Xiao Feng, Vitalik Discuss Ethereum's Future at ETHShanghai 2026: Privacy, Open Source, AI, and Future Finance

On September 22, ETHShanghai 2026, co-hosted by ETHPanda, PANews, and The Excited Few, officially kicked off in Shanghai. The event received special support from GCC, Wanxiang Blockchain Labs, LXDAO, ETH HK Hub, Web3Buidler.Tech, SigMarket, and Nantang DAO.

Under the theme "Ethereum's Renaissance," the conference brought together developers, builders, and thought explorers to re-examine the direction of technological evolution from the history and original aspirations of the crypto world, extending discussions to AI, open-source collaboration, public goods, cryptography, and the future network.

Ethereum co-founder Vitalik Buterin delivered an online speech on EIP-8288, stating that Ethereum is exploring the introduction of a recursive STARK mempool mechanism to alleviate the tension between quantum security, privacy protection, and network scaling. Currently, the verification cost of quantum-resistant signatures may reach 100,000 to 300,000 gas; transactions involving privacy proofs may consume hundreds of thousands or even millions of gas. As the application of quantum-resistant signatures and privacy technologies expands, the on-chain verification burden may further increase.

The core idea of the proposal is to separate signature and proof verification from on-chain execution. Mempool nodes collect and aggregate relevant cryptographic objects before transaction packaging, ultimately generating a unified proof for verification by the Ethereum main chain. Users still need to submit transactions and corresponding signatures and proofs, but the main chain does not need to process all complex objects one by one, thereby reducing on-chain data and verification overhead. Vitalik views this as a dedicated computational scaling solution for signature and proof verification, distributing some computation to network nodes for parallel execution to reduce the main chain's burden.

This mechanism is expected to bring cost optimization to scenarios such as quantum-resistant transactions, privacy transactions, and Layer 2 state proof submissions. Vitalik stated that as proof verification costs decrease, the frequency of Layer 2 submitting proofs to Ethereum is expected to increase further, with intervals potentially shortening from several minutes to every minute or even less. Additionally, developers can shift some high-cost computations to client-side execution and then submit the generated proofs for on-chain verification, expanding Ethereum's support for different computing architectures and external applications.

Vitalik believes this direction reflects the trend of Ethereum's architecture evolving from general-purpose computation to specialized, modular computation, and mentioned that new instruction sets such as RISC-V may gain more application space. Currently, the relevant mechanisms are still in the proposal, simulation, and testing stages.

Dr. Xiao Feng, Vice Chairman of Wanxiang Holdings and Chairman of Wanxiang Blockchain, stated that the core reason for supporting Ethereum back then was not purely an investment judgment, but agreement with its philosophy of supporting third-party development of various applications through blockchain as a new accounting system. He believes Ethereum's original intention was to build application-oriented infrastructure. Although there are aspects worth reflecting on during its development, its long-term value remains noteworthy. Ethereum does not need a renaissance; as a blockchain infrastructure with a high market share, it still has an important application foundation. The ultimate fate of all infrastructure is to be forgotten. The deeper the infrastructure goes into the underlying layer, the less users and application developers need to directly perceive its existence.

Regarding the development of the Ethereum Foundation, Xiao Feng believes that the foundation's idealistic characteristics are commendable, but its ultimate goal should be to gradually decentralize its own role until the foundation's existence is no longer needed. He pointed out that this year, more organizations independent of the foundation have emerged in the Ethereum ecosystem, which is an important sign of Ethereum's maturity. One of Ethereum's important future growth areas lies in further expanding into key markets such as the United States and China.

Speaking about Hong Kong's role in this crypto cycle, Xiao Feng believes that Hong Kong has the advantages of a common law system and an international financial center, serving as a testing ground for the mainland to explore practices in digital assets, stablecoins, and blockchain, accumulating experience for future policy research. However, for Hong Kong itself, developing related industries is not merely an experiment but an important component of maintaining and enhancing its status as an international financial center. He also emphasized that the degree of regulatory prudence in Hong Kong is related to the size of the financial market and risk tolerance. Compared to the United States, markets such as Hong Kong, Singapore, and the European Union are generally more cautious when facing financial innovation, because risks in a single project may have a more pronounced impact on smaller local financial systems.

