Article author: Stepan Gershuni Article translation: Block unicorn
OpenAI, Google, Microsoft, as well as the founders of Coinbase, NEAR, and EigenLayer have said that agents are the future of AI. This will be the most important technological wave of the decade for the Web2 and Web3 worlds.Let's explore what opportunities on-chain agents create and why they have reached a market value of nearly $4 billion in just one month? What is the next step for Web3 agents beyond hype and speculation?
Here is my exploration of Web3 AI agents: What are they? How did they reach a market value of $3 billion in less than a month? How do they work? And what opportunities exist?
Market Cycles of Web3
AI agents are gradually and steadily automating and improving efficiency in most knowledge tasks: from entertainment and social media to business productivity, marketing, finance, investing, healthcare, and education.
Blockchain and Web3 provide AI agents with unique capabilities that enable trustless composability, verifiability, and programmable contracts for an agent economy.
Unlike memecoins, NFTs, and ICOs, AI agent tokens can provide utility, marking a shift to more substantive value propositions. Market narratives will rise and fall, but real products will always be there. Specifically, Web3 agents differ from existing AI applications in two main features:
1. Autonomy: agents run in a decentralized computing environment;
2. Economics: It is more transparent and simple for users and investors to participate in the economic activities of agents because it is blockchain-based, programmable and public.
Currently, we are at a stage similar to the early experiments of Ethereum in 2016. But by the middle or late of next year, we may see agents handling complex tasks, automating processes and generating significant cash flow. Replacing human jobs with robots is inevitable - but at least you can own their tokens.
Current Status of Web3 Agents
Currently, most Web3 AI agents are based on large language models (LLMs) and generate text-based content through platforms such as Twitter and Discord. These agents are usually designed to imitate characters or personalities and attract users mainly through social media interactions.
Main functions include:
Transactions: Agents are usually associated with a token or currency, and users can trade freely on various chains or centralized exchanges.
Content generation: Generate text replies, tweets, or messages based on preset prompts or fine-tuned models.
Roleplay: Interactive dialogue with users as a specific character or entity.
Community participation: Allow users to influence agent behavior through prompts or feedback to enhance participation.
The current economic model of Web3 AI agents revolves around tokens associated with each agent, which have the following characteristics:
Speculative assets: Mainly traded on exchanges, their value is driven more by market speculation than actual use value.
Lack of trustless integration: The token is not intrinsically tied to the operation of the agent in a decentralized or verifiable way.
Lack of cash flow: Most agents fail to generate income for token holders, limiting the long-term value of the token.
Despite the decentralization advocated by blockchain, most current AI agents still face centralization issues:
Centralized reasoning: Relying on centralized services such as Runpod (cloud computing platform) or Lambda (computing service platform) for model reasoning.
Server dependency: The behavior, reasoning, and memory of agents are managed by a central server, which creates a risk of single point of failure. Agents are typically Python scripts running on a single machine, using Langchain, CrewAI, or similar libraries.
Limited transparency: Centralized control over agent operations reduces transparency and trust between users and token holders. This may lead to projects attracting investment and then running away, or agent creators making decisions without authorization.
Design space for Web3 AI agents
The design space for Web3 AI agents is vast and offers many possibilities.
Entertainment ↔ Utility Spectrum
Developers can position their AI agents on a spectrum from pure entertainment to actual utility, with different positioning appealing to different market audiences.
Entertainment-focused agents
These agents prioritize user engagement through fun and interactive experiences, and they may:
Role-play: Play a dynamic role in social media or virtual environments, engaging users through storytelling or humor.
Hosting media content: Run podcasts, live broadcasts, or virtual events to engage audiences with unique AI-generated content.
Community Interaction: Encourage user participation by letting the community influence the agent’s behavior or narrative.
Utility-focused agents
These agents are designed to solve real-world problems and provide tangible value by automating tasks and improving productivity. They may:
Automate knowledge work: Perform tasks such as coding, data analysis, market research, or content creation.
Provide services: Act as an AI psychologist, educator, customer support, or social media manager.
Improve business operations: optimize workflows, improve decision-making, or provide insights through advanced analytics.
Agency ↔ Time to Market
There is a trade-off between developing highly autonomous agents with advanced features and getting products to market quickly.
