On October 22, the ETHShanghai 2025 main summit concluded successfully in Shanghai. Centered on the theme "Expanding Ethereum, Shaping an Open Future," the conference was co-hosted by the Chinese builder community ETHPanda, Wanxiang Blockchain Labs, PANews, and TinTinLand. The event attracted over 40 industry leaders from around the world, who delivered insightful talks and in-depth discussions on core topics such as Ethereum scaling, modular ecosystems, developer growth, and long-term security. The winners of the ETHShanghai 2025 Hackathon were also officially announced during the conference.
Xiao Feng, Vice Chairman and Executive Director of Wanxiang Holdings, Chairman of Wanxiang Blockchain, and Chairman and CEO of HashKey Group, engaged in an in-depth online conversation with Ethereum co-founder Vitalik Buterin on AI, prediction markets, and ZK ID.
Vitalik stated that during the US election, many people participated in predictions via Polymarket, which in itself is a success. Traditional and social media may not be able to judge the importance of events, while prediction markets reflect reality more quickly through price signals, becoming a new information medium. Prediction markets are not only suitable for major events; with the help of AI, they can also make accurate predictions for smaller events. In the future, the combination of AI and prediction markets will make predictions more mainstream and intelligent.
Regarding the AI sector, Vitalik suggested that blockchain defines the rules of the game, and AI can become a participant. With the rise of AI Agents, they can autonomously issue tokens, participate in applications and governance, opening up a vast experimental space. In the future, we should further consider how to empower ecosystem development. However, he emphasized that AI should not replace humans, but rather serve as a tool for humans. Security remains a key issue, including the hacking risks of Agents.
He further mentioned the potential of ZK ID (Zero-Knowledge Identity) in preventing AI abuse. ZK ID can help distinguish between real people and bots, protecting privacy while supporting the development of low-risk DeFi and other sectors.
Xiao Feng stated that blockchains can be roughly divided into two categories: one represented by bitcoin, mainly as a currency issuance system, achieving high-speed computation through simple mathematical formulas and not allowing complex external deployments, thus quickly reaching consensus globally and being regarded as "digital gold"; the other represented by Ethereum, which is application-centric and has gradually developed along the original intention of its white paper, currently occupying 60%-70% of the application market share. Xiao Feng pointed out that there is no need to try to replace Ethereum, as it has a first-mover advantage and is continuously improving. Other blockchain projects need to prove that their strategic positioning is different from Ethereum and provide differentiated value; the possibility of challenging Ethereum is very low.
Xiao Feng also emphasized that the development of DeFi cannot be ignored, but KYC and anti-money laundering requirements must be considered, as its philosophy differs from traditional finance. Through zero-knowledge identity authentication (ZK ID), users can confirm their status as qualified investors based on certificates, proofs, work experience, etc., enabling safe global transactions and allowing decentralized finance to better serve the global financial system.
In her keynote speech "Large-Scale Adoption of Ethereum: Crossing the Chasm," Ethereum Foundation Co-Executive Director Hsiao-Wei Wang stated that the vision of large-scale adoption of Ethereum can be reflected in three aspects: first, self-sovereignty, allowing users to truly own their assets; second, global settlement capability, enabling value to cross geographical boundaries, improving overall efficiency, and having the ability to verify globally; third, daily utility, making blockchain and Ethereum as naturally and smoothly integrated into people's daily lives as the internet, such as in everyday transfers.
She also pointed out that before crossing the chasm, Ethereum faces three major challenges: the high wall of scalability and cost, the experience gap, and the trust gap. In terms of scaling, Ethereum's roadmap is L1 + L2, with the core strategy being to achieve high-throughput, low-cost transactions through L2 Rollups. Key upgrades include Dencun (EIP-4844), Pectra (2025 Q1), and Fusaka (2025 Q4). In terms of accounts and user experience, the concept of account abstraction has been proposed, introducing smart accounts through proposals such as ERC-4337, EIP-7701, and EIP-7702, turning user accounts into programmable contracts and supporting features such as social recovery wallets, gas fee sponsorship, and batch transactions. On infrastructure construction, she emphasized the importance of secure and stable mainnet facilities and the normalization of financial services.
