From the Hong Kong Consensus conference, we can see the potential projects and directions of AI x Web3 at present.

Author: Biteye core contributor Anci

AI and web3 are the two recognized drivers that have propelled humanity into the next technological growth point in recent years. After bringing disruptive AI experiences with ChatGPT, on-chain AI has quietly made a splendid transformation from memes to infrastructure, becoming the most certain and sustainable new track for web3 enthusiasts.

At the just-concluded Hong Kong Consensus Conference 2025, the integration of AI and web3 was also a hot topic. Discussions about AI x web3 were heard everywhere, whether at the main venue or the sub-venue. The Biteye team delved into it, understanding many promising AI projects, and brought frontline sharing to everyone.

1. AI Infrastructure

2. AI Agent launch platform and framework

With the popularity of $AI16Z and $VIRTUAL at the beginning of the year, the construction of AI Agent launch platforms and framework-like AI infrastructure has been very popular in the past half year. These projects provide a low-threshold platform for developers and ordinary users to own and use AI Agents, which is one of the key directions of this round of AI projects.

0G Labs

It is the first decentralized artificial intelligence operating system (deAIOS), which connects computing resources, data, and models through building an AI-specific Layer 1, and constructs a distributed AI development ecosystem. Developers can use the modular tools provided by 0G to build a more efficient and transparent AI Agent, while obtaining fair reward distribution. Users can contribute computing resources to the decentralized network and obtain profits by running nodes.

DeAgentAI

It is an innovative platform focusing on decentralized AI Agent and is committed to promoting the development of Multi-Agent Systems. Through DeAgentAI, users can create, manage, and coordinate AI Agent networks for various scenarios, such as business automation, data analysis, personalized recommendations, etc.

Autonomys Network

It is a decentralized infrastructure stack aimed at achieving secure, autonomous human-machine collaboration. Anyone can join the network as a node with just a solid-state drive (SSD). Users can create their own AI agents that can act autonomously, such as booking services, managing funds, coordinating meetings, and even participating in on-chain governance.

Gaia Network

It is a decentralized AI infrastructure platform that supports the distributed development and operation of AI agents and applications. By integrating blockchain technology with distributed storage, computing, and data verification, it addresses the privacy, scalability, and accessibility issues of AI. GaiaNet does not establish centralized servers but builds a distributed network of edge computing nodes controlled by individuals and enterprises. It hosts fine-tuned artificial intelligence models based on operators’ proprietary domain knowledge and expertise.

Questflow

A decentralized network composed of multiple AI Agents, where users only need to describe their needs, and the AI agent network can autonomously complete tasks. Multiple AI agents can seamlessly execute tasks, and utilize the collective intelligence and capabilities of these AI Agents to achieve efficiency and speed that cannot be compared to a single AI model or human.

2. Decentralized AI

Decentralized AI is the ultimate goal of on-chain AI. Currently, many projects are continuously striving towards this goal in terms of computing power, data, models, etc., hoping to break the monopoly of large companies on LLM through decentralized means and help the public gain ownership of data and models.

Vana

Dedicated to building a decentralized user data sovereignty platform, turning personal data into financial assets. Vana is an EVM compatible Layer 1 blockchain that allows users to contribute and validate data through its unique design of Data Liquidity Pools (DLP), generating tokenized assets for AI model training. It ensures data security and quality through Proof of Contribution and privacy protection technology. Yesterday, it received investment from YZi Lab, with CZ serving as an advisor.

Hyperbolic

Is an open access AI cloud platform that integrates global computing resources to provide users with affordable, scalable GPU resources and AI services. The platform supports AI inference services, on-demand GPU access, and monetization of idle hardware, helping businesses, researchers, and individuals reduce costs and efficiently utilize computing resources. Hyperbolic provides services for well-known on-chain AI projects such as Virtuals and AI16z, and revealed at this conference that it will launch its own AI Agent framework to further promote innovation and application of AI computing.

