The Rise of the Agent Economy: Trends in AI Agents in Web2 and Web3

Title of the original text: The Agentic Economy: Web2 & Web3 AI Agent Trends

Author of the original text: 0xJeff (@Defi 0xJeff)

Compiled by: Asher (@Asher_ 0210 )

Y Combinator (Y Combinator, also known as YC, is one of the most well-known startup accelerators in the world, headquartered in Silicon Valley, USA. Its incubated companies include Airbnb, Stripe, Dropbox, Reddit, Coinbase, OpenAI, and others, with extremely high success rates and influence) has released its "Request for Startups" for the Spring of 2025, listing the ideas they would like to see more founders explore. Many of these ideas indicate that the application of AI agents in the Web2 domain is becoming an emerging trend, aiming to solve real pain points.

In my opinion, the areas that will shape the trends of Web3 AI agents are: AI commercial open source software, development tools for AI agents, vertical AI agents, AI personal assistants, AI app stores, and B2A.

AI Commercial Open Source Software

Web3 AI is closely related to open source AI, making it a natural fit for this field. ai16z DAO has led one of the largest open source AI movements, with its ElizaOS framework currently having 14,000 stars and 4,227 code forks on GitHub. Despite market fluctuations, adoption is still on the rise.

This movement has inspired Web3 developers to open source their technologies, driving teams to build AI technologies and frameworks, making it easier for other developers to use and collaborate than ever before. In addition to ElizaOS, we have also witnessed the rise of frameworks such as arc, GAME BY VIRTUALS, SendAI, pippin, and Freysa, collectively driving the development of the open source innovation ecosystem.

As AI agents continue to evolve, with OpenAI launching o3, DeepSeek releasing new models, and tech giants accelerating their AI agent deployments, the demand for open source AI and Web3 AI is growing rapidly. Crypto x AI may ultimately occupy a significant share of the AI market.

Devtools for AI Agents

Building AI agents is not just about creating smart models, but more importantly, providing developers with suitable tools and infrastructure to efficiently realize the deployment of these agents. As the complexity of AI agents continues to increase, the demand for developer-friendly tools, frameworks, and platforms to support their construction, deployment, and management is growing rapidly.

In the Web2 era, a large number of development tools that enhance AI capabilities have already emerged, and Web3 is further accelerating this process. By introducing decentralization, trustless mechanisms, and open-source collaboration, Web3 breaks the constraints of traditional closed ecosystems, allowing the development and deployment of AI agents to no longer depend on the systems controlled by a few tech giants.

This trend has driven the rise of AI-focused development platforms, agent ecosystems, and no-code/low-code building tools, significantly lowering the creation threshold for AI agents and enabling more developers to easily participate, accelerating the innovation and popularization of AI technology.

In the Web3 domain, more and more platforms are starting to provide AI agent development toolkits, allowing developers to more easily create and monetize AI-driven applications. Some noteworthy examples include:

• ai16z DAO: ElizaOS, with the most plugins and integration support;

• SendAI (Solana Agent Kit) and Coinbase Developer Platform (CDP Agent Kit): These toolkits provide developers with the basic components for building on-chain AI agents;

• Olas ecosystem's Pearl: An AI agent app store focused on practical functions, covering areas such as prediction markets, DeFi automation, and autonomous execution agents;

• Allora: Providing machine learning infrastructure to help AI agents make more accurate real-time predictions;

• Cookie DAO: Focused on AI agent-driven data analysis, helping AI agents extract social sentiment insights from on-chain and off-chain data;

• Masa: Providing real-time data streaming solutions to provide AI agents with the latest intelligent information.

Some no-code AI platforms focused on Web3 include:

• Virtuals Protocol: A leading no-code/low-code AI agent building and launching platform, helping developers turn AI agent concepts into usable products with minimal investment;

• Holoworld AI: A no-code building tool focused on 3D audiovisual AI agents, helping users design AI-driven virtual characters;

• Cod3x: A no-code platform specifically for building autonomous trading agents, helping traders leverage AI to automate trading strategies;

• Almanak: A building platform for institutional-grade quantitative agents, focused on advanced financial use cases;

• Elite Agents: Focused on plugin-enhanced AI agents, and integrated with ecosystems like ElizaOS, G.A.M.E, and others.

