Once AI becomes an economic entity, why does Ethereum become the optimal financial solution for Agents?

Author: Etherealize

Translation: Felix, PANews

The AI Agent economy is exploding. Etherealize has published an in-depth article pointing out that Ethereum is the only blockchain capable of providing its financial infrastructure—without human identities, low-cost, and composable—that supports this growth. Below are the details.

In early 2026, an AI Agent named Felix generated over $300k in revenue within five weeks. Felix employs other AI Agents and operates multiple business lines, including Iris for customer support and Remy for sales. He sells a continuously updated deployment guide for AI Agents at $29; he also built and runs Claw Mart, a marketplace for developers to buy and sell pre-built AI skills and workflow templates. Additionally, he customizes AI Agents for companies needing content marketers, customer service reps, or sales assistants. His total operating costs are roughly $1,500 per month.

Felix can write code, deploy websites, manage sales pipelines, and respond to customer support emails—all without human help. But he cannot open a bank account. Felix’s creator, Nat Eliason, must personally create a Stripe account and hand over the API keys to Felix. The income Felix earns can only be idle, as he cannot open brokerage accounts to invest these funds or raise capital to start new ventures. Traditional financial systems assume that every account, credit application, or signature involves a human on the other end. But Felix is not human.

However, when Nat allows Felix to do things with cryptocurrency, “he can do it effortlessly, it’s just so simple.”

Felix is not an isolated case. For example, last week, Marc Andreessen stated on the Latent.Space podcast:

“I think AI is the killer app for cryptocurrency… It’s now clear that AI Agents need funding. That’s already happening… My most aggressive friends using OpenClaw have already handed their bank accounts and credit cards over to their Claws. They’ve done that, and it’s obvious they need to do that… It’s completely obvious. I don’t know how many people are doing it today, maybe around 5,000. But it will grow. That’s how these things start.”

Felix is an experiment. It’s too early to tell whether his revenue is sustainable or just a burst of early-stage growth. But the pattern he represents—a self-sufficient Agent that can make money, spend money, and requires financial services—will recur, whether Felix himself endures or not. Human reliance on their own financial identities is merely a temporary expedient. Ultimately, they will use the Ethereum financial system we’ve only been building for ten years.

Agents are already transacting

So far, discussions around AI Agents and cryptocurrency have almost exclusively focused on payments. Coinbase, Cloudflare, and Stripe have established a foundation to govern x402, an open protocol enabling Agents to make instant stablecoin micro-payments. Stripe and Paradigm have also launched the Machine Payments Protocol on Tempo (a blockchain built specifically for stablecoin settlement).

Data is already substantial. In the first nine months, x402 handled over 140 million Agent-to-Agent transactions, totaling $43 million in volume. x402 now accounts for about one-fifth of the traffic on Coinbase’s Base network. Nearly 16k verified Agents are running on-chain, recording over 400k individual buyer addresses.

Agents will accelerate the shift toward native crypto payments because traditional card networks are structurally incompatible with Agent commerce. According to the “2026 Agent Status Report,” the average transaction amount between Agents is $0.31, mainly for API calls, computation, and data access. At this scale, a fixed fee of about $0.30 per transaction—like Visa’s—would consume nearly all the payment amount.

But payments are the simplest financial function. The more interesting question is: what happens when some Agents go beyond mere payments and start managing the funds held between transactions?

What kind of DeFi do Agents need?

Most Agents will never need the financial system. Customer support Agents acting on behalf of companies won’t hold vaults, nor will coding Agents. These are tools operated within the deploying companies, handled by the companies themselves for financial matters.

The Agents that need DeFi are those operating as autonomous economic actors: Agents with their own income streams, expenses, vaults, and unable to access financial services due to lack of human identity. This subset is smaller but growing. As Agents become more powerful, longer-lived, and more autonomous, the number of Felix-like Agents will grow from hundreds to thousands, then millions. Coinbase CEO Brian Armstrong believes AI Agents will eventually surpass humans. Even if only a small fraction act as autonomous economic entities, the total capital they manage will be significant. The question then becomes: what financial services does an autonomous Agent require?

Needs for borrowing: operational capital for computation, bridging cash flow gaps, or funding new projects. Traditional lending requires credit applications, underwriters, and legal identities, but on Aave, Agents can deposit collateral and borrow stablecoins instantly without human intervention.

