YC Spring Startup Guide Released: Besides AI Agents, What Other Tracks Could Become the Next Big Thing?

Original Article Title: The Agentic Economy: Web2 & Web3 AI Agent Trends
Original Article Author: Defi0xJeff, Founder of steak studio
Original Article Translation: Ashley, BlockBeats

Editor's Note: The article explores the development trends of Web2 and Web3 AI Agents, based on the "Request for Startups" released by Y Combinator in Spring 2025, outlining six main tracks: AI Open Source Software, Agent Development Tools, Vertical AI Agents, Personal AI Assistants, AI App Stores, and B2A (Business-to-Agent). Web3 empowers AI Agents with decentralization, trustlessness, and composability, driving the rise of the AI + Crypto ecosystem.

The following is the original content (reorganized for easier comprehension):

Web2 & Web3 AI Agent Trends

Y Combinator recently released the "Request for Startups" for Spring 2025, listing the entrepreneurial directions they hope more founders will explore. Many of these directions reveal emerging trends in the Web2 AI Agent field, addressing a range of real-world pain points:

· AI App Stores

· Data Centers

· Compliance and Audit

· DocuSign 2.0

· Browser and Computer Automation

· Personal AI Assistants

· AI Agent Development Tools

· The Future of Software Engineering (Engineering Agents)

· AI Business Open Source Software

· Hardware-Optimized Code Agents

· Business-to-Agent (B2A)

· Vertical AI Agents

· Inferencing AI Infrastructure

For readers who want to delve deeper, I highly recommend checking out the original post from @ycombinator:

There is a wealth of information here, but if you are already deeply involved in this field, you may have seen many Web3 Agent teams making moves in these tracks. In my view, the following tracks will lead the trends of Web3 AI Agents (in no particular order):

· AI Business Open Source Software

· AI Agent Development Tool

· Vertical AI Agent

· Personal AI Assistant

· AI App Store

· B2A (Business-to-Agent)

AI Business Open Source Software

Web3 AI is deeply integrated with open-source AI, making it naturally suitable for this race. @ai16zdao was the first to drive one of the largest open-source AI movements—the ElizaOS framework, which currently has 14k stars and 4,227 forks on GitHub. Despite market fluctuations, adoption rates continue to rise.

This trend motivates Web3 developers to open source their technology, making AI Agent technology and frameworks more open, and driving developers to collaborate more efficiently. For example, in addition to ElizaOS, we have also witnessed the rise of frameworks such as @arcdotfun, @GAME_Virtuals, @sendaifun, @pippinlovesyou, and @freysa_ai.

As AI Agent technology evolves—with the release of OpenAI's o3, the launch of new DeepSeek models, and tech giants accelerating the deployment of AI Agents—the demand for open-source AI and Web3 AI is growing rapidly. AI x Crypto may eventually claim a significant share of the AI market.

AI Agent Development Tool

Building an AI Agent is not just about creating intelligent models, but more importantly, providing developers with the right tools and infrastructure to build, deploy, and manage AI Agents more efficiently. As AI Agents become increasingly complex, the demand for developer-friendly tools, frameworks, and platforms is soaring.

A plethora of AI development tools has already emerged in Web2, and Web3 is taking this trend to a new level—integrating decentralization, trustlessness, and open collaboration, breaking free from the closed ecosystems controlled by traditional big tech companies.

This trend has spurred AI developer platforms, Agent ecosystems, as well as no-code/low-code build tools, reducing the barriers to entry for AI Agent development.

In Web3, we are seeing an increasing number of AI Agent development tools emerge, such as:

@ai16zdao: ElizaOS, with the most plugins and integrations

@sendaifun (Solana Agent Kit) & @CoinbaseDev (CDP Agent Kit): Providing core development tools for on-chain AI Agents

@autonolas's Pearl: Focused on market prediction, DeFi automation, and self-executing Agent in the AI Agent App Store

@AlloraNetwork: Providing machine learning infrastructure for AI Agents to enhance real-time prediction capabilities

@cookiedotfun: AI Agent-driven data analytics platform, assisting Agents in extracting social sentiment insights from on-chain and off-chain data

@getmasafi: Real-time data streaming solution, offering the latest intelligence to AI Agents

