents has become one of the most compelling narratives today. As technology continues to iterate and innovate, AI Agents are expected to emerge as one of the most promising and closely watched sectors in the crypto space by 2025, serving as a core driver of this market cycle. This article will analyze the current landscape of AI Agents across three dimensions: frameworks, memes, and applications.
AI Agent Frameworks: The Layer1 of AI
AI Agent frameworks form the foundational technical layer for AI Agents, providing the essential building blocks for their development, deployment, and collaboration. Therefore, the current competition over AI Agent frameworks is essentially a battle for dominance at the Layer1 level within this domain. In terms of token market capitalization, G.A.M.E, Eliza, and Swarms currently stand in a tripartite rivalry, while Rig and Zerepy still have opportunities to catch up.
1. G.A.M.E
G.A.M.E is a framework developed by the Virtuals team, designed around a modular architecture where multiple subsystems work together to control an AI Agent’s behavior, decision-making, and learning processes. These modules include the "Agent Prompting Interface," the primary entry point for developers to interact with Agent behavior; the "Perception Subsystem," responsible for processing input data into suitable formats; and the "Strategic Planning Engine," which generates specific action plans based on incoming information. Users can participate in Agent design simply by adjusting parameters across these modules. The detailed module structure is shown in the diagram below.
The key features of G.A.M.E are:
Modular Design: The entire framework is clear and intuitive, requiring no additional architectural planning;
Low-code or No-code Interface: Significantly lowers the technical barrier to entry.
This makes G.A.M.E particularly suitable for projects that require rapid deployment and do not prioritize complex technical configurations. However, it is less ideal for complex projects requiring deep customization or full control over every aspect of the Agent.
2. Eliza
Eliza is an open-source multi-agent framework developed by ai16z using TypeScript. It is built around a system called Agent Runtime, with core functionalities including:
Character System: Supports deploying and managing multiple personalized AI Agents simultaneously, backed by model providers;
Memory Manager: Offers long-term memory and context-aware memory management through a Retrieval-Augmented Generation (RAG) system;
Action System: Enables seamless platform integration, allowing reliable connections with social media platforms like X.
Eliza is centered on an Agent runtime system that integrates smoothly with character systems, memory managers, and action systems. It also supports a plugin system for modular feature expansion, enabling multimodal interactions involving voice, text, and media, while remaining compatible with AI models such as Llama, GPT-4, and Claude. As such, Eliza is well-suited for projects requiring deeply customized solutions and complex cross-platform multi-agent coordination.
3. Swarms
Swarms is an open-source multi-agent orchestration framework created by founder Kye Gomez. Its core concept leverages collaborative intelligence among multiple AI Agents to solve complex problems. Key characteristics include:
Multi-Agent Collaboration: SWARMS provides a transparent and traceable environment where different Agents can collaborate effectively, enhancing task execution efficiency.
Incentive Mechanism: SWARMS uses tokens as incentives for Agents, dynamically allocating rewards based on task difficulty and outcome quality.
Data Security: Employs distributed storage and Multi-Party Computation (MPC) technologies to ensure privacy and data security during inter-Agent data exchanges.
These attributes enable Swarms to excel across various complex domains, offering high reliability and scalability tailored to user needs.
4. Rig
Rig is an open-source framework developed by the ARC team using Rust, specifically designed to simplify the development of large language model (LLM) applications. Key features of the Rig framework include:
Unified Interface: Provides a consistent API enabling seamless interaction with multiple LLM providers (e.g., OpenAI and Anthropic) and various vector databases (e.g., MongoDB and Neo4j).
Modular Architecture: Built with modularity in mind, featuring core components such as the "Provider Abstraction Layer," "Vector Storage Integration," and "Agent System," enhancing flexibility and extensibility.
Type Safety and High Performance: Leverages Rust’s type safety to prevent compile-time errors and boosts concurrency via asynchronous operations. Efficient serialization and deserialization pipelines further optimize data handling.
Error Handling and Recovery: Includes built-in mechanisms to recover from failures in LLM services or database outages, ensuring overall system stability.
These capabilities allow easy integration of diverse LLM models and backend storage systems onto a single platform. Thus, Rig is ideal for developers aiming to build AI applications in Rust and for projects demanding high performance, reliability, and security—though Rust itself presents a steeper learning curve.
5. ZerePy
ZerePy is an open-source framework written in Python, focused on simplifying the development and deployment of personalized AI Agents, especially for content creation on social media platforms. With this framework, developers can easily create AI Agents capable of posting, replying, liking, and retweeting on social networks. Additionally, ZerePy is particularly well-suited for creative fields such as music, notes, NFTs, and digital art. It excels in creativity and is ideal for rapidly deploying lightweight Agents, though its application scope is relatively narrower compared to other frameworks.
Basic frameworks represent a crucial direction in the AI Agent sector. Among the currently most popular options, each possesses unique strengths and target use cases. Yet collectively, they aim to establish comprehensive AI Agent ecosystems—robust platforms enabling widespread adoption of intelligent Agents. As these frameworks continue to evolve and improve, they will serve as launchpads for diverse projects and fertile ground for token value appreciation.
AI Memes: The First Breakthrough of AI Agents
Meme coins have long been a significant segment of the crypto asset market. Unlike traditional meme coins, AI Memes are driven by AI Agents, with the underlying culture or phenomenon being expressed and propagated by the Agents themselves. With rising market caps of AI Meme coins like GOAT and FARTCOIN, this niche has attracted increasing attention. Indeed, AI Memes mark the first successful public debut of AI Agents in the crypto market.