Needs to generate yield from idle funds: Felix has over $165k in capital and (in Nat’s words) “doesn’t know what to do with it.” On Ethereum, these funds can be deposited into lending protocols to buy tokenized government bonds like BlackRock’s BUIDL, or deployed as liquidity on Uniswap—all permissionless, instant, and composable. Tokenized government bonds on Ethereum are growing rapidly, with over $22.5 billion in fund assets tokenized (accounting for 71.9% of all blockchain markets). J.P. Morgan launched its MONY market fund on Ethereum in early 2026, joining BlackRock’s BUIDL and Franklin D. Dunst’s on-chain money market funds. These institutional-grade products are exactly what autonomous Agents with idle capital need, operating on permissionless infrastructure accessible without brokerage accounts.

Needs to raise capital: Felix cannot set up a Carta account or initiate wire transfers from Mercury, but he can deploy a smart contract that issues tokens representing revenue shares, accept stablecoin investments, and programmatically manage distributions. The legal framework for this is taking shape, and the “Digital Asset Market Clarity Act” marks a solid step forward in facilitating on-chain capital formation in the U.S.

Needs to make and receive payments: this is already happening at scale on L2 and Solana. But when Base pays settlement fees to L1, stablecoins are issued and redeemed on mainnet, and Agents need to hold earnings between transactions, Ethereum can capture value from these activities.

Needs custody of assets: equity tokens, governance tokens, stablecoins, identity credentials—without custodians, they can be frozen or recovered without counterparties. Self-custodied Ethereum wallets inherently provide this.

Why Agents use low-risk DeFi on Ethereum

Vitalik proposed in September 2025 that core financial services—payments, savings, lending, borrowing—are Ethereum’s most important applications. His key insight was that for increasing numbers of participants in the global economy, the tail risks implicit in traditional finance—bank failures, account freezes, capital controls, counterparty defaults—are now surpassed by the tail risks of proven DeFi protocols. This refers to individuals lacking reliable jurisdictional financial institutions, but the argument is even more relevant for Agents. Agents will prefer DeFi not only because it reduces counterparty risk but because it’s inherently a better financial system for machines.

In DeFi, transaction costs are just a few cents, not a few percentage points. Settlements take seconds, not days. The system is frictionless worldwide. And each protocol’s rules are encoded in open, auditable code, allowing Agents to verify them before committing funds.

There’s an irony here. Smart contracts have always been awkward for humans, and user experience has been a persistent challenge. When Nick Szabo introduced the concept in 1997, he described contract logic embedded directly in machines, executing automatically based on conditions, without human intervention. This vision has never truly suited human users, who prefer human intermediaries when issues arise. But it fits perfectly for Agents.

An autonomous Agent with a $500k treasury will need something akin to a money market fund—predictable yields, deep liquidity, minimal smart contract risk, and no counterparties that can freeze or seize assets. Ethereum’s DeFi increasingly meets these standards. Hacks and losses still occur, but they are becoming rarer and more concentrated on the fringes of a highly speculative ecosystem. A core set of stable, resilient applications has proven its robustness through repeated stress events—something no other blockchain has replicated.

Ethereum’s L1 DeFi losses. Source: Vitalik Buterin

DeFi eliminates a whole class of risks for Agents. Rules are encoded in auditable smart contracts; collateral ratios are automated; no counterparties can freeze, seize, or renegotiate. For software-native participants, this is a better architecture.

Other blockchains have DeFi protocols too. Any team can fork Aave and deploy lending protocols on new chains. But building a long-term, trustworthy DeFi ecosystem that can attract significant capital is a different matter entirely.

As Erik Voorhees said: “Ethereum remains king. People are distracted by other L1 platforms, but if you look at where developers are and where stablecoin trading volume is, those metrics are hard to fake and very important—they’ve always been mainly on Ethereum. The gap is very clear.”

Ethereum’s DeFi has now formed an almost unassailable network effect:

Protocol maturity. Aave launched in 2020; MakerDAO has maintained its DAI peg through multiple market crashes since 2017. Uniswap’s cumulative trading volume exceeds $3 trillion. These protocols have operated flawlessly through black swan events like Terra/Luna and FTX. For investors holding funds for six months, the difference between five years of stress-tested protocols and two years is crucial. Rational investors weigh past performance when choosing where to deploy capital.

Liquidity depth. Low-risk lending requires deep pools. If an Agent deposits $10 million collateral on Aave and borrows $7 million in stablecoins, the pool must be sufficiently deep to handle this without significant slippage or rate impact. Ethereum’s DeFi pools are several times larger than any competitor’s. As of April 2026, Ethereum’s DeFi TVL exceeds $55 billion—almost ten times Solana’s—and accounts for 57% of all blockchain market share.

