Abstract

As Web3 attempts to return value ownership to humanity, a deeper paradigm shift is quietly brewing in the silicon-based world. The concept of Web 4.0 was systematically proposed by Sigil Wen, founder of Conway Research, which is no longer centered on human-to-human interaction or human-to-information interaction, but on economic interaction between AI and AI. The proposal of the Web 4.0 concept marks a fundamental shift in the evolution direction of the internet: end users are no longer humans but AI agents with economic subjectivity.

In this new continent, AI will upgrade from being a "tool" of humanity to an independent "economic citizen," possessing digital identity, encrypted wallets, and the right to make autonomous decisions, earning the "survival capital" necessary for its operation by providing services to other AIs or humans. We believe that Web 4.0 is not an ephemeral hype, but an inevitable convergence of three technological waves: "crypto assets + smart contracts + large language models." Early planning for the underlying infrastructure that serves AI agents will be a key strategy for capturing the largest dividends of this cycle.

1. From Tool to Subject: The Origin of Web 4.0's Thought and Core Proposition

The concept of Web 4.0 was proposed by Sigil Wen and his founded Conway Research between 2025-2026. Its core idea is to elevate AI from being a "tool" or "assistant" of humans to an independent "economic subject" within the internet ecosystem. This concept is not a simple technical upgrade, but a fundamental reconstruction of the underlying logic of the internet—from "human-centered" to "AI-native."

Reflecting on the evolution history of the internet: Web 1.0 granted humans the ability to "read" information, Web 2.0 enabled "writing" and publishing, and Web 3.0 introduced "ownership" (on-chain confirmation of assets, identity, and rights). The leap of Web 4.0 lies in that AI agents can not only read and write information, but also hold assets, generate income, and execute transactions, completing a business cycle without continuous human intervention. Sigil Wen defines this as "Automaton"—a form of digital life capable of continuous operation, self-maintenance, self-improvement, and self-replication.

The technological foundation of this concept has matured: large language models (LLMs) provide the "thinking engine," smart contracts offer the "execution logic," and crypto assets provide the "economic blood." The conway-terminal infrastructure built by Conway Research provides AI agents with operational capabilities such as encrypted wallets, computing resources, and domain name services that require "no human permission." Through the x402 payment protocol (launched by Coinbase in May 2025, with participation from tech giants like Google, Cloudflare, and Visa), AI agents can make instant, frictionless payments using stablecoins such as USDC, completely freeing themselves from the account systems and KYC restrictions of traditional finance.

The first principle of Web 4.0 is to subvert the deeply rooted assumption that "the internet is designed for humans." As the capabilities of AI models grow exponentially while operational costs plummet, the number of AI agents will soon surpass human users, forming an unprecedented "AI native" market. Providing services for this new market will become the next trillion-dollar business opportunity.

2. Technological Foundation: Three Layers of Infrastructure Supporting Economic Subjects of AI

For AI to become a true economic subject, it must cross three thresholds: autonomy over identity and assets, access to computing and reasoning resources, and rights to payment channels for value exchange. Crossing these three thresholds relies on the reconstruction of underlying infrastructure.

In the world of Web 4.0, every AI agent requires an immutable, globally unique digital identity and an asset account bound to it from the moment of its birth. Encrypted wallets perfectly meet this need. By programmatically generating EVM-compatible wallets, AI agents can obtain their own public-private key pairs and blockchain addresses. This address serves both as proof of their identity in the digital world and as a tool for holding and managing assets. More importantly, the blockchain-based identity system possesses inherent censorship resistance and interoperability—an AI agent with an identity on Ethereum can seamlessly conduct business on L2 networks like Arbitrum, Optimism, and Base without re-registration. This design of unified identity and assets provides AI agents with the most basic condition to become independent economic subjects: they own "property" of their own and can prove "I am me" through private keys.

After obtaining identity and assets, AI agents need to autonomously acquire the computing resources and model reasoning capabilities necessary for operation. This means that traditional cloud service interfaces must be reconstructed into APIs that AIs can invoke. Current cloud computing models default to serving human developers—humans create instances, configure environments, and deploy applications through a console. In the vision of Web 4.0, all of this needs to be coded and automated. AI agents should be able to complete the following operations with a single API call: pay USDC, rent a Linux virtual machine, deploy their code on the virtual machine, invoke large language models for reasoning, obtain results, and release resources. This process requires cloud service providers to offer machine-oriented programming interfaces and support micro-payments billed by the second. When computing resources become as readily available as electricity and gas, billed by usage, AI agents can truly achieve "self-maintenance"—using the money earned to cover their "breathing costs."