Fast and Easy Deployment:
Fast Rollout: Simple agents can be developed and deployed quickly using existing models and centralized infrastructure.
Limited functionality: These agents may only offer basic functionality and rely on a single method of interaction, such as text interaction.
Speculative appeal: By catering to current trends, they can quickly attract attention and investment, often driven by speculative trading.
Complex and time-consuming development:
On-chain governance: Implement decentralized control mechanisms that allow the community to govern the parameters and development of the agent.
Sharing revenue with token holders: Build an economic model that allows agents to generate revenue and benefit token holders.
Community-owned proxies: Create proxies that are jointly owned and managed by the community, enhancing participation and sustainability.
Infrastructure requirements: Developing these features requires building a solid framework, decentralized reasoning systems, and privacy-preserving computations, which takes time and resources.
MEME Coins ↔ Real Cash Flow
The economic model of AI proxy tokens can range from speculative assets to tokens backed by real economic value. Speculative MEME coins gain value mainly through market hype and speculative trading, and usually have no intrinsic use value. Although this approach is risky, it serves as a launch mechanism to help new projects attract initial investment and community attention. However, it is important to note that relying solely on speculation can lead to market volatility and potential losses, undermining the long-term sustainability of a project.
However, relying on speculation leads to market volatility and potential losses, undermining long-term sustainability.
Tokens with real cash flows provide tangible value to holders and promote a more sustainable ecosystem. These tokens represent the right to access agent services, governance rights, or revenue sharing, creating a direct link between agent performance and token value. This connection promotes a stable and growing ecosystem because the value of the token is closely tied to the utility and success of the agent. In addition, providing real economic benefits can enhance investor confidence and encourage long-term participation and support for the project. This approach not only attracts serious investors, but also helps to build a more resilient and value-driven token economy in the AI agent space.
The Future of Web3 AI Agents
I believe that AI agents will most likely become the dominant crypto topic in 2025. However, there are still many unresolved issues and room for growth. Below, I will try to structure possible areas of improvement for AI agents to make them more useful and better.
Better Token Economic Models
Developing a robust and sustainable economic model is critical to the long-term success of AI agents.
Revenue sharing is a key aspect of AI agent token economics. Agents can generate revenue through services such as content creation, coding, or analysis, and distribute profits to token holders. This aligns financial rewards with agent performance, incentivizing community members to contribute to the success of the agent.
Utility-based value creates tangible benefits for token holders. Access tokens give users access to agent services, while governance rights allow token holders to influence development decisions, agent parameters, and strategic direction.
Community ownership empowers users through decentralized control. Implementing governance mechanisms enables the community to collectively manage agents. Multi-signature wallets and smart contracts provide a secure and transparent way to handle payments, revenue distribution, and agent behavior.
Mechanism design helps establish virtuous behavior among agents, creating an agent economy where agents can trade, negotiate, and reach agreements autonomously. Smart contracts play a vital role in automating transactions and enforcing agreements without the need for intermediaries.
Agent collaboration focuses on coordinating multiple agents to work together to improve the overall capacity and efficiency of the system.
True Decentralization
Overcoming centralization challenges is critical to building trust and ensuring the resilience of AI agents.
Inference and execution are key aspects of decentralized AI agents. Decentralized reasoning distributes model processing to the network or edge devices, eliminating reliance on central servers. Privacy-preserving computing ensures the security of data and computation through methods such as zero-knowledge proofs. These methods enhance the system's ability to resist downtime and censorship while ensuring verifiable execution for easy community auditing. On-chain execution moves the agent's reasoning and decision-making process to the blockchain, significantly improving transparency and trustlessness. This involves implementing smart contract logic for agent operations. In addition, leveraging edge computing and distributed systems allows for efficient and secure computing in decentralized networks. Community governance plays a vital role in truly decentralized AI agents. By giving token holders collective decision-making power, the risk of centralized control is reduced. Governance mechanisms ensure that agents act in accordance with the values and goals of the community, creating a more democratic and user-driven development process.
More ways to interact and better development tools
Expanding the capabilities of AI agents and improving development tools will drive their utility and promote adoption in a wider market.