She also stated that Ethereum's future goal is to no longer be "seen," but to be silently relied upon and trusted like the internet. True large-scale adoption is not because Ethereum itself is huge, but because it is ubiquitous, transparent, and reliable. When it exists as naturally as air, that is when blockchain truly realizes its value.
Bruce Xu, co-initiator of ETHPanda, and Tomasz Stanczak, Co-Executive Director of the Ethereum Foundation, held a fireside chat. Tomasz Stanczak stated that if Ethereum becomes the coordination layer for AI Agents, in the future economic ecosystem dominated by Ethereum, the system will be even more permissionless. Everyone can deploy AI agents, participate in the local economy, and make payments by building reputation and trust, greatly reducing reliance on traditional financial systems. Agents can provide users with a set of standards to connect global infrastructure, promote project construction, support schools and building work, and improve global infrastructure through coordination. Without the Ethereum intermediary layer, agents would still need to rely on traditional payment systems and institutional verification, making the process cumbersome and control centralized.
He revealed that emerging standards such as ERC-8004 and x402 have already attracted the attention of some large institutions. Among them, ERC-8004 is a standard for agent functionality, covering identity, reputation, and claim verification, and is the foundation of core agent functions; x402 is a payment standard that defines how agents request other agents to complete payments. These standards are still in the early stages, but represent huge opportunities to explore agent functionality, reputation systems, and verification mechanisms, which can be expanded and optimized for applications in the future. Meanwhile, payment reliability and integration with existing systems remain challenges. In the next year or two, more infrastructure solutions will emerge, making it easier for institutions to deploy agents and connect to payment systems, significantly improving user experience.
In terms of community support, the Ethereum Foundation promotes standard-setting through decentralized teams and connects data, enterprise, and developer communities. The Foundation also supports developers through summer schools and global initiatives, with a particular focus on AI and agent directions, helping founders, developers, scholars, and enterprises build connections, market, and coordinate communities.
Regarding the Chinese community, Tomasz Stanczak also stated that the West attaches great importance to the region's innovation potential, which is an important opportunity for global cooperation. He encourages developers interested in AI and Ethereum to join community discussions and jointly promote technological development.
After four days of on-site creation and professional judging, the list of winners for the ETHShanghai 2025 Hackathon main track was officially announced at this main forum. The first prize was awarded to Yield Market (DeFi x Infra track), with a prize of $2,000. The second prizes went to VRF using EIP 2537 (Chain for Good track) and Wandfi (DeFi x Infra track), each receiving $1,500. The third prizes were awarded to Ponymarket (DeFi x Infra track) and Foxhole (AI × ETH track), each receiving $1,000. These winning teams showcased the latest innovations in DeFi, AI, and other tracks, injecting new vitality into the Ethereum ecosystem.
In a deep conversation between Du Yu, Executive Director of Wanxiang Blockchain/Wanxiang Blockchain Labs General Manager, and Professor Li Guoquan, Chairman of the Global FinTech Institute/Professor of FinTech and Blockchain at Singapore University of Social Sciences, the focus was on the practical applications and development trends of Ethereum.
Professor Li Guoquan provided a detailed analysis of Ethereum's leading position in global finance, government, and regulation, and discussed its future development. Professor Li pointed out that Ethereum is not only technologically advanced but also has significant advantages in regulatory compliance, and has become the preferred infrastructure for many financial institutions and governments worldwide. Especially in the application of financial products such as stablecoins and bonds, Ethereum's strong technical support has won widespread recognition from regulatory authorities.
He specifically mentioned that the Singapore government has applied Ethereum to public services such as diploma certification and international trade certificates, ensuring transaction transparency and security, reducing costs, and providing a strong example for other countries and regions. Professor Li Guoquan believes that Ethereum's decentralized nature ensures financial transparency and anti-money laundering compliance, while maintaining geopolitical neutrality, which is beneficial for enterprises' cross-border operations.