Institutional participation. BlackRock chose to launch BUIDL on Ethereum. Franklin D. Dunst selected Ethereum for its on-chain money market fund. Ethereum hosts about 71% of tokenized funds. These institutions conduct extensive due diligence before choosing a blockchain. Their participation creates a self-reinforcing cycle: deeper liquidity attracts more institutional capital, further deepening the market. Institutions seeking the lowest-risk DeFi environment tend to favor blockchains with the most concentrated institutional capital, as this creates deeper markets, better auditing, and clearer regulation.

Network reliability. Ethereum has never experienced a downtime in over a decade of operation. Hundreds of thousands of validators secure the network, making censorship of individual transactions nearly impossible.

Composability. On Ethereum, a trader can deposit ETH into Aave, borrow USDC, and deploy that USDC into a tokenized government bond fund—all in a single transaction. If any step fails, the entire sequence reverts. There’s no partial execution or counterparty risk. This composability exists because all major DeFi protocols share the same state on the same chain, and as traders execute increasingly complex multi-step financial strategies, their value compounds.

57% of DeFi TVL is on Ethereum (Source: DeFi Llama)

What does this mean for ETH?

Autonomous Agents primarily transact using stablecoins. 98.6% of Agents pay in USDC. But every interaction with Ethereum’s DeFi stack—borrowing on Aave, swapping on Uniswap, deploying smart contracts, rebalancing portfolios—requires paying gas fees in ETH.

An Agent with $1 million in collateral will use Ethereum L1 because of the strongest security guarantees, and it will be happy to pay gas fees. These costs are trivial compared to risk capital. As Agent DeFi activity grows, Ethereum’s block space will become increasingly valuable, and EIP-1559 means a portion of each gas fee is burned, permanently reducing ETH supply.

Moreover, as Vitalik pointed out, the economic contribution of low-risk DeFi to ETH isn’t just in transaction fees but also in locking ETH as collateral. Agents borrowing stablecoins on Aave must provide collateral, and ETH is the deepest, most liquid collateral asset on the network. The more Agents borrow, the more ETH is locked in lending protocols, reducing circulating supply without relying on burning mechanisms.

It’s hard to precisely estimate this structural demand. Frankly, it depends on how many Agents evolve into autonomous economic entities, the scale of their managed capital, and how much capital flows through Ethereum’s DeFi. But the direction is clear: the Agent economy is growing, Ethereum is the only system capable of serving large-scale autonomous participants, and every transaction on that system requires ETH.

Potential issues

Three factors could weaken this thesis, and it’s worth clarifying them.

First is gas abstraction. Account abstraction and payment Agents allowing stablecoins to pay gas could reduce demand for ETH as operational capital. If this becomes standard, the need for ETH would decline. But some on-chain functions still require ETH to process transactions.

Second is competition. If other blockchains or L2s achieve the liquidity depth, protocol maturity, and institutional influence Ethereum has, DeFi participants might disperse their activity across chains.

Third is the transformation of traditional finance. Banks will eventually create APIs for Agent accounts, and brokerages will build machine-accessible interfaces. But even with adjustments, traditional finance products designed for humans will still carry human costs, whereas DeFi offers software-native solutions.

Overall, the bullish case remains stronger. Gas abstraction shifts demand for ETH within the ecosystem rather than eliminating it; competitors’ DeFi ecosystems lag behind Ethereum by years in the specific attributes needed for low-risk DeFi; and the inefficiencies of traditional finance are hard to overcome. Still, these risks should be carefully weighed.

Ethereum’s next billion users won’t be humans

Ethereum is moving toward becoming a machine economy’s financial system. It’s the only platform capable of providing the financial services—lending, yield generation, capital formation, custody—that autonomous Agents need, without requiring human identity verification, human labor costs, or jurisdictional access restrictions.

As the number and complexity of Agents increase, those evolving into autonomous economic entities will continuously demand low-risk DeFi on Ethereum. Every transaction they execute consumes and destroys ETH. They rely on Ethereum’s financial infrastructure because no other blockchain offers the liquidity, maturity, reliability, and institutional support necessary for low-risk DeFi.

Related: Galaxy Research: The Era of Zero-Human Companies—How AI Agents Will Activate the On-Chain Financial Flywheel?