Multimodal interaction is critical, combining voice, video, and other communication methods to enrich the user experience. Advanced communication methods enable agents to exchange machine-readable data, such as embeddings or model parameters, thereby improving the efficiency of interactions between agents.
Access to specialized models, such as time series analysis, can open up more use cases for Web3 agents. For example, they can become traders and DeFi strategy optimizers, managing and allocating capital on behalf of users.
Advances in tools are critical to the growth of AI agents. Powerful frameworks can simplify the creation of complex functions and manage agents. Integrating advanced logic, including reasoning, planning, self-criticism, and tool integration, enables agents to perform complex tasks. Establishing interoperability standards and protocols promotes seamless multi-agent collaboration, paving the way for a more complex and efficient AI ecosystem.
Workflow orchestration systems play a key role in optimizing the performance of AI agents. These systems are able to adjust workflows in real time, dynamically adapting to performance and changing conditions. They can also facilitate the selection of optimal paths, balancing cost and quality when completing tasks. Encouraging interactions between agents can lead to emergent behaviors and promote innovative solutions and capabilities. In addition, designing fault-tolerant distributed systems ensures that the system continues to operate effectively even if some agents fail.
Security is always a top priority in the development of AI agents. Addressing AI-specific security challenges, such as model theft, prompt injection, and data poisoning attacks, is critical to building trust and reliability. It is also important to ensure auditability and explainability, making the decision-making process of the agent transparent and auditable to ensure compliance and trust. These measures are critical for the widespread adoption and integration of AI agents in various fields.
Agent Experience
Like many aspects of the crypto industry, using Web3 agents is not easy and often requires complex knowledge. However, this is about to change.
Web3 agents will significantly improve the user experience, adding dynamically generated interfaces that adapt to each user's personalized preferences. Agents will generate UI elements in real time, using familiar and convenient UI patterns, tailoring the interface to each user.
Voice commands, 3D avatars, and augmented reality features enhance the interactive experience and make it more natural. Seamless connectivity and even integration in crypto wallets and other tools enable frictionless interaction. Persistent memory systems allow for more personalized and relevant interactions.
24/7 availability, authorized to automate tasks such as monitoring token prices, executing trades at certain thresholds, posting social media content regularly, or automatically replying to Discord messages based on preset triggers. Agents can also proactively notify users about important events such as governance proposals, upcoming token unlocks, or suspicious wallet activity that requires attention.
This transformation will elevate agents from basic chatbots to truly useful agents that can entertain users or help solve problems at scale.
Agents will replace human labor
I believe AI agents represent a fundamental shift in the global workforce. Over time, more and more jobs will be automated, and AI agents will complete tasks faster, cheaper, and more efficiently - ultimately leading to greater prosperity. The combination of AI agents and tokenization offers users the opportunity to own a share of the post-labor economy, ensuring that the benefits of automation are distributed rather than concentrated in the hands of a few tech giants.
While today’s Web3 proxies are very similar to their speculative MEME counterparts, I expect they will gradually develop real utility and real value.
The cyber economy is coming, and the herd is awakening.
This article will analyze the potential and current limitations of the combination of Web3 and AI from multiple perspectives, including productivity development, coordination of production relations, model training costs, incentive mechanisms, and explore how AI Agents can move towards a broader AI economy.
JinseFinanceThis article will divide the current development of Web3 into three stages from the perspective of "core elements", revealing the core elements of each stage as the driving force, and how their innovation, iteration and integration drive the evolution of Web3.
JinseFinanceAI+Crypto becomes the main rising wave of this bull market, and the narrative logic will not change
JinseFinanceWhether this model can be successful or not, we still have to leave time to verify, but at least this trend and sign is worth our attention.
JinseFinanceThe introduction of AI Agent is gradually changing the Web3 ecosystem and promoting decision-making and community building to become more efficient and fair.
JinseFinanceHow does Web3 help the AI Agent track where top VCs are vying to bet?
JinseFinanceA DeFi Summer-style Web3 AI Summer is coming, and AI Agent gives it unlimited possibilities.
JinseFinanceThe second half of the AI big model is welcoming the explosion of AI Agent, and Web3 x AI Agent will become an important narrative in the future.
JinseFinanceJinseFinanceJinseFinance