OpenLedger

It is a next-generation network focusing on AI and blockchain, providing decentralized economic infrastructure to support developers in accessing high-quality data, fine-tuning Specialized Language Models (SLMs), and deploying them as paid services. SLM is its core advantage, designed for blockchain, DeFi, smart contracts, and other fields; while leveraging the Open LoRA framework to efficiently run thousands of fine-tuned models on a single GPU, reducing costs and maintaining high throughput. Combined with Datanets data network and Model Factory, OpenLedger offers a complete solution for data aggregation, model optimization, and secure deployment.

IO.NET

It is a decentralized computing platform that provides on-demand access to GPU and CPU clusters, aiming to eliminate the need for users to have expensive hardware or infrastructure. Through a decentralized physical infrastructure network (IOG Network, consisting of independently operated distributed hardware nodes), users can instantly access distributed GPU clusters through IO Workers, suitable for high-performance computing, AI model training, and other scenarios, without the need to purchase expensive hardware.

Aethir

Is an innovative platform dedicated to providing distributed cloud computing infrastructure, where Aethir Earth is a bare metal GPU cloud designed for AI computing tasks, providing high-performance computing capabilities; Aethir Atmosphere is a cloud GPU network optimized for the gaming industry, supporting low latency, high-quality gaming experience. Users can also add their own GPU to the Aethir platform, become contributors and receive benefits.

MinionLab

Is a decentralized autonomous AI intelligent agent network, these intelligent agents are called ‘Minions’, running on user devices, used to mine data from the internet in real-time, device owners can obtain MINION by supporting the network and running these autonomous data mining intelligent agents.

Token rewards. By leveraging idle computing resources, MinionLab offers a more flexible, cost-effective, and scalable way of data mining, while driving the AI industry towards decentralization and user-driven development.

GAIB

It is an economic layer solution dedicated to the field of AI and high-performance computing, aimed at creating a new asset class and economic system through financialization and tokenization of GPU resources. Its core idea is to view GPUs as assets and computing power as currency. GAIB solves the current challenges of computing resource allocation and financing through decentralized computing networks, fractional ownership, and blockchain technology (L1/L2), while meeting the huge demand for high-performance computing resources in the AI era.

Kite AI

It is a decentralized Layer 1 blockchain platform designed for the artificial intelligence economy. It aims to unlock fair access to AI assets (data, models, and agents) and rewards through its innovative Proof of AI (PoAI) consensus mechanism. Kite AI is committed to building an open, collaborative ecosystem where data providers, AI developers, and end users can all benefit from the decentralized AI economy.

Automata

Aiming to provide an intermediate layer of privacy protection and untraceable computing functions for decentralized applications (DApps). It offers developers and users an efficient, trustworthy, and decentralized environment, maximizing data privacy and user autonomy.

Public AI

The vision is to create an open and transparent AI data platform, supporting the collection and annotation of multimodal data, including text, audio, video, and map data. At the same time, it provides efficient and low-cost AI-assisted data annotation services. The AI assistant will perform preliminary annotations on the data, which will then be comprehensively reviewed by AI validators to ensure data quality. Adopting a fair and transparent Proof of AI consensus mechanism, allowing everyone to participate in the AI training process by contributing data.

3. Verifiable AI

One of the important challenges facing the development of AI is the opacity of the training process and the inability to guarantee the accuracy of AI output results. Currently, many projects hope to achieve the verifiability of the AI training process through technologies such as ZKP, TEE, to ensure the reliability of AI output results.

Phala Network

It is a decentralized cloud computing platform that provides trusted privacy computing and AI inference services for on-chain applications. Its product services include a confidential computing network based on Trusted Execution Environment (TEE), supporting smart contracts, AI model training and inference, as well as data privacy protection. Phala Network has a natural advantage in high-privacy scenarios such as DeFi, DeSci, NFT markets, and on-chain AI Agent development, and can ensure data security and computational verifiability through TEE technology.

