Ethereum is positioning its base layer to coordinate autonomous agents, a movement that puts the machine, the trade machine on a direct route towards the settlement in the chain in next year.
This month, the Ethereum Foundation formed a DAI team dedicated with a mandate to advance the identity of the agent, trust and payments, including support for ERC-8004, a standard draft for agents credentials and verification that would anchor identity and baskets at the protocol level.
The initiative frames Ethereum as a layer of settlement and coordination for agents’ economies, with open censorship resistance and access as central design objectives, while community drafts around ERC-8004 describe how identity and confidence in the chain could allow automated systems to negotiate, publish ties and execute the custody without custody intermediaries.
The short -term delivery is the progress of the investigation and the standards that can be adopted by wallets, Middleware and DAPPS in 2026, creating a shared trust substrate for agent applications.
Tokens flows already reflect an inclination of AI in cryptography markets.
The tokens focused on AI as bittensor, fetch.ai (like), Internet computer and render have maintained the activity in the chain and the relative stability of prices until the third quarter, overcoming the largest Altcoins during the recent market reduction.
Koinly market summaries point to the continuous demand for decentralized computing, inference and agents frames, while ecosystems reports show the impulse of ICP for the accommodation of native applications and the render GPU market, causing a constant use of AI workloads.
According to tokens metrics, the total value of blocked defi has been recovered from approximately $ 72 billion in early 2025 to the area of $ 100 billion, with new AI defining rails such as Blackhole Dex in Avalanche, Sahara AI and Moby Ai that contribute to volume and generation of fees that pursued through volatility. Tokens metrics place this in a broader rotation towards automated liquidity and the execution of agents that can operate through the chains through Omnichain messages and abstractions.
The payment pile is converging in cases of use of agents in the limit of the protocol. Google introduced agents in the Payment Protocol, or AP2, in September to allow software agents to request and confirm consumer payments through standardized flows, a construction block for billing patterns and subscription of machine by machine that can interact with cryptographic settlement rails.
According to Google Cloud, AP2 is designed around the explicit consent of users, identities of verifiable agents and reversible transactions for compliance, and the first pilots include integrations of Ethereum and ICP through third -party connectors that join Fiat accounts with chain transfers.
As these pilots mature, wallets can treat agents as first-class actors, with ERC-8004 style certifications that allow policies that limit spending by time frame, restrict counterparts or require human co-firms for high value thresholds.
Advance models now unite plumbing updates to measurable network demand.
September stage work projects of Token Metrics Intelligent agents that reach 15 to 20 percent of the volume of transactions defi at the end of the fourth quarter, which, if the DAI Dai roadmap of Ethereum are maintained and amplified, places protocols integrated in AI in the TVL range of $ 200 to 300 billion dollars at the end of 2016.
The same analysis frames a food in the use of the base layer, with the use of gas for the identity and execution contracts of the agent that increase from 30 to 40 percent trimester in the quarter in 2026 once the standards such as ERC-8004 see a wide adoption through custody, consumption wallets and the DAO measurement.
In practice, this means governance, treasurement of the Treasury, rates and cross -chain liquidity management could be executed by software agents operating with risk limits, insurance and credentials verifiable in the chain.
The security results are another lever in the adoption curve. Academic and industrial research on adaptive contracts and assisted by INI points to a strong fall in successful exploits when contracts can detect anomalies, tune in parameters and in quarantine suspicious flows in real time.
The first models show reductions of up to 70 percent in successful attacks for systems that combine controls based on rules with learned heuristics, in relation to static parameter schemes. This result depends on the transparent update policies and monitorable behavior in the chain in the chain to avoid creating opaque control surfaces, a point that fits with supervision care on the auditability of the smart contract and incident reports.
The macro context is changing from the concept to the pilot.
Regulatory agendas in the United States and Europe include work transmissions on automated financial agents, transparency for adaptive contracts and disseminations on the risk of the model.
The brief and other legal trackers of September of DLA Piper describes a path where the identities of the agents, the policies of use and the management of exceptions must be legible for the regulators and the counterparts, a requirement that is aligned with the identity and the impulse of certification of Ethereum instead of contradicting it.
