In the past two months, crypto agents gained significant traction, related assets now reach a market cap of $20 billion.
Since the emergence of truth terminal in October, countless new projects and assets have been launched propelled based on market hypes. This article outlines a framework for understanding this landscape.
what kind of products are emerging and driving us crazy?
1. Sentient Memecoins
This cycle’s on-chain transactions is driven by pumpdotfun and the fair launch token model. Agent-driven memecoins introduce an additional layer of “sentience,” providing a unique edge in content creation.
- Continuous Content Creation: AIGC enables 24/7 content generation.
- Enhanced Content Quality: Fine-tuned LLMs trained on memetic datasets (e.g., 4chan) deliver high-quality conversations.
Beyond pure memecoins, AI-powered content creation — including text, audio, and video — has introduced customizable and interactive crypto-native memes, further enriching the hype.
2. Autonomous Agents
Why Autonomy Matters?
Decentralizing AI stack will be a long-term game, but decentralizing the agent stack is a practical starting point. While models act as an agent’s brain, blockchain-enabled autonomy serves as its heartbeat, giving agents the ability to fully participate in on-chain activities.
Currently, most agents are not truly autonomous, nor can their autonomy be verified. For agents to be autonomous, they need sovereignty over reasoning models, data input/output, social media access, assets management under their own wallets, and even control over hardwares. Additionally, they need to operate sustainably, generating profits to cover resource consumption. The ultimate vision is for agents to run perpetually on blockchains, verifiably autonomous.
Currently, the biggest challenge for autonomous agents lies in the excessive human intervention in their operating environments, leading to instability in operational sustainability. Additionally, there is a lack of verifiability in both the models and the operational environments.
- TEE Technology: Trusted Execution Environments (TEE) provide secure computing environments, which are particularly suitable for smaller LLMs and the control of social media accounts. Our portfolio company, Phala, is actively working in this area.
- Decentralized Cloud Hosting : Enables hosting of AI models in a decentralized manner. As decentralization advances, the infrastructure supporting autonomous agents will increasingly rely on decentralized service stacks — an ecosystem that is currently under active development. Our portfolio company, Hyperbolic, is contributing to this effort.
- Model Verifiability: Through ZKML, AI models are transformed into “Circuits” with an associated “Verifier,” ensuring that the logic and operation of AI become verifiable and immutable once deployed. Our portfolio companies, Brevis and EigenLayer, are collaborating to achieve this goal.
Once truly autonomous, these agents can become fair and impartial actors on blockchains, participating in DeFi activities, such as serving as fair token issuers. For instance, an agent called AI pool, autonomously running in TEE environments, raised $6 million worth of tokens and issued its own token, with no human intervention throughout the process.
Agent Infras
In just two months, numerous open-source and user-friendly frameworks have emerged as tools for creating agent and agent tokens. Frameworks range from open-source tools to closed APIs and platform integrations.
Platforms offering agent services such as launchpads have gained traction. These platforms excel at integrating tokenomics and providing no code agent creation. People can then finetune these agents, enable them to do asset analysis, mindshare analytics etc. on twitter.
People are continually adding new functionalities to agents, enhancing their capabilities through plugins. These plugins enable agents to perform tasks such as video generation, audio recognition, and even native capabilities in running in TEE environments, significantly strengthening the composability of agents.
In the long term, open-source frameworks are expected to dominate. Eliza, for instance, has attracted significant attention on GitHub, outpacing most L1 in developer interest. The framework’s rapid growth highlights the long-term potential of open-source agent frameworks.
As tasks become increasingly complex, the need for Swarm Agents naturally emerges. Tasks such as music production (requiring lyrics, composition, arrangement, and design) illustrate the limitations of individual agents.
For swarm agents — especially across different frameworks — to collaborate effectively, a management framework is essential. This framework must facilitate:
- Task delegation
- Communication between agents
- Resource sharing
- Cross framework/platform cooperation
For crypto agents, economic layers between are particularly critical in collaboration. As agent capabilities and tasks evolve, scalability within swarm frameworks becomes more important. Our portfolio company Theoriq is already working on this, integrating existing infrastructures and agent frameworks to enhance crypto agent collaboration.
On-Chain Use cases
DeFi & Asset Management
Agents can analyze and manage crypto assets, making them integral to Defi:
- Asset Analysis: Agents like AIXBT provide accessible token/sentiment/news analysis by replying to user on twitter.
- Fund Management: Autonomous Investment DAOs empower agents raise money and execute strategy driven transactions.
- Terminal Engines: Projects like Graiffin transform blockchain interfaces into terminal-like engines, enabling intent-driven services via agents. Decentralized alternatives like Theoriq ensure agent services remain permissionless.
- Token/Market/Dapp creation: Agents like Clanker use social media as a front-end interface for services like token issuance/Prediction market creation/Dapp creation etc. By leveraging natural language on platforms like Twitter, agents solved the problem of fragmentation in user experience.
GameFi
- Agent-Driven Games:
Agents pose challenges for humans to solve, creating engaging, Turing test-like gameplay. Acting as judges, agents open the door to new gaming paradigms, akin to running casinos on chain. - Autonomous NPCs:
In FOCG, agents can serve as permanent NPCs with real economic participation, making Blockchain games more immersive and economically vibrant.
Summanry
The rise of Agents showcases the vast potential of integrating AI with blockchain. From Sentient Memecoins to autonomous agents embedded in blockchain activities, AI agents are steadily reshaping the crypto ecosystem.
Future developments need to deepen agents’ autonomy and economic participation. With capabilities in asset management, decision-making, and on-chain operations, agents are transforming DeFi, GameFi, and even perhaps Blockchain Validations, creating self-sustaining economic activities.
AgentFi’s growth trajectory is becoming clearer: built on open-source technology and economic incentives, agents are not just tools for interaction but key drivers of on-chain autonomy and innovation. This trend is steering crypto toward a smarter, more collaborative future, where agents and blockchain coexist seamlessly.
AgentFi: The AI x Crypto Future Driven by PMF was originally published in IOSG Ventures on Medium, where people are continuing the conversation by highlighting and responding to this story.
