BlackRock’s Machine-Native Economy: Why AI Agents Might Pay in Stablecoins

BlackRock’s The Machine-Native Economy report presents a possible next step for digital commerce: autonomous AI agents buying data, software access and computing capacity from other services, with programmable digital assets helping them pay. The idea is plausible as a direction for payment infrastructure, but current evidence does not show agent-to-agent commerce at scale. 211

Why BlackRock thinks agents could use stablecoins

An AI agent may need to make small payments automatically while completing a task—for example, to access an API or data feed. BlackRock’s case is that digital payment systems designed for software could make these transactions easier to initiate and settle programmatically. Conventional payment rails are not ruled out: the report also discusses adapting them for agent workflows. 11

Stablecoins are a proposed fit because they aim to maintain a steadier value than volatile cryptocurrencies while remaining usable on blockchain networks. That can make the cost of a purchase easier for software to assess. This does not make stablecoins risk-free: issuer, regulatory and blockchain risks remain. 15

The distinction is not simply “blockchains work, traditional payments do not.” The report describes both emerging blockchain protocols and adaptations to traditional rails. Its broader thesis is that machine-to-machine commerce could benefit from payment infrastructure that software can access and use without a human handling each transaction. 11

What Ethereum, Arc and x402 could do

These names describe possible pieces of the infrastructure, not a settled set of winners.

• Ethereum is discussed as a potential blockchain rail for programmable settlement and digital assets. 7
• Circle’s Arc is presented as another possible rail, oriented toward stablecoin payments. Its mention is not evidence that it will become a dominant network. 7
• Coinbase’s x402 uses the web’s “Payment Required” mechanism to let software request a resource and make a payment as part of that flow. It is one example of how API or data access could be paired with a machine-initiated payment. 1012

The report also frames computing capacity as a possible digital-asset use case. If claims on compute were standardized and tokenized, they might be bought, financed or settled programmatically. That is a proposed market structure, not a claim that cloud providers’ revenue will itself be tokenized. 911

What the market figures show—and what they don’t

BlackRock cites stablecoin market capitalization above $300 billion and adjusted stablecoin transaction volume above $11 trillion in 2025. Those figures show that stablecoins already have substantial market activity. But adjusted volume is not directly comparable with traditional payment-network spending, and neither figure measures payments made by AI agents. 213

The report also points to a projection of roughly $1.1 trillion in combined 2030 revenue for AWS, Google Cloud and Microsoft Intelligent Cloud. That projection suggests the scale of the cloud businesses around which future compute markets could develop; it does not show how much computing capacity agents will buy or how much of it could be tokenized. 613

The evidence gap: agent payments remain limited

Reports about x402 and stablecoin payment infrastructure show that companies are exploring ways to support software-driven transactions. Reports of agents on the XRP Ledger using Ripple USD (RLUSD) also illustrate why an agent might prefer a stable-value asset to a volatile native cryptocurrency for payments. But those reports do not establish broad adoption across AI agents or payment networks. 523

A more direct test comes from TRM Labs’ analysis of x402 activity. The firm estimated that likely autonomous agents accounted for 0.6%–7.5% of screened x402 payment value. That is a share of activity on one protocol, not a measurement of every AI-agent payment—but it underscores how much of the broader thesis remains unproven. 30

The strongest reading of BlackRock’s report, then, is as a forecast about what payment infrastructure might be useful if autonomous agents begin transacting widely. Stablecoins, blockchain settlement, protocols such as x402 and tokenized compute could serve that future. Today’s market figures support the existence of relevant infrastructure and a large cloud sector; they do not yet demonstrate large-scale agent commerce.