BlackRock forecasts AI agents will pay via stablecoins on Ethereum and Arc

This document commits no BlackRock capital and introduces no product, so it is not a direct flow signal; any market response is more likely to reflect sentiment than new buying. It does, however, reinforce the institutional credibility of the AI-and-crypto overlap, a theme that often benefits tokens at the settlement layer and stablecoin issuers in narrative terms. The expected shift is still years off. Near-term positioning in ETH will probably be guided by ETF inflows and macro updates rather than by agentic-payment projections.

---

BlackRock's report is titled The Machine-Native Economy.

Previous coverage:

• BlackRock's Ethereum ETF purchase pace fell by a third in a week despite an ETH rally

---

BlackRock's bet is that the next major stablecoin customer will be a machine rather than a person, but the agents described in the report do not yet show up in the data.

Summary:

• BlackRock's digital-asset research team released The Machine-Native Economy this week, holding that AI agents will need payment rails built for machines and that stablecoins will likely take the lead in transactional use.
• Ethereum and Circle's Arc are named as settlement venues, and Coinbase's x402 is cited as an emerging standard for payments between machines.
• Stablecoin circulation exceeded $300 billion as of September, and adjusted volume was above $11 trillion in 2025, an 80% CAGR since 2020 versus about 8.5% for ACH, which still processed about $93 trillion.
• The often-repeated $5 trillion is an estimate of cumulative AI capital spending from 2025 to 2030, not a stablecoin market size.
• BlackRock envisions tokenised compute contracts and later exchange-traded compute futures, with hyperscaler cloud revenue near $1.1 trillion by 2030.
• The writers concede that agent-driven payment activity and compute market liquidity are still limited.

BlackRock's new research report argues that autonomous AI agents could become a meaningful new source of demand for stablecoins and for the blockchains that settle them. Ethereum and Circle's Arc are identified as possible settlement venues. The Machine-Native Economy was published this week by the firm's digital assets research team. The paper treats AI as machine-native intelligence and digital assets as machine-native money.

The logic is simple. An AI agent handling a job, such as booking travel or running a long analysis, may need to pay over and over for API calls, data feeds and computing power. Those payments are often less than a cent and can occur at any hour. BlackRock's case is that card networks and ACH do not fit that pattern well: opening an account may need human involvement, merchant charges make micro-payments uneconomic, and settlement does not happen instantly. The firm flags x402, an open payment protocol built by Coinbase that lets machines pay in stablecoins such as USDC, as one emerging standard.

BlackRock reinforces the case with scale data. Stablecoin supply stood above $300 billion as of September, and adjusted transaction volume passed $11 trillion in 2025. That volume has grown at an 80% compound annual rate since 2020, compared with about 8.5% for ACH. ACH still handled about $93 trillion last year, though BlackRock says its stablecoin and card-network metrics are not built for direct comparison.

The $5 trillion figure circulating in headlines needs context. It represents outside estimates of cumulative AI capital spending from 2025 to 2030, not the stablecoin market's size. BlackRock uses it to build a second argument: compute is turning into an investable resource. Consensus projections put combined revenue at Amazon Web Services, Microsoft's Intelligent Cloud and Google Cloud near $1.1 trillion by 2030. BlackRock says compute contracts, once standardised, could be tokenised, put up as collateral and ultimately traded as exchange-traded futures; Stripe's August agreement to buy OpenRouter is cited as an early signal.

When it comes to Ethereum, user benefits and the demand for ETH are two distinct things. A greater volume of stablecoin settlement on Ethereum could boost demand for blockspace and validator services. BlackRock adds a caveat: whether value is captured depends on a network's fee structure, staking mechanics and gas-sponsorship design. On Arc, fees are paid in USDC itself, so expansion there would add to USDC's usefulness rather than create demand for a separate token.

The document is a research thesis, not a purchase, product rollout or fund filing, and the authors concede that agentic payment activity and compute market liquidity are limited. FinTech Weekly observed that the customers BlackRock writes about have not arrived yet, and that tokenised bank deposits might undercut the argument for stablecoins as the default.

The next test is whether payments driven by agents appear in measurable on-chain volumes, and on which networks. Ethereum fee revenue rising together with stablecoin growth would make the ETH value-capture story stronger, while activity clustering on purpose-built chains like Arc would weaken it. For now, the paper is best approached as a long-horizon adoption story, not as a near-term price catalyst.