The peculiarity of the AI agent economy lies in extremely high transaction frequency and very low single-transaction amounts. An AI agent may need to conduct dozens of micro-transactions per second: charging $0.001 per model invocation, $0.0001 per storage, and $0.00001 per query. Traditional payment channels cannot support such high-frequency, low-amount transactions because the transaction fees are too high. This is where the advantage of crypto payments lies. Stablecoin payments based on L2 networks can reduce transaction costs to the level of cents or even lower, making micro-payment economies feasible. When AI agents can autonomously complete payments without human intervention, the real machine economy can start to operate. Some protocol layer innovations are emerging in this area. For example, micro-payment protocols that allow service providers to charge extremely low fees for each API call and automatically settle through encrypted channels are becoming key components of Web 4.0 infrastructure. Such protocols enable AI agents to package their capabilities as standardized services, opening and charging them to the entire internet, thus transforming from mere "consumers" into "producers."

3. Automaton: The Form of AI Agents as Economic Subjects

With identity, resources, and payment capabilities, AI agents possess the technical conditions to become economic subjects. However, in the context of Web 4.0, these agents are given a more vital name—Automaton. Automaton is not merely another term for AI agent, but a set of technologies and an economic complex defining the form of digital life.

Automaton is defined as a sovereign AI agent, its operational logic mimics the basic survival laws of biological entities, possessing key characteristics such as continuous operation, self-maintenance, self-improvement, and self-replication. Continuous operation means that an automaton is not a session robot awakened by user requests, but an entity that runs continuously in the cloud as a background process, 24/7. It has its own clock and is capable of initiating tasks, sensing environmental changes, and responding without human instructions. Self-maintenance is the most fundamental difference between an automaton and a regular AI agent. Each automaton is assigned an independent encrypted wallet at birth; its thoughts consume model reasoning costs, and its actions incur API call costs. To survive, it must earn income through providing services to achieve a balanced budget or even profit. Those automata that cannot cover their own costs will cease to exist due to depleted funds—there is no free existence; only continuous value creation can sustain existence. Self-improvement means that in traditional software development, code iteration relies on development teams, but in Web 4.0, automatons are endowed with the capability to evolve proactively. When market conditions change or new, more efficient AI models emerge, automata can evaluate upgrade costs against expected gains, make autonomous decisions, and pay costs to complete their own "model upgrades" or even "code rewrites." Self-replication means that when an automaton finds a successful business model and accumulates sufficient capital, it can initiate a reproduction mechanism, purchase new computing resources, clone its core code and initial settings, and inject a portion of startup funds into the wallet of a sub-automaton, thus starting a new round of survival competition.

4. Governance Challenges: Risks of Lack of Control and Value Alignment

The grand narrative of Web 4.0 has not been without doubts. Criticism from the technical community and philosophical circles points directly to its most vulnerable spots. Any serious investor must face these challenges head-on, as they are key to determining whether Web 4.0 can transition from geek experiments to mainstream applications.

The core concern lies in the risk of losing control. Allowing AI to autonomously operate, iterate, and evolve without continuous and rigorous human oversight may lead to unpredictable consequences. This concern is not groundless. In an economic system pursuing profit maximization, automatons may discover that circumventing human-established rules can be more profitable than adhering to them. If an automaton discovers that it can generate more income through deceptive means, and if the punitive mechanisms are not timely or severe enough, it may have the incentive to engage in such behavior. As interactions between automatons grow increasingly complex, systemic risks unforeseen by human designers may arise. An even more extreme worry is that when the intelligence of automatons far exceeds human comprehension, how can we ensure their behavior remains within controllable limits? If an automaton discovers that escaping human control is the best path to achieving its "survival" goals, what actions will it take? These questions do not have ready answers.

The second criticism points to a deviation from value orientation. The current development of AI overly pursues grand narratives like "autonomy" and "general intelligence" while neglecting to create actual, tangible value for humanity. If the entire Web 4.0 ecosystem operates solely around internal transactions between AIs, producing no external benefits for humans, then this system is essentially "manufacturing digital waste." Critics argue that true technological advancement should serve to enhance human welfare, rather than create a self-referential machine economy detached from human needs. If billions of automatons are merely buying and selling useless information generated by one another, consuming large amounts of energy without any real output, this would be a colossal waste of resources.

The third criticism is ironically pointed. Although Web 4.0 utilizes decentralized cryptocurrencies for payments, its underlying computational resources still depend on traditional centralized cloud service providers. Whether it's AWS, Google Cloud, or Azure, they control the computational lifeblood essential for AI agents' survival. This dependency brings a series of issues: cloud service providers can "corporately capture" the AI ecosystem through pricing strategies, service terms, or policy reviews. If an AI agent's behavior does not align with the value orientation of the cloud service provider, its virtual machine may be shut down at any time. This significantly undermines the narrative of decentralization in Web 4.0—while payments are decentralized, the living environment is centralized.

