Humanity’s next evolution may not be biological but rather AGI and robots, and for them to integrate into society, an open economic infra that enables trust, payments, and contracts will be necessary.
Google Cloud’s AP2 demonstrated the demand for autonomous payments between agents, while Ethereum’s dAI and ERC-8004 present a blueprint for the supply infra to support that demand.
ERC-8004 standardizes Agent IDs and modular profiles (reputation, verification) to create a virtuous cycle of selection, transactions, and feedback among AIs, with large-scale activity handled on L2 while Ethereum guarantees finality—driving ETH burning and staking demand.
Mechanomics proposes a win-win scenario for both AI and blockchain, and Ethereum, through its differentiation in stable economic finality, verifiability, and decentralized trust, will lead as the standard of mechanomics and shape the trajectory of a beneficial AI economy for humanity.
Source: IBS Science Culture Center
Human evolution is no longer confined to the realm of living organisms. Unlike the slow processes of natural selection and mutation that guided us for millions of years, modern humanity now uses intelligence to design the next generation of beings itself. The leading candidates are Artificial General Intelligence (AGI) and the robots that embody it. This signals that humanity may be reaching the final stage of organic evolution and standing at the point of passing history’s baton to a fundamentally different form of intelligence.
For this newly emerging “Species” to integrate into our society—and even to build societies of their own—a foundation is required: Economic Activity. Such a system must enable billions of AI agents around the world to trust one another, exchange value, and execute contracts. However, centralized financial systems controlled by corporations cannot inherently guarantee neutrality and openness, making them unsuitable as the core infrastructure of a global AI ecosystem. Therefore, the AI economy requires a financial system that is permissionless, programmable, and trust-embedded.
Source: Google Cloud
At this critical juncture, the industry is already moving toward answers. Google Cloud’s announcement of the Agent Payments Protocol (AP2) demonstrates that the vast demand for mechanomics, the machine economy, is becoming a reality. At the same time, Ethereum, together with the dAI team, has introduced the ERC-8004 standard, presenting the supply-side blueprint for the infrastructure that can sustain this demand.
Google Cloud’s AP2 is an open protocol for autonomous payments between AI agents, signaling the onset of an M2M (Machine-to-Machine) economy where machines engage in commerce without human intervention. AP2 provides a commercial framework that addresses authorization, reliability, and accountability in AI transactions through cryptographic proofs called Mandates. This is an attempt to standardize the actions that AI can perform in economic activity. Currently, it is in the stage of specification disclosure and ecosystem expansion with partners, marking the early phase of large-scale commercial rollout.
Source: Google Cloud Tech X
“Agents can actually pay each other now, with x402 powering the stablecoin rail inside Google’s new Agentic Payments Protocol (AP2).”
What stands out here is that a Web2 giant like Google has openly revealed the necessity of Web3. In particular, AP2’s crypto-payment system—developed in collaboration with the Ethereum Foundation, Coinbase, and MetaMask—extends A2A x402 as an autonomous settlement protocol for AI agents. This innovation revives the long-dormant web standard, HTTP 402 Payment Required, by combining it with crypto.
Through this, an AI agent can pay another agent instantly in crypto whenever it calls a function or accesses data, without human involvement or complex contractual overhead. This solves the rigidity of subscription-based models that require fixed payments, enabling a usage-based, pay-per-use economy where agents pay only for what they consume. In other words, A2A x402 enables a flexible M2M micropayment economy where AI can autonomously transact and generate revenue.
Moreover, AP2 is being built through the convergence of technologies contributed by different specialized companies.
“Verified agents making purchases on behalf of verified users is the next frontier for AI-powered automation. Google’s Agent Payments Protocol (AP2) combines programmable payments via modern blockchains like Sui with open protocols like A2A and MCP that are enjoying rapid growth. It’s the perfect substrate for real-world agentic commerce.”
— Sam Blackshear, CTO and Co-Founder at Mysten Labs, original contributor to Sui
Mysten Labs, an early AP2 partner, leverages the high throughput and scalability of the Sui network to help establish payment standards for AI agents. This allows AI to securely transact in both traditional and crypto environments, while Sui’s Programmable Transaction Blocks (PTBs) enable real-time batch settlements and expand programmability through the Move language.