The recent issues of application emphasize the effectiveness of control, not the prohibitions of technology, which supports a track track for the operations of agents that meet as mature standards.
Data hiring remains supported, and Recrutblock registered an increase of 22 percent year after year in 2025 for roles at the intersection of AI and Blockchain, which covers protocol engineers, data infrastructure and applied cryptography, a pipe that matters if agents frames must reach the production scale throughout the consumer and the points of contact of the company.
Cross market, the machine’s economy lens is not limited to a single pile. Avalanche organizes liquidity governed by AI through Blackhole Dex, Ethereum focuses on identity and agreement, application accommodation in the nearby chain and ICP in the court and the inference of low latency, and represents GPU resources for training and models service.
The coverage of Koinly’s metrics and the token places them in complementary roles instead of direct substitutes, with a thesis that the demand for decentralized inference and market coordination expands as agents become predetermined actors in payments, realizations and protocol operations.
If the ICP growth model for the native to accommodation, the inference cycles in the chain could reduce the latency by half by 2026, which would make the interactivity of the agent viable for user -oriented applications such as intent rings, real -time coverage and supply chain settlement or IoT.
| Protocol | Primary function | Chain volume or TVL, September 2025 | Final focus |
|---|---|---|---|
| Ethereum | Agent identity and liquidation, ERC-8004, DAI team | $ 38b+ | TRUST AND COORDINATION LAYER FOR AGENTS |
| Bittensor, Tao | Decentralized training and inference markets | $ 1.4b est. | Open AI computing exchange |
| Fetch.ai, fet | Autonomous Economic Agents, DAPP Infrastructure | $ 640m est. | Machine -by -machine coordination |
| Render, LND | Decentralized GPU and inference | ~ $ 985m | Calculation of the spine for AI in the chain |
| Internet computer, ICP | Native to chain application accommodation | $ 800m+ | Lower latency for agent DAPPS |
| Blackhole Dex, avalanche | AMM and Liquidity of AI | $ 193m | Agent trade without permission |
The scenarios fall into three cubes.
A base case has Ethereum that consolidates the identity and trust layer, since at least a quarter of the new DAPPs adopt the automation of agents by 2026, converging governance, treasury, rates and payments in programmable policies anchored in the certifications.
A bull route lights a more complete machine economy where agents handle negotiation and bilateral compliance in the contexts of consumers and companies, with defi TVL that moves beyond the API markets of the decentralized the decentralized of $ 300 billion and that reach the critical mass for long -tailed services.
A bear case focuses on the regulatory licenses of agents and the continuous centralization of the calculation and access to the model, which would limit the participation and innovation of bottlenecks to a small number of well -financed equipment.
The general description of DLA Piper and policy trackers point to transparency and control standards such as the support point, not direct prohibitions, but calculating centralization remains a known restriction.
Investors and builders change tokens stories to measurable adoption triggers.
On the standards side, ERC-8004 is a central clock element, since wallets and custody suppliers must implement certification controls, recovery flows and policy application so that agents operate safely in consumer contexts.
On the side of the payments, the AP2 pilots, if they extend to the cryptographic rails at scale, would provide the first pattern repeatable for subscriptions, billing of use and compliance between non -human actors, and press bridges and stackings of abstraction of accounts to expose limits and approvals of fine grain.
On the security side, the field evidence that adaptive controls reduce the loss made would unlock a more autonomous governance, especially for the tuning of parameters in volatile markets. Each of these tracks has public milestones that can be monitored without depending only on price lists.
The open question is beyond whether agents will transact; It is where the agreement and trust controls occur.
If identity, certifications and policies live in the chain, the machine economy will breach public accounting books, and Defi will become the operating system for non -human economic activity. If those checks remain on closed platforms, the role of Crypto will collapse on bridges and payment rails.
With the DAI mandate of Ethereum, the AP2 road for agents payments and a measurable change in the hiring of developers towards the crypto roles of AI X, the center of gravity is moving towards the verifiable coordination in the chain that treats the agents as first class participants in the markets.