Ethereum is evolving its core structure to facilitate the operation of independent AI agents, potentially bringing machine-to-machine transactions directly onto the blockchain within the next year.
The Ethereum Foundation recently established a specialized Decentralized AI (dAI) Team. This team will focus on enhancing agent identification, establishing trust protocols, and improving payment systems, which includes promoting the use of ERC-8004.
ERC-8004 is a proposed standard designed for agent authentication and verification, aiming to secure agent identities and certifications at the foundational protocol level.
This initiative positions Ethereum as a central platform for agent-based economies, prioritizing unrestricted access and resistance to censorship in its design. Community discussions around ERC-8004 highlight how built-in identity verification and trust mechanisms could allow automated systems to autonomously negotiate deals, deposit guarantees, and manage secure transactions without needing traditional intermediaries.
The initial goals involve developing research and standards suitable for adoption by digital wallets, middleware, and decentralized applications (dApps) by 2026. This aims to create a reliable infrastructure for agent applications to operate within.
AI Impact Evident in Crypto Market Token Activity
Cryptocurrencies with an AI focus, such as Bittensor, Fetch.ai (ASI), Internet Computer, and Render, have demonstrated consistent on-chain activity and maintained stable prices throughout the third quarter, performing better than many other alternative cryptocurrencies during recent market declines.
Koinly’s market analyses indicate ongoing interest in decentralized computing, AI inference, and agent frameworks. Ecosystem reports also reveal consistent use of Internet Computer Protocol (ICP) for native app hosting and Render’s GPU marketplace for AI-driven tasks.
According to Token Metrics, the total value locked (TVL) in decentralized finance (DeFi) has increased from approximately $72 billion at the start of 2025 towards $100 billion. New AI-integrated DeFi solutions like Blackhole DEX on Avalanche, Sahara AI, and Moby AI are contributing to significant transaction volumes and fee generation, even during periods of market volatility. Token Metrics attributes this to a broader shift towards automated liquidity management and agent-driven operations across different blockchains, utilizing messaging and omnichain technologies.
Payment technologies are increasingly adapting to accommodate agent-based use cases at the foundational protocol level. Google introduced the Agents to Payments (AP2) protocol in September. AP2 enables software agents to initiate and confirm customer payments using standardized processes. This is a crucial component for machine-to-machine billing and subscription models that can integrate with cryptocurrency payment rails.
Google Cloud states that AP2 is designed with user consent, verifiable agent identities, and reversible transactions to ensure regulatory compliance. Initial tests include Ethereum and ICP integrations via third-party connectors, linking traditional financial accounts with blockchain-based transfers.
As these trials advance, digital wallets could start treating agents as primary users, with attestation methods similar to ERC-8004. This allows for the implementation of spending limits, restrictions on counterparties, or the requirement of human approval for high-value transactions.
Advanced Models Link Infrastructure Upgrades to Tangible Network Demand
Token Metrics’ September projections indicate that AI-driven smart agents could account for 15% to 20% of all DeFi transaction volume by the end of the fourth quarter. If sustained and boosted by Ethereum’s dAI roadmap, this could push AI-integrated protocols to a TVL range of $200 billion to $300 billion by the close of 2026.
The same analysis suggests this will drive increased usage of the base layer, with gas consumption for agent identity and execution contracts increasing by 30% to 40% each quarter in 2026, particularly as standards like ERC-8004 are widely adopted across custody solutions, consumer wallets, and DAO middleware.
In practice, this means governance, treasury management, fee distribution, and cross-chain liquidity management could be handled by software agents operating with pre-set risk limits, insurance, and verifiable on-chain credentials.
Improved security is another factor driving adoption. Academic and industry research on AI-assisted smart contracts indicates a significant reduction in successful exploits when contracts can detect irregularities, adjust parameters, and isolate suspicious activities in real-time.
Early models demonstrate up to a 70% decrease in successful cyberattacks for systems using a combination of rule-based controls and AI-driven heuristics, compared to systems with static parameters.
This outcome relies on transparent update procedures and monitorable on-chain behavior to prevent hidden control mechanisms, a concern that aligns with regulatory focus on smart contract auditability and incident reporting.
Shift from Conceptual to Practical Application
Regulatory initiatives in the United States and Europe are addressing automated financial agents, transparency in adaptive contracts, and disclosure requirements for model risk.