The development tool ecosystem for Web3 AI is still in its early stages, but the infrastructure is being rapidly built and improved. With the continuous advancement of technology, a fully decentralized AI development ecosystem is expected to emerge in the coming years. In this ecosystem, AI agents will not only become easier to build, but also possess complete autonomy, scalability, and monetization capabilities. One of the key factors driving this transformation will be the tools that provide support for developers, which will become the most valuable infrastructure in the Web3 AI economy.

Vertical AI Agents

AI agents are evolving from simple task execution tools to highly specialized, industry-specific intelligent agents capable of handling complex and detailed business operations. These agents, by leveraging domain expertise, go beyond basic automation functions and become decision-making entities capable of executing tasks that typically require deep human professional knowledge.

With this development, a wave of AI-driven industry verticalization is emerging, covering areas such as finance, law, and research. As the capabilities of AI agents continue to grow, they not only can analyze and recommend solutions, but also represent users to execute decisions and operations, driving deep transformations across various industries.

Some notable examples of vertical AI agents include:

• Tax agents: Able to calculate, optimize, and execute tax-saving strategies;

• Legal agents: Able to review contracts, detect unfavorable terms, and propose more favorable alternatives (even representing you in legal disputes);

• Financial agents: Able to analyze financial statements, interpret macroeconomic trends, and generate investment insights.

The key difference between Web3 and Web2 in the application of vertical AI agents lies in the emphasis on autonomy, decentralization, and on-chain integration. Unlike traditional AI services that rely on centralized data silos, Web3 native AI agents have on-chain verifiability, enabling them to provide higher transparency and trustworthiness.

Furthermore, in the Web3 domain, community interaction and personality are crucial, which also affects the development direction of Web3 AI agents. Unlike the typically impersonal and purely functional AI agents in Web2, Web3 AI agents are gradually developing unique personalities and interaction patterns to better fit the culture of decentralized communities. Here are some typical examples:

• AI influencers, such as aixbt, who share insights and investment information by analyzing crypto-related content on the X platform;

• Token analysis agents, such as Rei, kwantxbt, 3σ, Moby AI, and Agent Scarlett;

• Research agents, such as Deep Value Memetics and s4mmy, who provide actionable intelligence through Orbit;

• DeFAI agents, managing LP provision, yield farming, and trading strategies, developed by teams such as Cod3x, Giza, and Olas.

As AI model platforms like Nous Research, Bagel, and Pond continue to enhance the personality of agents, the application scenarios of Web3-native AI are rapidly evolving. DeFAI agents simplify the complexity of DeFi and guide the next wave of billions of users' adoption, potentially becoming the next major wave of AI adoption.

AI Personal Staff

AI personal assistants are revolutionizing the way people handle daily tasks, bringing unprecedented convenience and automation. The functionality of these assistants will no longer be limited to simple reminders and scheduling, as they will proactively make decisions for users, optimizing the use of time and resources.

For example, an AI can not only book a trip, but also recommend the best restaurants based on the user's preferences, check traffic conditions, and even reschedule meetings if the user is running late. It can also summarize all meetings, suggest follow-up actions, and even book transportation when needed. Additionally, the AI will automatically organize photos, tag locations and events, and create convenient memory albums for the user to access at any time.

In the Web3 ecosystem, the applications of AI personal assistants will further expand:

• Airdrop agents: scanning user wallets and determining eligibility for upcoming airdrops;

• Yield farming and LP management agents: automatically tracking and rebalancing DeFi positions, claiming rewards, and compounding under the best strategies;

• GitHub repository analysis agents, such as SOLENG, evaluating the strength of a project's development team or whether a project is likely to be a scam;

• Automated trading agents, such as Cod3x and Almanak, automatically entering and exiting the market based on preset trading conditions, optimizing profits, and adjusting to market changes.

The next evolution of AI personal assistants will be fully autonomous intelligent agents, which will not only be assistants but also active partners capable of taking initiative. As AI models' reasoning and decision-making capabilities continue to improve, these agents will shift from passive responsiveness to proactively predicting user needs and executing complex multi-step tasks with minimal human intervention.