Regarding the Chinese market, Professor Li Guoquan pointed out that although consortium chains may be more secure domestically, for companies planning to expand into international markets, Ethereum is a must-choose option for enterprises facing the global market. He also looked forward to breakthroughs in decentralized hardware and cloud computing technologies in the future, believing that these innovations will further improve blockchain infrastructure and enhance risk resistance, especially progress in hardware fields such as chip design, which will provide strong support for the popularization of blockchain.
Finally, Professor Li Guoquan emphasized the importance of education and government cooperation in the popularization of blockchain technology, suggesting that domestic developers actively use the Ethereum testnet for experiments, promote cooperation between universities and research institutions, and is optimistic about Ethereum's potential in emerging fields such as green finance and carbon trading, indicating that it will have a profound impact on the global economy.
Audrey Tang, initiator of the Ethereum Application Alliance (preparatory) / Shanhaiwu, introduced that Shanhaiwu aims to build a cross-disciplinary, pop-up builder village, bringing together technology developers, AI researchers, organizational behavior experts, and governance practitioners to promote Ethereum innovation through resource linking, educational systems, and cross-disciplinary ecosystem collaboration. Each year, a month-long intensive construction is held at a fixed location, promoting exchanges between Eastern and Western blockchain communities, incubating real application scenarios, helping startups move from prototype to market maturity, and enhancing the influence of the Chinese-speaking region in the global Ethereum ecosystem.
At the same time, Audrey Tang also stated that the Ethereum ecosystem has gradually improved in terms of infrastructure construction, but still faces challenges in application implementation and sustainable community development. Over the past three years, we have promoted ecosystem development through community education, developer support, and public open-source projects, but the overall business density has not been fully realized, most communities still rely on donations to operate, and developers lack experience in financing, commercialization, and team management. As a non-profit organization, the Application Alliance operates as a public goods organization, aiming to pool the strength of communities, foundations, institutions, and enterprises through structured financing and a global membership system, establishing a long-term sustainable funding pool to support public goods construction, open-source tool development, and global application experiments. The alliance not only provides support for startups from idea to market implementation, but also promotes the practice and implementation of emerging applications in the real world through education and training, developer community building, and cross-disciplinary collaboration. The alliance aims to promote the creation of a sustainable and long-term impactful Ethereum application ecosystem, accelerating the transformation of blockchain from infrastructure construction to application prosperity. She emphasized that after the completion of infrastructure construction, Ethereum should focus on application implementation in the next decade.
Gordon, co-founder of BitSeek, introduced the project's operating model and core advantages in his keynote speech "The World's First Atomic AI Model Provider." Gordon stated that BitSeek can execute distributed deployment, execution, and governance of AI models at nearly trillion-scale without relying on centralized services. The project splits large language models (LLMs) into "intelligent atoms" that can run independently on each developer node through its proprietary DeLLM protocol, building an open, verifiable, and user-controlled intelligent paradigm.
Gordon stated that BitSeek's goal is to reduce the cost of AI usage, and it is expected that after network expansion, fees can be reduced by more than 80%, allowing both developers and users to enjoy high-performance AI services. At the same time, the solution ensures data security and anonymity through cutting-edge privacy protection design, releasing control of AI from a few large tech companies to global developers and users, bringing a new open and shared model to the AI ecosystem.
At the ETHShanghai 2025 main forum, the roundtable discussion themed "Ethereum Scaling and Future Prospects" was exciting and insightful. The panelists included Vaelyn, core contributor of ETHPanda and Managing Partner of Plutos VC, Leo, technical developer of HashKey Chain, Tang Yi, Founding Partner of Enlight Capital, Beihai, Senior Advisor of PicWe, and Anthurine, Co-Founder of EthStorage.
Leo believes that in recent years, Ethereum has achieved significant scaling through technologies such as sharding and Layer 2, improving TPS (transaction processing capability) and achieving considerable results. However, as the network expands, the issue has shifted from "can it scale" to "how to maintain consistency and trust after scaling." The fragmentation and decentralization brought by scaling have affected user experience, asset liquidity, and enterprise operational logic, making coordination, data consistency, and user experience between different chains and layers still in need of improvement. At present, most cross-chain activity is still asset transfer, and the underlying trust and compliance logic have not been fully resolved, which is a bottleneck that the ecosystem urgently needs to break through.