DLA Piper’s September report, along with other legal analyses, suggest that agent identities, usage policies, and exception handling need to be understandable to regulators and other parties. This aligns with Ethereum’s focus on identity and attestation, reinforcing rather than undermining its approach.
Recent enforcement trends emphasize the effectiveness of controls over outright technology bans, which creates opportunities for compliant agent operations as standards develop.
Hiring data remains positive, with Recruitblock reporting a 22% year-over-year increase in roles related to the intersection of AI and blockchain in 2025. These roles span protocol engineers, data infrastructure specialists, and applied cryptography experts, which is crucial for the widespread adoption of agent frameworks across consumer and enterprise applications.
The concept of a machine economy is not confined to a single platform. Avalanche supports AI-governed liquidity through Blackhole DEX, Ethereum prioritizes identity and transaction settlement, NEAR and ICP are attracting on-chain app hosting and low-latency AI inference, while Render provides GPU resources for AI training and model deployment.
Analyses from Koinly and Token Metrics suggest these platforms play complementary roles rather than being direct competitors. The consensus is that demand for decentralized inference and marketplace coordination will grow as agents become fundamental participants in payments, order fulfillment, and protocol operations.
If ICP’s growth model for native AI hosting continues, on-chain inference cycles could reduce latency by half by 2026. This would make agent interaction viable for user-facing applications like intent routers, real-time hedging, and supply chain or IoT settlements.
| Protocol | Primary AI function | On-chain volume or TVL, Sept 2025 | Forward focus |
|---|---|---|---|
| Ethereum | Agent identity and settlement, ERC-8004, dAI Team | $38B+ | Trust and coordination layer for agents |
| Bittensor, TAO | Decentralized training and inference markets | $1.4B est. | Open AI compute exchange |
| Fetch.ai, FET | Autonomous economic agents, dApp infrastructure | $640M est. | Machine-to-machine coordination |
| Render, RNDR | Decentralized GPU and inference | ~$985M | Compute backbone for on-chain AI |
| Internet Computer, ICP | Native on-chain AI app hosting | $800M+ | Lower latency for agentic dApps |
| Blackhole DEX, Avalanche | AI-governed AMM and liquidity | $193M | Permissionless agent trading |
Three Potential Scenarios
In a base scenario, Ethereum solidifies its position as the primary identity and trust layer, with at least a quarter of new dApps adopting agent automation by 2026. This would integrate governance, treasury management, fee structures, and payments into programmable policies, secured by verifiable attestations.
In a bullish scenario, a fully developed machine economy emerges, where agents handle bilateral negotiations and order fulfillment across both consumer and enterprise sectors. This would lead to DeFi TVL exceeding $300 billion, and decentralized AI API marketplaces achieving widespread adoption for long-tail services.
In a bearish scenario, regulatory licensing of agents and ongoing centralization of computing power and model access could restrict open participation and limit innovation to a few well-funded organizations.
DLA Piper’s analysis and policy trackers indicate that transparency and control standards, rather than outright prohibitions, are the key determinants. However, the centralization of computing resources remains a significant concern.
Investors and Developers Focus on Measurable Adoption Triggers
Regarding standards, ERC-8004 is a crucial element to watch, as digital wallets and custody providers will need to implement attestation verification, recovery processes, and policy enforcement to ensure agents can operate safely in consumer contexts.
On the payment side, scaled-up AP2 trials that extend into cryptocurrency rails could establish a repeatable model for subscriptions, usage-based billing, and order fulfillment between non-human entities. This would push bridges and account abstraction technologies to offer granular limits and approval mechanisms.
From a security perspective, tangible evidence that adaptive controls reduce real-world losses would encourage more autonomous governance, especially for parameter adjustments in volatile markets. Each of these areas has publicly visible milestones that can be tracked independent of price fluctuations.
The fundamental question is not simply whether agents will transact, but where the transaction settlement and security checks will occur.
If identity verification, attestations, and policy enforcement happen on the blockchain, the machine economy will naturally gravitate towards public ledgers, and DeFi will become the operating system for automated economic activity. If these checks remain within closed platforms, crypto’s role will be reduced to just bridges and payout systems.
With Ethereum’s dAI initiative, the AP2 pathway for agent-based payments, and a clear increase in developer hiring in AI-crypto roles, the focus is shifting towards verifiable, on-chain coordination that treats agents as integral participants in the financial ecosystem.