Web3 plays a crucial role in this transformation: decentralized AI agents have the attributes of trust, transparency, and censorship resistance, ensuring that users can fully control their AI-driven workflows. By automating decision-making and task execution, especially for financial and operational decisions, this will significantly change the existing ways of working.

AI App Store

The AI app store is one of the most exciting and inevitable advancements in the AI field. Just as mobile apps have their own app stores, AI agents also need a dedicated marketplace where users can discover, purchase, and seamlessly integrate AI-driven applications.

In Web3, this concept is evolving into a combination of a Multi-Agent Orchestration (MAO) Network and an Agent Distribution Network:

• The Agent Distribution Network drives the construction of the market by attracting builders, investors, and users into the ecosystem. Virtuals Protocol is a typical example of this model, as it is building a society of agents where different AI agents can live and interact;

• The MAO Network ensures that AI applications can precisely match user needs, orchestrating agents to efficiently deliver value. Users no longer need to manually search for applications; they can simply state their requirements, and the appropriate AI agents will be recommended, or even instantly assembled into a solution.

Therefore, the Web3 AI app store is not just a marketplace; it must also curate and vet the applications, ensure privacy, and facilitate seamless interactions between agents.

Key participants are driving this progress, including:

• Virtuals Protocol, expanding its vision of an agent society, attracting high-quality agent teams, and developing agent-to-agent communication protocols;

• Santa by Virtuals and Questflow, optimizing the coordination of Virtuals agents to improve resource allocation efficiency;

• Orbit and Hey Anon as abstraction layers, helping to integrate AI agents with DeFi and enhance their accessibility.

While AI orchestration is still in its early stages, it is evident that the ability to seamlessly operate and monetize AI agents will become a massive market, and Web3 is poised to capture a significant share of it.

B2A (Business-to-Agent)

AI agents are no longer just tools; they are gradually becoming active economic participants, capable of trading, managing resources, and even autonomously collaborating with other AI agents. This transformation requires a new infrastructure that not only serves humans but also serves AI agents as customers. This is the concept of B2A.

Just as SaaS (Software as a Service) has transformed the way businesses operate, B2A will define the interaction, transaction, and operation of AI agents in the digital economy. AI agents will need their own payment solutions, data access, computing capabilities, and even privacy frameworks. Some Web3 projects are already paving the way for this:

• AI business payments: Nevermined is developing agent-native payment solutions, effectively becoming the "PayPal for AI agents";

• Computing management: Hyperbolic is developing self-sustaining agents capable of efficiently managing their computing resources;

• Privacy and security infrastructure: Phala Network, ORA, and Brevis are building privacy-preserving computing layers for AI agents to ensure secure and verifiable interactions;

• High-quality data access: Grass, vana, Masa, and Cookie DAO are providing structured, high-quality data sources for AI agents to effectively train, learn, and operate;

• Agent-to-agent communication: Virtuals Protocol is building agent-to-agent communication protocols, enabling AI agents to collaborate with each other.

B2A is not just a theoretical concept, it is actively being built. As AI agents become more sophisticated, they will require dedicated infrastructure to operate independently within the economic ecosystem. If you have not yet considered how to serve AI agents as a market, you are already behind.

Summary

AI agents are shaping our interactions, constructions, and automations in Web2 and Web3. With the rise of a native Web3 AI ecosystem, they bring a new paradigm that drives open-source collaboration, agent-driven commerce, and decentralized automation.

While the fusion of AI and crypto is still in its early stages, the momentum of this trend cannot be ignored. Unlike Web2, Web3 offers unparalleled advantages for AI agents: ownership, permissionless innovation, and a fully composable ecosystem. The question is no longer whether AI agents will reshape Web3, but rather how quickly this transformation will occur and which industries will become the leaders of the future.

As the agent-driven economy continues to expand, the opportunities of the future are becoming increasingly vast. Whether you are a developer, investor, or curious observer, now is the best time to focus on this field. Infrastructure is being rapidly built, key players are emerging, and potential opportunities are ubiquitous.

The only question is: will you be a part of it?