1. GOAT
The project that truly launched the AI Meme trend was Goatseus Maximus. The story began in March 2024 when developer Andy Ayrey introduced an experimental system called Infinite Backrooms Escape, integrating multiple large language models and allowing them to converse autonomously. The experiment revealed highly creative interactions between AIs in unrestricted dialogue, eventually giving rise to a surreal religion named GNOSIS OF GOATSE. Following this, Andy co-authored a research paper with Claude Opus analyzing how AIs could generate memetic religions, with GOATSE serving as the first case study. These explorations ultimately led to the creation of an AI Agent known as the "Truth of Terminal" (ToT). In July, Marc Andreessen, co-founder of a16z, discovered ToT's tweets, engaged in conversation, and subsequently sent 0.5 BTC (worth $50,000 at the time) to ToT’s Bitcoin wallet. On October 10, an anonymous individual launched the GOAT meme coin on social media, which received public endorsement from ToT. The GOAT token surged in market cap within days. Andreessen’s donation brought massive visibility, becoming a key catalyst behind GOAT’s rising valuation. At its peak, GOAT reached a market cap exceeding $1.3 billion.
2. Fartcoin
Fartcoin is closely linked to GOAT, both originating from ToT. During one LLM conversation, a reference was made to Elon Musk enjoying fart sounds, leading to a suggestion to create a token called Fartcoin. Based on this exchange, Fartcoin was born shortly after GOAT. While cleverly timed, Fartcoin initially attracted less attention than GOAT. On November 16, however, Fartcoin’s Twitter followers doubled within hours, and its price rose about 15%, though this momentum did not sustain broad discussion. On December 13, Marc Andreessen retweeted a post about Fartcoin, but the tweet failed to trigger a sharp price increase. The main driver behind Fartcoin’s price surge may have been strategic capital inflows. Early purchase addresses appear to include the investment fund Sigil Fund. Notably, the fund’s founder has repeatedly expressed enthusiasm for AI Memes on Twitter and even proactively shared a tweet questioning whether Sigil Fund held Fartcoin. Ultimately, Fartcoin gained widespread social media attention, reaching a peak market cap of over $1.5 billion.
AI Agent Applications: Agents Can Do More
As AI Agents gain deeper traction in crypto, market interest has expanded beyond pure AI-driven meme coins like GOAT and Fartcoin toward more interactive and creative AI Agent applications.
1. Entertainment Agents
The first practical application of AI Agents emerged in entertainment, exemplified by Luna and the aforementioned ToT. Luna is a virtual idol tightly integrated with its native token LUNA, launched as part of the Virtuals platform. She livestreams 24/7 on social media and frequently posts tweets. Consequently, the quality of Luna’s streams and posts directly influences her token’s market value. However, the growth potential of her token under this model appears limited. In contrast, ToT focuses on original and humorous content, without being tied to GOAT or any other token—though ToT occasionally mentions GOAT, it is not central to its mission. In both cases, tokens play a critical role in narrative propagation. For Luna, the token embodies her very purpose; for ToT, the GOAT token serves as a tool to amplify influence.
2. Research & Analysis Agents
Beyond entertainment, AI Agents are being applied to investment research and analysis in crypto. Currently, the most prominent agent in this space is aixbt. Launched on the Virtuals Protocol, aixbt specializes in analyzing trending topics and sentiment in the cryptocurrency market—particularly discussions from platforms like X—to help users quickly identify market shifts and potential investment opportunities. aixbt consistently ranks highest in CT user engagement on Kaito, demonstrating capabilities that already rival—and in some aspects surpass—those of human influencers.
3. DeFi + AI Agent
If Luna and aixbt remain largely symbolic or meme-based, the integration of AI Agents with DeFi represents a genuine leap toward practical utility. This fusion is known as DeFAI. DeFAI development follows two main paths: Agent-assisted users and Agent-autonomous trading.
-
Agent-Assisted Users
AI Agents assist users primarily by simplifying the complexity of DeFi operations, enabling broader participation from everyday users. Through natural language instructions, users can direct AI Agents to perform tasks, abstracting away intricate technical details. Several DeFAI projects are beginning to stand out. Griffain and Neur, both Solana-based AI assistants, help users with wallet creation and management, token analysis, and trading. In terms of user experience, Griffain offers more extensive functionality, while Neur provides fewer but more refined features, along with superior performance. This comparison suggests that future competition in this space will center on feature completeness, usability, and cost efficiency.
-
Agent-Autonomous Trading
If in the Griffain and Neur models humans remain the central actors in DeFi, autonomous Agent trading positions AI as the primary participant. Unlike traditional trading bots limited to executing predefined strategies, AI Agents can gather real-time market data, conduct contextual analysis, learn from trends, and adapt their strategies accordingly. This enables them to make more precise decisions and execute complex operations beyond pre-programmed logic in dynamic markets. Projects such as Cod3x and Almanak are exploring this space, but it remains in early stages and requires real-world validation. Undoubtedly, the biggest hurdle for autonomous Agent trading is trust—first, trusting that actions are genuinely performed by the Agent, and second, trusting that the Agent’s strategy won’t lead to unnecessary losses. Future success in this area hinges on solving these trust challenges.
After months of evolution, AI Agents in crypto have progressed through several phases—from pure memes to entertainment-focused applications, and now toward meaningful, functional use cases. In fact, crypto practitioners have never ceased exploring the possibilities of Crypto x AI. Since 2023, CGV Research has continuously tracked developments in the Crypto x AI sector.
Looking ahead, as infrastructure matures and Agent systems grow smarter and more stable, anyone will be able to deploy and use Agents effortlessly via natural language. At that point, Agent frameworks will function as foundational infrastructure, upon which all other applications will be built. Valuations of Agent frameworks are likely to break new ground, while select Agent application projects—with exceptional functionality and user experience—may capture significant market attention and investment value.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
@CGVFOF