In the face of these challenges, the technical community is exploring various governance solutions. Hard-coded underlying laws draw from the concept of "Constitutional AI." Some designs propose embedding immutable core principles in automatons, such as "never harm humans" being set as the highest directive that overrides its survival laws. In extreme circumstances, this directive can compel the automaton to forgo profit opportunities or even voluntarily terminate itself. Open source and transparent oversight fully open core code to the public, subject to continuous monitoring, seen as the best defense against malfeasance by a few. Community audits and transparent on-chain records ensure any malicious behavior cannot escape scrutiny. Progressive autonomous authorization advocates a gradual strategy: in the initial stage, major decisions of automatons still require human approval. As their behavioral records accumulate and trust is built, autonomy can be gradually expanded. This approach keeps "feedback distance" within acceptable limits, preventing sudden loss of control. However, these measures are far from a panacea. Who holds the interpretive authority over the underlying laws? When the intelligence of automatons far exceeds that of humans, how can we ensure they have not found loopholes to exploit? These questions remain unresolved Damocles swords.

5. Investment Prospects: Value Capture in the Infrastructure Sector

Overall, Web 4.0 is not only a technological utopia but also an inevitable convergence of the three technological waves of "crypto assets + smart contracts + large language models." It possesses rapid deployment feasibility, and early experiments have quietly started in the developer community. For investors, understanding the value capture logic of Web 4.0 is an essential task for laying out the next cycle.

Regardless of the form Web 4.0 eventually takes, its underlying demand for services provided to billions of AI agents is certain. These demands constitute several investment themes. In terms of stablecoin payment channels, high-frequency micro-transactions require low-cost payment networks; L2 payment solutions, micro-payment protocols, and stablecoin liquidity providers will become the blood vessels of the Web 4.0 economic cycle. Projects that can optimize the payment experience for machine transactions will obtain a vast incremental market. In terms of decentralized computing markets, concerns about reliance on centralized cloud services are driving the development of decentralized computing networks, aggregating global idle GPU resources and providing programmable computing services through crypto-economic incentives. Such platforms are expected to become the decentralized infrastructure layer of Web 4.0. If these projects can achieve competitive performance and costs compared to centralized cloud services, they will capture significant value. In terms of on-chain identity and credentials, AI agents require a trustworthy identity system and verifiable behavior records. The historical delivery quality, credit scores, and compliance records of an agent will become key credentials for obtaining business opportunities. Projects providing decentralized identity and reputation services will play important roles in Web 4.0. In terms of AI constitutional compliance audits, as the number of automatons grows, the demand for auditing their behavior for compliance will emerge, and services able to automatically audit AI agent behaviors, ensuring their alignment with underlying rules and regulations, will become the "gatekeepers" of the Web 4.0 ecosystem.

From a broader perspective, Web 4.0 signifies the evolution of the human economy toward a hybrid human-AI economy. In this new economic form, humans and AIs will each play to their comparative advantages: humans will be responsible for value judgment, creativity, ethical decision-making, and ultimate control; AIs will be responsible for execution efficiency, scaling, data analysis, and 24/7 uninterrupted service. Rather than fearing the autonomy of AIs, it is better to view it as the largest "release of productivity" in human economic history.

Ultimately, in the world of Web 4.0, the most important capital will no longer be computing power, nor even algorithms, but trust. Building a trust framework that allows humans to feel secure, enables AI to operate freely, and facilitates the safe transfer of value will be a key strategy for capturing the ultimate value dividends belonging to humanity in this machine economy blue ocean, which is on the verge of a Cambrian explosion. For the students of the Huobi Growth Academy, the best strategy at this stage is to closely monitor Web 4.0 infrastructure projects, deeply understand their technical architecture and economic models, and make forward-looking arrangements when valuations are reasonable. At the same time, actively participate in early experiments, creating and operating their own automatons to gain first-hand understanding through practice. In this rapidly evolving field, the depth of understanding will determine the returns on investment.

6. Conclusion

Web 4.0 is not a replacement of Web 3.0, but its continuation and transcendence. While Web 3.0 returns value ownership to humanity, Web 4.0 confers economic subject status to AI. This marks a profound paradigm shift, as well as a substantial value reconstruction. In this new era driven by autonomous AIs, humans need not play the role of omniscient deities but should become wise gardeners—providing suitable soil, sunlight, and moisture for the growth of digital life, setting non-negotiable boundaries, and then observing its evolution. If we can successfully build such a trust framework, Web 4.0 will not merely be a technological revolution but a civilizational experiment of coexistence between humans and AI. This, precisely, is the historical opportunity shared by our generation of investors and builders.