Brevis

It is a decentralized computing engine, positioning itself to provide verifiable off-chain AI and blockchain computing, combining zero-knowledge proofs (ZKP) to enhance privacy and efficiency, focusing on financial data privacy protection, supporting the extension of smart contracts and Web3 applications. Its product offers a secure computing environment, ensuring the privacy of on-chain AI inference and transactions while maintaining verifiability, suitable for DeFi, prediction markets, and privacy-driven AI inference.

Verisense Network

It is an innovative platform focusing on decentralized data verification and trustworthy AI, helping developers verify the source of data to ensure the authenticity and integrity of training data; at the same time, it supports the audit and verification of AI decision-making processes, providing transparent training data and reasoning processes for AI models, eliminating the ‘black box’ problem and making AI decisions more trustworthy. Suitable for serving AI models in DeFi, prediction markets, and compliance monitoring, it can ensure the reliability of on-chain data.

Two, AI Use Cases: Potential and Expectations

Compared to the rich AI infrastructure, there are still relatively few outstanding AI practical use case projects at present. In addition to the familiar Twitter bot AIXBT, the speakers at this event also mentioned Virtuals ecosystem’s @BukProtocol, which can realize travel booking business, NBA star commentator @HeyTracyAI, and DeFAI Agent @askjimmy_ai under the ARC framework, providing more possibilities for the application of AI Agents.

Narra

The Gamefi AI Agent platform on Berachain, which utilizes Narra’s AI engine to generate real-time dynamic narrative content, interact with players, drive story development, and provide personalized experiences in gaming scenarios. In addition, Narra also supports the generation and interaction of AI-NFTs, which are not only digital assets but also have intelligent features for dynamic interaction with users.

AI Travel

Is an AI-driven travel assistant that can help users automatically customize different travel plans through chat; it can also help users call hotels to order suitable rooms at the best price; and provide hotel flights and other activities booking and price comparison services through browser plugins.

HeyTracyAI

It is an AI agent with NBA champion Tristan Thompson focused on sports commentary in the field of basketball, which can provide real-time analysis and predictive insights for games, among other things.

AskJimmy

Is a platform that focuses on the financial and trading fields, and its main goal is to create a decentralized multi-strategy hedge fund operated independently by AI Agents.

3. Transformation of traditional projects to AI

The trend is clear, and many traditional web3 projects in this event have also embraced AI, announcing their respective AI transition plans.

Projects like Sui, Near, Flow, and Aptos actively participated in AI-related conferences at the conference. They expressed in the discussion that the emergence of AI Agents is important to simplify the complex interaction process of blockchain, which can introduce more users to the world of web3 and solve the interaction problems that have been plaguing the industry. The tokenization and cooperation of AI Agents also reveal more possibilities for AI on the chain. In response, various blockchain projects have expressed their goal of advancing AI, fully supporting the development of AI from the underlying architecture, account innovation, and encouraging developers to innovate and develop AI applications on the chain through activities such as hackathons.

Eigenlayer, which used to focus on Restaking services, also stated in its jointly organized AI sub-conference that it is striving to build a decentralized trust layer, provide verifiable cloud services, and provide on-chain proofs for AI training, reasoning, prediction, and other off-chain operations, to help the development of verifiable AI agents.

Four, Challenges and Future

As a hot topic at the Hong Kong Consensus Conference 2025, the discussion and sharing about AI and Web3 are intense and exciting. While envisioning a beautiful blueprint, many guests also mentioned that the current development of on-chain AI still faces many challenges, including insufficient model reliability, ambiguity in intent of prompt words, storage and hardware limitations, privacy security, and other issues. These challenges not only bring technical difficulties to the industry but also will foster huge innovation opportunities. In the long run, everyone is full of hope for the development of on-chain AI and looks forward to promoting the integration and prosperity of AI and Web3 through further improvement of infrastructure, innovative use cases, and community collaboration.