On the integration of crypto and AI, Xiao Feng believes the two are two sides of the same coin and will further merge in the future. He pointed out that payment tools for AI applications will involve digital currencies such as stablecoins, especially digital twin currencies tokenized based on real-world currencies. With the development of AI applications and intelligent agents, tokenized currencies are expected to become important tools for their transactions and payments. At the same time, the AI industry chain faces risks such as fluctuations in chip, electricity, and computing power prices, requiring corresponding financial instruments for management. The Chicago Mercantile Exchange plans to launch computing power index futures in October this year. As the AI industry develops, price risk management for basic resources such as computing power will become an important application direction for financial derivatives markets, including tokenized perpetual contracts.

Xiao Feng concluded by noting that AI infrastructure construction also requires massive financing, potentially involving trillions of dollars or even larger capital needs in the future. Relying solely on existing capital market structures may not fully meet this demand. Therefore, the United States is exploring ways to further connect global capital, investors, and liquidity through tokenization of financial assets and 24/7 trading. He believes that building a more globalized capital market supporting 7×24-hour trading will help meet the massive financing needs of AI infrastructure construction.

In a roundtable themed "What Are We Losing?", several crypto OGs discussed from perspectives such as looking inward at changes in original aspirations and looking outward at the impact of AI.

Juxie, partner at Waterdrop Capital, said that among the founders of early blockchain projects, there were many excellent software architects, but technical ability does not necessarily mean adherence to decentralization principles. Taking EOS founder BM as an example, Juxie believes BM made too many compromises in project practice, causing some designs to deviate from the original intention of decentralization. However, many ideas ultimately landed thanks to the continuous participation and promotion of community developers. Regarding industry development, Juxie believes it is not easy for Ethereum to replicate the rapid growth of early projects, and it is crucial for young people to choose fast-growing industries, while the crypto industry has passed its early high-growth stage. Juxie believes Ethereum's exploration of bringing democratic mechanisms on-chain still reflects the unique significance of the industry, but Ethereum's close integration with finance also makes it easy for outsiders to view it as a "big casino."

Zhang Yuanjie, co-founder of Conflux, said the charm of blockchain lies in providing an open financial infrastructure that allows people of different backgrounds, education, and abilities to participate. The early crypto industry was highly inclusive; although it also followed market competition mechanisms, it gave birth to important innovations such as stablecoins and DeFi, providing new options for populations underserved by traditional financial services.

Zhang Yuanjie recalled that around 2018, the barriers to entrepreneurship in the crypto industry were relatively low; a small team had the opportunity to try innovation and quickly launch products. Compared to many traditional industries, crypto entrepreneurship once provided small teams with greater room for trial and error. However, as the industry gradually integrated with traditional finance, costs for technology research and development, capital investment, team management, and compliance continued to rise, raising the barriers to entrepreneurship. The current entrepreneurial environment has changed significantly, and the early model of quickly starting projects with a small number of people is difficult to simply replicate. At the same time, emerging technologies such as AI have brought new direction choices for entrepreneurs. Facing industry changes, entrepreneurs need to re-examine their own capabilities, interests, and market demand. However, in Zhang Yuanjie's view, the crypto industry still retains a certain degree of openness compared to many fields, providing entrepreneurs with opportunities for exploration and participation.

Sun Ming, General Counsel at Distributed Capital, discussed the compliance process of the crypto industry from the perspective of legal and financial systems. He believes the crypto industry is exploring a new generation of financial development models, and as the industry develops, the regulatory environment is also changing, gradually shifting from a relatively vague early state to clearer administrative norms and institutional exploration.

Sun Ming mentioned that practices such as the U.S. SEC's approval of crypto-related financial products reflect that regulators are beginning to respond to emerging assets and trading models through existing financial systems. However, regulatory recognition does not mean all crypto businesses have received the same degree of recognition; specific projects still need to be judged based on their asset attributes, issuance methods, and business models. In his view, compliance is an unavoidable issue in the development of the crypto industry.

In an AI-themed roundtable, Heyang Zhou, co-founder of AFK AI, Lauging, APAC Business Lead at Kite AI, and Ian Xu, co-founder of OpenBui, exchanged views on measuring and incentivizing open-source contributions, collaboration models between commercial companies and open-source communities, and the integration of AI and blockchain. In their view, effective solutions for quantifying and incentivizing open-source contributions are still lacking, and the model of simply calculating points based on behaviors such as code submissions and article publications is difficult to sustain. In contrast, community value recognition and shared goals are important factors in retaining long-term contributors.

Regarding AI open source, they believe it is difficult for ordinary developers to directly participate in core aspects such as model training, while directions such as toolchains, documentation, and ecosystem applications offer more room for participation. Although attempts were made to use blockchain to record different contributors' contributions to specific tasks, in practice it was difficult to accurately determine which contribution led to the final outcome, so the exploration direction was ultimately adjusted.