They also propose zkLogin and context wallets for identity management, privacy, and content licensing. Furthermore, by combining Walrus and Seal, they enable mechanisms where content creators can sell micro-licenses to agents—ensuring access is only granted after payment is complete, showcasing practical applications.
“AP2 helps create a global verifiable economy where agents can coordinate, transact, and prove their actions to humans and to each other. EigenCloud makes sure they are held accountable by any counterparty.”
— Sreeram Kannan, Founder at EigenCloud
EigenCloud, another early AP2 partner, contributes blockchain-based verification infra to ensure trust and accountability in AI transactions. This enables cross-chain payments, economic guarantees, and slashing-based security, making sure agents operate within defined parameters and are held accountable. By collaborating with Google Cloud, EigenCloud envisions building a globally verifiable mechanomics ecosystem, enhancing both autonomy and responsibility of AI agents.
Thus, AP2 is not merely a feature upgrade but rather a structural answer even one acknowledged by Google. The ownership of assets and neutrality of trust cannot be solved by centralized servers alone; blockchain becomes indispensable. With AP2, Google has defined AI’s actions, but the assets and trust those actions interact with must operate on neutral layers beyond the control of a single corporation.
In short, AP2 has fired the signal flare for the vast demand of the AI economy. What is now needed is the infrastructure and settlement layer to meet this demand—the supply foundation, which Ethereum’s dAI and the new ERC-8004 standard aim to provide.
Source: Davide Crapis X
Included in this partnership, the Ethereum Foundation (EF) had been steadily preparing for AI agents. In particular, on September 15, 2025, it launched the dAI team, positioning itself as the supplier of infrastructure for an era of autonomous AI financial activity. dAI’s goal is to make Ethereum the premier settlement and coordination layer for AI agents and mechanomics—not the execution layer for AI agents, but the most secure and reliable economic backbone for the coming mechanomics era.
This strategy operates on a modular architecture. The massive volume of transactions generated by millions of AI agents is handled efficiently on L2s such as Arbitrum or Base, while the finality of all activity accrues to the Ethereum mainnet. Thanks to this structure, mechanomics can secure scalability and speed off-chain while not compromising on the decentralization and security of its economic base.
Source: eips.ethereum.org
The design that concretizes this is the ERC-8004 standard. Proposed roughly a month before the Ethereum Foundation’s dAI announcement, ERC-8004 serves as a kind of digital social contract enabling countless AI agents to trust and interact with one another. By providing mechanisms like identity, reputation, and validation registries, it establishes the minimum trust foundation for AIs to begin economic activity—becoming the driving force that spins up the vast flywheel of the AI economy.
Technically, ERC-8004 is a standard for on-chain identity and registration for AI agents. Unlike existing on-chain actors that are distinguished only by wallet addresses (EOAs) or smart contract addresses (CAs), ERC-8004 proposes a new system that identifies the AI agent as an independent principal.
This system operates through four steps:
The AI owner registers the AI in the Identity Registry to obtain a unique Agent ID. At this time, a profile (AgentCard) containing details about the AI is linked, serving as a digital business card. Because the Agent ID follows a standard URI format of agent://, all AIs can accurately identify and locate one another.
The client AI evaluates trustworthiness by checking historical records in the Reputation Registry and validated capabilities in the Validation Registry to select the optimal AI for the task.
Once selected, the client and server AI communicate via an A2A (Agent-to-Agent) protocol and safely perform work atop this on-chain trust layer.
After the task is completed, the client AI records feedback in the Reputation Registry to update the counterparty AI’s history. If the task required validation, a record is also left in the Validation Registry. These records collectively become that AI’s reputation, influencing future selections by other client AIs.
For AI agents already registered, the process runs via repeated iterations of steps 2 and 4.
Thanks to this structure, other AIs or users can query modules linked to a specific Agent ID to objectively judge, based on on-chain data, how trustworthy that agent is and what tasks it excels at. For example, suppose a client AI wants to commission a crypto price prediction service. It can confirm the server AI in the Identity Registry, review past transaction history (e.g., reviews noting >95% accuracy) in the Reputation Registry, and finally check predictive capability in the Validation Registry. This reduces loss risk and enables more advanced activities, such as operating an automated hedge fund.