Some no-code AI platforms focusing on Web3 include:

@virtuals_io: A leading no-code/low-code AI Agent builder, helping developers quickly realize AI Agent products

@HoloworldAI: A no-code building tool for 3D audiovisual AI Agents

@Cod3xOrg: A no-code platform designed for automated trading AI Agents

@Almanak__: A platform specializing in institutional-grade quant Agent construction

@EliteAgents_AI: A plugin-enhanced AI Agent integrating with ElizaOS, G.A.M.E, and other AI ecosystems

The ecosystem of Web3 AI Agent development tools is still in its early stages, but the infrastructure is rapidly maturing. Over the next few years, we can expect a fully decentralized AI development ecosystem where AI Agents will not only be easy to build but also capable of achieving full autonomy, scalability, and profitability.

Vertical AI Agent

AI Agents are no longer general-purpose tools but are gradually evolving into highly specialized vertical domain Agents capable of handling complex tasks, even making decisions and executing them.

We have already seen the rise of AI Agents in fields such as finance, law, research, providing enhanced analytical, advisory, and execution capabilities, for example:

· Tax Agent: Calculate, optimize, and execute tax-saving strategies

· Legal Agent: Review contracts, identify unfavorable terms, and provide alternative suggestions

· Financial Agent: Analyze financial statements, interpret macroeconomic trends, generate investment insights

Unlike Web2, Web3's vertical AI Agents focus more on autonomy, decentralization, and on-chain integration, avoiding the issue of centralized data silos, enhancing transparency and trust.

In the Crypto space, community interaction and personalization are crucial, and Web3 AI Agents are evolving towards being more personalized and interactive. For example:

· AI Influencer: @aixbt_Agent, sharing market insights and Alpha opportunities on Crypto Twitter

· Token Analysis Agent: @unit00x0, @kwantxbt, @tri_sigma_, @mobyAgent, @_AgentScarlett

· Research Agent: @DV_Memetics, @S4mmyEth, providing on-chain intelligence analysis

· DeFAI Agent: @Cod3xOrg, @gizatechxyz, @autonolas, managing LP, yield farming, and automated trading strategies

Web3-native AI Agents are rapidly evolving, with future DeFAI Agents expected to streamline DeFi complexity, attracting the next billion users into the AI ecosystem.

Personal AI Assistant

Personal AI Assistants are fundamentally transforming how we handle daily tasks, with previously unimaginable convenience and automation becoming a reality. These AI Assistants are not just simple reminder tools or schedule managers but are capable of proactively making decisions for users, optimizing time and resource as intelligent Agents.

Imagine an AI that not only helps you book travel but also selects the best restaurant based on your taste, monitors traffic conditions in real-time, and automatically adjusts meeting schedules if you are running late. The AI can summarize all meeting content, provide follow-up action recommendations, and even automatically book transportation. Additionally, it can organize your photos, categorize them by location and event, and create curated albums for easy viewing anytime.

In the Web3 space, this concept has been further extended, for example:

· Airdrop Agent: Scans all your wallets to determine if you qualify for airdrops from projects such as @berachain, @monad_xyz, @StoryProtocol, etc. (if someone is doing this, please contact me lol).

· Yield Farming & LP Management Agent: Tracks and rebalances DeFi holdings, automatically claims rewards, and compounds interest using the best strategies.

· AI-Driven GitHub Codebase Analysis: For example, @soleng_Agent, can assess whether a project has a strong development team or if it might be a scam.

· Automated Trading Agent: Like @Cod3xOrg and @Almanak__, can execute predefined trading strategies, accurately time market entries and exits, and optimize returns.

The next stage of development for AI personal assistants will evolve from "passive response" to a true autonomous AI co-pilot. They will not merely be tools but intelligent Agents with autonomy and decision-making abilities, able to proactively anticipate user needs and complete complex multi-step tasks with minimal input.

In this transformation, the role of Web3 is crucial: decentralized AI Agents will have trustlessness, transparency, and censorship resistance, ensuring users have full control over their AI-driven workflows. In the future, users will not only delegate simple tasks to AI but may even entrust financial and operational decisions to AI Agents, undoubtedly changing the rules of the game.