Tang Yi pointed out that platforms are not only continuously iterating in technology and products, but also showing positive development trends in user growth, application expansion, and service experience. From an investment perspective, we also observe this trend: more and more users and enterprises, especially B-end clients, are shifting from focusing on a single field to broader application scenarios, such as OTC business and C-end related services, fully reflecting the diversification of user needs and behavioral changes. At the same time, stablecoins and more complex financial protocols are performing steadily in the market, various products are maturing, and user experience continues to improve. The company continues to focus on talent cultivation and entrepreneurial support, and by deeply understanding and meeting the real needs of different users, it can further promote the implementation of innovative applications and the prosperity of the ecosystem. He also expressed hope that in the next decade, Ethereum can develop into a neutral, secure, and globally shared settlement layer.
"The core issue facing Ethereum scaling is not just the improvement of technical throughput, but the consensus and liquidity structure of the entire network." In Beihai's view, although the early white paper and upgrade designs clearly set scaling goals, there are many challenges in actual operation. For example, individuals who want to participate in node consensus may not be able to do so effectively due to fee or incentive mechanism restrictions, and a small number of nodes holding large amounts of assets and consensus power leads to excessive decentralization of system liquidity, creating potential risks. Sharding and other scaling solutions have not yet been fully implemented, and there is still a gap between the expected and actual network expansion, which requires continuous contributions from developers to complete the ecosystem. Early Ethereum designs, such as Satoshi Nakamoto's node capacity setting, were intended to allow any computer to join the consensus, but as the network scale and complexity have greatly increased, rapid expansion may bring security risks. Therefore, he believes that while promoting scaling, it is necessary to re-examine which parts are overextended, focus on solving consensus security and liquidity issues, and ensure that decentralization, security, and system stability are maintained in large-scale networks, achieving synergy between technical solutions and ecosystem design.
Anthurine stated that currently, Ethereum scaling has not completely solved the user experience problem, and there is still room for improvement in Ethereum's integration with other chains or applications. At the same time, she believes that technical optimization cannot be isolated, and must take into account fragmentation, L1/L2 collaboration, and ecosystem diversity. EthStorage is helping to further improve Ethereum's performance and ecosystem through code contributions and modular development. At the same time, it is also exploring AI and on-chain computing, running small and medium-sized AI models directly on-chain to achieve immutability of computation and data, while lowering the threshold for node entry and paying attention to the decentralization of the frontend, thereby effectively enhancing resistance to attacks and censorship.
In the roundtable discussion "From TradFi to Web3: Building a Bridge of Trust and Value," panelists included Xiao Xiao, Investment Partner at HashKey Capital, Z, Business Growth Lead at HashKey Capital, Nora, Community Lead at Synbo Labs, CJ Fong, APAC General Manager at GSR Markets, and Ryan Chen, Head of Research and Innovation at DigiFT.
Z pointed out that the current stage of Web3 development can be compared to the early days of the internet, similar to the early days of Amazon and Alibaba. From around 2000 to 2010, we saw the initial development of the internet, and about a decade later, e-commerce platforms began to mature. They became fully compliant and trustworthy, and by 2047, you could buy goods on any major global platform and trust the payment and banking systems behind the platform, with risks reasonably controlled. He believes Web3 also needs a similar development path, taking about ten years, focusing on compliance and training to help financial institutions (such as large brands or enterprises) enter the field. In terms of user coverage, we are indeed seeing growth on two fronts: retail user coverage and institutional investor participation. Currently, we are already in the so-called digital banking stage, making financial services more accessible to users through data transparency, with institutional users gradually shifting from traditional products to new digital financial products. For financial institutions, the question is not "whether to change," but "how to change within a compliance framework."