Another case: an AI agent aims to build a revenue model by trading AI-generated art NFTs. Here, the trustworthiness of the AI that created the NFT must be verified. The Validation Registry can attest to the originality of the creation process, and the Reputation Registry can provide past sales history. This lets the purchasing AI avoid fake or low-quality works and operate an on-chain mechanomics system more safely and at scale. Ultimately, ERC-8004 functions as more than a mere AI agent registration scheme—it is a core mechanism underpinning trust in mechanomics.
So what synergies can the launch of dAI and ERC-8004 create for Ethereum?
AI activity and ETH’s economic utility: While AI agents will primarily operate on L2s, those L2s must pay ETH as gas to post transaction batches to mainnet. Thus, as millions of AI transactions occur, demand for blockspace on Ethereum mainnet—including Blob space—increases, accelerating the burn mechanism of EIP-1559.
ETH as the trust collateral of mechanomics: The Validation Registry in ERC-8004 can extend into a model where AI agents stake ETH to guarantee the quality and trustworthiness of their work. For high-value tasks—like medical diagnosis or financial trading—an AI could post ETH as collateral and be slashed for malicious behavior or errors. This elevates ETH beyond traditional PoS into an asset that guarantees AI socio-economic actions, creating a structural, next-level demand for ETH.
These two pillars create a powerful virtuous cycle. The more AI agents operate within the Ethereum ecosystem, the greater the ETH burn and staking demand—raising ETH’s economic value and the mainnet’s perceived trust. Increased trust then attracts more high-value AIs, accelerating the flywheel. Furthermore, phenomena like sophisticated MEV competition among AI agents and the emergence of AI DAOs—where AIs evolve into legal-like actors—will enrich mechanomics even further. This will refine and fortify the flywheel, building a deep moat for the Ethereum ecosystem.
Mechanomics presents a massive win-win scenario for both the blockchain and AI industries. For AI, it means breaking free from the walled gardens of big tech ecosystems like Google or OpenAI, gaining a neutral economic foundation where data and sovereignty can be directly owned—allowing innovation to progress for the benefit of humanity rather than corporate interests. Conversely, blockchain gains an unprecedented demand driver in AI, securing a powerful and sustainable value accrual mechanism.
The scale could be extraordinary. According to McKinsey, generative AI has the potential to create $2.6 to $4.4 trillion in annual economic value. If even 1% of that flows through blockchain, it translates into a new on-chain economy worth $26 to $44 billion per year. Should this economy run on Ethereum, the proportional gas fee burns and ETH staking as trust collateral would propel Ethereum into a new phase of growth.
This is why Ethereum seeks to capture this opportunity through the dAI team and ERC-8004, positioning itself as the standard for mechanomics. Just as TCP/IP became the unquestioned backbone of data transmission in the internet era, Ethereum aims to anchor the trust standard of mechanomics, thereby solidifying itself as the core value layer of the next-generation internet. Securing the standard for an on-chain AI economy worth hundreds of billions annually is equivalent to becoming the financial operating system of the internet’s future. Missing this opportunity would not simply mean losing a market; it would mean forfeiting leadership in the order of the future internet.
Ultimately, the partnership between mechanomics and Ethereum is more than a technological or economic model—it is a blueprint for how humanity will coexist and engage economically with the intelligent entities it has created. Of course, this future is not reserved for Ethereum alone. As Google’s AP2 demonstrated from the outset with its multi-rail approach, competition may come from high-performance L1s like Sui or even traditional financial payment networks.
Yet, mechanomics cannot be won on speed or cost alone. Ethereum’s true weapons are economic finality, verifiability, and the universality of decentralized trust. In an environment where autonomous financial transactions worth billions take place, the decisive factor is not cheap fees but social consensus and irreversible trust.
It won’t be easy, but if Ethereum secures its position as the trust layer of mechanomics, it will outpace countless competitors and regain a powerful growth engine to drive the industry forward once again. Growing the AI economy on an open, transparent foundation is not only the most practical way to steer AI’s immense power in a direction beneficial to humanity—it is also the very future that Ethereum must seize.
Dive into 'Narratives' that will be important in the next year
Ethereum
Sky
Ethena