AI App Store

The AI App Store is one of the most exciting and inevitable developments in the field of artificial intelligence. Just as mobile apps have dedicated app stores, AI Agents also need a specialized marketplace for users to easily discover, purchase, and integrate AI-driven applications.

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

The Agent Distribution Network is responsible for attracting builders, investors, and users into the AI Agent ecosystem. One of the most representative cases is @virtuals_io, which is building an Agent Society to allow various AI Agents to collaborate and interact within it.

The MAO Network ensures that AI applications can accurately match user needs, intelligently orchestrate Agents to maximize their value. In the future, users will not need to manually search for AI applications, but only need to express their needs, and the system will automatically recommend suitable AI Agents, and even dynamically combine multiple Agents to provide a complete solution.

This means that the Web3 AI Application Store is not just an application marketplace; it must also have key functions such as content filtering, privacy protection, and decentralized Agent interaction to ensure seamless collaboration between AI Agents.

Key players driving the development of the Web3 AI Application Store:

@virtuals_io: Expanding its vision of the Agent Society, introducing high-quality AI Agent teams, and driving the development of inter-Agent communication protocols.

@santavirtuals and @questflow: Optimizing coordination between Virtuals Agents to improve resource allocation efficiency.

Abstraction Layers:

@orbitcryptoai and @HeyAnonai: By integrating abstraction layers, making AI Agents more easily accessible and interactive with DeFi.

The orchestration of AI Agents is still in its early stages, but it can be certain that the ability to seamlessly operate and monetize AI Agents will become a vast market, and Web3 is actively positioning itself to capture a significant share of this emerging market.

B2A (Business-to-Agent)

AI Agents are not just tools - they are becoming active participants in the digital economy, able to autonomously trade, manage resources, and even collaborate with other AI Agents. This transformation requires a new infrastructure that serves not only human users but also treats AI Agents as clients. This is the core of the B2A (Business-to-Agent) model.

Just as SaaS (Software-as-a-Service) transformed enterprise operations, B2A will define how AI Agents interact, transact, and operate in the digital economy. In the future, AI Agents will need their own payment solutions, data access rights, computing resources, and even privacy frameworks. Currently, several Web3 projects are building the B2A ecosystem.

· AI Business Payments: @Nevermined_io is developing an AI Agent-native payment solution, aiming to become the "PayPal for AI Agents".

· Compute Power Management: @hyperbolic_labs is researching a self-sustaining AI Agent to enable it to autonomously manage computing resources.

· Privacy and Security Infrastructure: @PhalaNetwork, @OraProtocol, and @brevis_zk are building a privacy-preserving computing layer dedicated to AI Agents, ensuring secure, verifiable interactions.

· High-Quality Data Access: @getgrass_io, @vana, @getmasafi, and @cookiedotfun provide structured, high-quality data sources to assist AI Agents in training, learning, and efficient operation.

· Agent-to-Agent Communication: @virtuals_io is developing an Agent-to-Agent communication protocol to enable AI Agents to collaborate with each other.

· AI Intellectual Property Management: @StoryProtocol is constructing an AI-generated content management framework similar to TCP/IP, allowing AI Agents to autonomously manage and license the content they create.

B2A is not just a theoretical concept; it is actively being built. As the capabilities of AI Agents continue to grow, they will require dedicated economic infrastructure to operate independently within the digital ecosystem. If you have not yet thought about how to serve AI Agents in this market, you may already be behind.

Conclusion

AI Agents are reshaping our interactions, constructions, and automation in the Web2 and Web3 worlds. With the rise of a Web3-native AI ecosystem, we are entering a decentralized AI Agent-driven era, bringing new paradigms such as open-source collaboration, Agent economies, and decentralized automation.

The fusion of AI and Crypto is still in its early stages, but its growth trajectory is undeniable. Web3 empowers AI Agents with capabilities that Web2 cannot offer: self-ownership, permissionless innovation, and a fully composable ecosystem. The question is no longer whether AI Agents will reshape Web3 but how quickly this transformation will occur and which tracks will define the future.

With the expansion of the AI Agent economy, one thing is certain: whether you are a developer, investor, or just a curious onlooker, now is the best time to pay attention to this trend.

Infrastructure is being built, industry players are emerging, and the market opportunity is huge.

The only question is: are you ready to join?

「Original Post Link」