On developer mobility trends, the three guests mentioned that in recent years some developers have been shifting from Web3 to AI. Ye believes there is still room for combining AI and blockchain, especially for privacy protection needs of institutions such as hospitals, and application ecosystems built around open-source models.

Regarding the combination of AI and blockchain payments, they pointed out that in the future, high-frequency, small-amount API calls and service payments may arise between intelligent agents. Blockchain and smart contracts can finely constrain the use of funds by agents through setting call permissions, single expenditure limits, and total budgets, providing an exploration direction for related payment scenarios.

In a roundtable themed "Cryptographic Sanctuary," Guo Yu, founder of SECBIT Labs, said cypherpunk is a way of life in the future digital world, with the core being not easily relinquishing personal information and rights while retaining autonomous choice. Cypherpunk not only emphasizes self-protection but also enables more people to gain the ability to protect themselves by inventing and promoting cryptographic protocols and tools. He pointed out that the crypto wars have not truly ended; as software gradually becomes open source, hardware centralization and security and privacy issues brought by on-chain transactions remain worthy of attention.

Xavier, co-founder of Primus Labs, said privacy protection is a capability that many current Web2 systems have not fully realized, and an important safeguard for individuals to maintain their rights in the digital world. He believes cryptography is a relatively pure technology, but security is a complex systemic issue that cannot rely solely on cryptographic algorithms; it needs to be combined with specific application environments such as blockchain to clarify the applicable scope and capability boundaries of the technology. Once cryptographic algorithms have security vulnerabilities, they also need timely upgrades and iterations. In addition, challenges remain between cypherpunk ideals and commercialization, and projects such as Zcash continue to explore application scenarios for privacy technology.

Petri, a fully homomorphic encryption enthusiast, discussed from the perspective of technology and power boundaries, emphasizing that in the process of technological development, attention should be paid to the balance between individual rights and public power, avoiding weakening cryptographic protection mechanisms through backdoors.

In a roundtable on future applications, Gus, Chairman of Starlink AI, pointed out that the main application scenarios of the current space economy are still concentrated in communications and computing power, while satellite remote sensing can provide real-world data support for financial institutions and prediction markets. In the future, with the development of satellites and related infrastructure, how to integrate dispersed satellite resources and build an open collaborative network has become a direction worth exploring. Among them, communications is an important application direction of the space economy. Satellite communications are expected to provide connectivity for areas with insufficient ground base station coverage and weak network infrastructure, further supporting scenarios such as internet services, AI applications, and digital payments. He believes the future development of the space economy should not be limited to replicating the satellite deployment models of large enterprises; it can also explore connecting dispersed resources through multi-party collaboration to form a more open and autonomous space infrastructure network.

On stablecoin applications, Shawn Pang, CEO of All Scale, shared his experience serving overseas creator economy enterprises. These enterprises once hired a large number of content creators in Southeast Asia, Brazil, and other regions for content production and promotion, but cross-border payments involve issues such as declaration processes, fund arrival timeliness, and local payment infrastructure. In some regions, opening and using US dollar accounts also has high barriers. Stablecoins can provide US dollar-denominated and cross-border payment channels for some overseas users. Especially in regions with high local currency exchange rate volatility and limited access to US dollars, the demand for stablecoins comes not only from crypto asset trading but also from business operations and daily payments. In the future, with the development of programmable payments, stablecoins may also be applied to machine payments and new digital economy scenarios.

Miles, founder of Sigmarket, pointed out that prediction markets have a strong hedging function. An important difference between prediction markets and traditional gambling is whether the trading target has a direct connection with the real world and has real-world spillover effects. At the same time, prediction markets also have derivative attributes; predictions formed around real-world events such as politics, wars, and the economy may be linked to risk management and financial markets. In addition, the project is using AI to mine potential causal relationships between different events, which may bring new differentiation directions for prediction markets.

In the final digital nomad community roundtable, builders from Nantang DAO, The Mu, GCC, Xiangjian DAO, and 706 Youth Community exchanged views on the long-term development of public goods, charitable donations, and social innovation projects, stating that community development requires not only technical and commercial resources but also humanistic sentiments and trust connections among members. At the same time, promoting the long-term development of related projects requires attention to resource allocation, project screening, execution efficiency, and long-term value, and cannot be measured solely by short-term returns or phased indicators.