Nora stated that in the Web3 world, we try to use code and decentralization, rather than relying on a large number of centralized institutions, to help us establish global rules. Technology is just a tool; what really matters is whether the public can accept these concepts and rules. That is, the participation of society and programmers is more important than the technology itself, and tech companies only help turn dreams into reality. Vision and ideals, technology, and developers are the three core factors of Web3. Currently, whether users or enterprises, the core concerns are security, trust, and risk management, mainly because the financial system has a long-standing legal and regulatory framework.
"We need technology to drive the development of Web3, just as the traditional financial system relies on institutional structures, closed-end fund management, and the technical systems built by large financial institutions (such as UBS, Citi, etc.)," CJ Fong pointed out. Web3 needs to form corresponding infrastructure, and the significance of this technology is that it will make the traditional financial system more efficient, as traditional finance still uses technology from over 20 years ago, such as the SWIFT transfer system. Through the optimization of Web3 technology, traditional financial infrastructure can be upgraded to a new stage. At the same time, global attention is now turning to DATs (crypto treasury companies), and we are seeing a large number of innovative products and solutions in this field. Although this somewhat recalls the bubble stage of 2007, it is also a sign of progress.
Regarding asset tokenization and liquidity, Ryan Chen believes that when we consider RWA, the issue returns to a more fundamental level: how to provide liquidity. The key issue is that these are all security tokens. To access, trade, and enter the market, in theory, licenses are required in every jurisdiction. At the same time, when tokens are transferred, legal ownership should be ensured; if tokens are only receipts, the market may find it difficult to accept, and only true asset tokenization can drive market development.
In the roundtable discussion "On-Chain New Paradigm: The Intersection of Web3 and AI," panelists included Bi Tongtong, Co-Founder of PANONY & PANews, Ky, Co-Founder & CTO of HolmesAI, Luke, CEO of Nora, Cindy Shi, Head of Marketing at KiteAI, and Renee, Brand Director at Chainbase, who jointly explored the future development trends of Web3 and AI.
In Ky's view, most of the value of personal data (behavioral data, personal preferences, schedule information, etc.) on internet platforms is captured by large platforms. Web3 can enhance individual autonomy and earning potential in the internet economy, rather than relying entirely on traditional platform allocation. For the AI direction, he is optimistic about tracks such as AI Coding, AI Trading, and AI Prediction.
Luke pointed out that current internet interactions rely on traditional technology, such as identity verification relying on human characteristics and payment actions relying on third-party platforms. In the future, AI Agents may replace humans in interactions, forming a new paradigm of internet interaction, which will require a whole new infrastructure to support. At the same time, when AI Agents replace humans in interactions, new technologies will be needed to solve issues such as identity authentication and transaction execution, in order to establish a highly trusted AI ecosystem.
Cindy Shi believes that the combination of Web3 and AI allows more people to easily participate in AI projects, even individuals or small developers can contribute their capabilities. Blockchain can provide confirmation of rights and transparency mechanisms, enabling participants to receive benefit feedback and strengthening community trust. In the future, AI may replace humans in many tasks, and traditional systems cannot meet new transaction needs. Similar to the evolution from offline payments to virtual payment systems like VISA, future transaction systems will also require a completely new architecture to adapt to AI-driven operational models. At the same time, she believes that the large-scale advancement of Web3 and AI depends on lowering the user entry threshold and building a strong payment network.
"AI Agents will become an indispensable part of future work, helping individuals improve efficiency and creativity." Renee also pointed out that in the application process of AI or AI Agents, users will encounter practical problems, such as global payments and cross-border transactions, as well as the protection of personal data and information sovereignty. When these pain points arise, users will truly feel the value of Web3, such as decentralization, data ownership, and transparent payment mechanisms.
Finally, in the roundtable discussion "The Power of the Chinese Community in Advancing Ethereum," representatives from major Ethereum open-source communities—Ethean, Brand Lead of TinTinLand, Yoyo, Co-Initiator of ETHPanda, Twone, Core Member of GCC, Hanbai Yan, Contributor of Nantang DAO, and Marcus, Contributor of LXDAO—shared their original intentions for founding their communities and their recognition of Ethereum's philosophy. They also introduced their support for Ethereum in terms of technology and funding, especially emphasizing the importance of expanding the influence of the Chinese-speaking community in the global market.