AI and Web3 Integration: Insights from the Hong Kong Consensus Conference 2025

AI and Web3, as the two most prominent technology fields today, are continuously driving human society towards new technological peaks. With the revolutionary AI experience brought by ChatGPT, on-chain AI is gradually moving from concept to practice, becoming one of the most promising emerging tracks in the Web3 field.

At the recently concluded Hong Kong Consensus Conference 2025, the integration of AI and Web3 became a hot topic, with related discussions permeating the entire event. Let's take a look back at the exciting viewpoints and cutting-edge projects regarding AI x Web3 that were presented at this conference.

1. AI Infrastructure

1. AI Agent platform and framework

In the past six months, the launch platforms and framework-based infrastructure for AI Agents have been very popular. These projects provide developers and ordinary users with low-threshold platforms to use AI Agents, making them one of the key directions for current AI projects.

• 0G Labs: The first decentralized artificial intelligence operating system ( deAIOS ), by building an AI dedicated Layer 1, connects computing resources, data, and models to create a distributed AI development ecosystem.

• DeAgentAI: An innovative platform focused on decentralized AI Agents, dedicated to promoting the development of multi-agent technology. Users can create, manage, and coordinate AI Agent networks for applications in business automation, data analysis, and other scenarios.

• Autonomys Network: A decentralized infrastructure stack that enables secure and autonomous human-machine collaboration. Users can create dedicated AI agents to perform tasks such as reservation services and fund management.

• Gaia Network: A decentralized AI infrastructure platform that supports distributed development and operation of AI Agents and applications, integrating distributed storage, computing, and data verification through blockchain.

• Questflow: A decentralized multi-AI Agent network, where users describe their needs and the AI agent network can autonomously complete tasks, leveraging the advantages of collective intelligence.

2. Decentralized AI

Decentralized AI is the ultimate goal of on-chain AI. Currently, multiple projects are working on computing power, data, models, and other directions, hoping to break the monopoly of large companies on LLM through decentralization and help the public gain ownership of data and models.

• Vana: A decentralized user data sovereignty platform that transforms personal data into financial assets through Data Liquidity Pools.

• Hyperbolic: An open-access AI cloud platform that integrates global computing resources, providing affordable GPU resources and AI services.

• OpenLedger: A next-generation network focused on AI and blockchain, supporting developers in obtaining high-quality data, fine-tuning dedicated language models, and deploying them as paid services.

• IO.NET: A decentralized computing platform that provides on-demand access to GPU and CPU cluster services, eliminating the need for users to purchase expensive hardware.

• Aethir: A distributed cloud computing infrastructure platform, including Aethir Earth for AI-specific computing and Aethir Atmosphere optimized for gaming.

• MinionLab: Decentralized Autonomous AI Agent Network for real-time mining of internet data, where device owners can receive token rewards.

• GAIB: AI and high-performance computing economic layer solutions, viewing GPUs as assets and computing power as currency.

• Kite AI: A decentralized Layer 1 blockchain platform designed for the AI economy, achieving fair access and rewards through the Proof of AI mechanism.

• Automata: Provides middleware privacy protection and non-tracking computing capabilities for decentralized applications.

• Public AI: An open and transparent AI data platform that supports multimodal data collection and annotation, utilizing the Proof of AI consensus mechanism.

3. Verifiable AI

One of the important challenges facing AI development is the opacity of the training process and the inability to guarantee the accuracy of the results. Some projects are attempting to achieve verifiability of the AI training process through technologies like ZKP and TEE, ensuring the reliability of the output results.

• Phala Network: A decentralized cloud computing platform that provides trusted privacy computing and AI inference services for on-chain applications.

• Brevis: Decentralized computing engine that provides verifiable off-chain AI and blockchain computation, combining zero-knowledge proofs to enhance privacy and efficiency.

• Verisense Network: A platform focused on decentralized data verification and trusted AI innovations, helping to verify data sources and AI decision-making processes.

2. AI Use Cases: Potential and Expectations

Currently, there are relatively few standout AI use case projects, but some emerging projects provide more possibilities for the application of AI Agents:

• Narra: The Gamefi AI Agent platform on Berachain generates real-time dynamic narrative content and supports the creation and interaction of AI-NFTs.

• AI Travel: An AI-driven travel assistant that can automatically customize travel plans, book hotels, and provide price comparison services.

• HeyTracyAI: An AI Agent in the basketball field featuring NBA champion Tristan Thompson, providing real-time analysis and predictive insights.

• AskJimmy: An AI Agent platform focused on the finance and trading sector, aiming to create a decentralized multi-strategy hedge fund operated autonomously by AI Agents.

3. Traditional Projects Transforming to AI

Many traditional Web3 projects are also starting to embrace AI, announcing their respective AI pivot plans:

• Public chains like Sui, Near, Flow, and Aptos actively participate in AI-related conferences, indicating their comprehensive support for AI development from aspects such as underlying architecture and account innovation.

• Eigenlayer is working hard to build a decentralized trust layer and verifiable cloud services, providing on-chain proof for off-chain computations such as AI training, inference, and prediction.

4. Challenges and Future

Despite the bright prospects, the development of on-chain AI still faces numerous challenges, including insufficient model reliability, ambiguous prompt intentions, storage and hardware limitations, and privacy security issues. These challenges are both technical difficulties and give rise to enormous opportunities for innovation. In the long run, the industry is full of hope for the development of on-chain AI, looking forward to jointly promoting the integration and prosperity of AI and Web3 through improved infrastructure, innovative use cases, and community collaboration.