BlackRock Says AI Agents Could Spark a New Wave of Crypto Demand

Highlights

BlackRock's new research, titled "The Machine-Native Economy," suggests that as AI agents gain autonomy they will require continuous, tiny-value payments — a use case where stablecoins and on-chain settlement excel. The paper proposes tokenized claims on compute as a potential new asset class and notes growing stablecoin transaction volumes. While pilot systems and protocols (like x402) already exist, current agent-driven payments remain a small share of activity. BlackRock views the existing infrastructure as groundwork for much larger future demand from machine-to-machine payments.

Sentiment Analysis

• The overall tone of the article is cautiously optimistic. It emphasizes innovation and potential market growth tied to AI-driven, machine-native payments while acknowledging current usage is limited. The piece highlights technical and economic reasons stablecoins may be better suited than traditional rails for sub-cent, 24/7 automated payments. It also frames tokenized compute as a plausible, emerging asset class, and notes sizeable cloud-revenue projections that could underpin demand for such instruments. Risks and adoption hurdles are acknowledged but not heavily stressed.

Article Text

BlackRock's Digital Assets Research team recently published a paper arguing that the next meaningful driver of crypto demand could come from artificial intelligence systems that pay autonomously. The report—presented under the framework "The Machine-Native Economy"—posits that agentic AI, meaning systems capable of planning and executing multi-step tasks with little human intervention, will increasingly need to make frequent, very small payments for services such as API calls, data feeds, or compute resources. Traditional banking and card networks, designed around human account holders and fee structures that suit larger transactions, are poorly matched to this type of machine-to-machine commerce.

To address that gap, BlackRock highlights stablecoins—cryptocurrencies pegged to fiat values—as an efficient instrument for continuous, low-value settlement. Stablecoins can transfer value 24/7, settle rapidly, and avoid the overhead and identity requirements tied to opening bank accounts or issuing cards. The paper suggests these attributes make stablecoins a natural fit for payments that agents will perform autonomously on behalf of users or services.

BlackRock also explores the idea of tokenized claims on compute capacity. As demand for cloud resources rises, the report outlines how compute could be standardized into tradeable contracts—similar to commodities or futures—that an AI agent might purchase, sell, use as collateral, or settle automatically on-chain. In this envisioned ecosystem, an agent could compare offerings, select the most cost-effective server capacity, and pay for it programmatically, reducing friction and manual contracting.

Some of the supporting infrastructure is already in development. Protocols such as x402, which implements a payment-on-request flow modeled on the HTTP 402 concept, enable software to pay for a data or service request within the same transaction. Large cloud providers and payment platforms have experimented with integrations that let agents pay for API calls or services mid-task using stablecoins or other digital payments. For example, partnerships and pilots have been reported where cloud or AI tooling accepts on-chain or tokenized payments, and major firms have explored agent-to-agent payment layers that handle cards, stablecoins, and instant transfers.

Despite these technical demonstrations, real-world agent-driven payments remain a small fraction of current activity. Blockchain analytics cited in the report show that while protocol volumes exist, a relatively modest percentage is attributable to genuine autonomous agents, with much of the traffic resembling ordinary automated scripts. BlackRock doesn’t claim the transformation is imminent at scale; rather, the firm argues the plumbing and standards are being put in place now, ahead of a future inflection point when autonomous agents become far more prevalent.

Economics underpin the thesis. Cloud computing is costly and demand for compute continues to grow rapidly, with large cloud providers projected to generate substantial revenue in the coming years. If compute can be commoditized and transacted programmatically, it creates a recurring need for reliable, programmable settlement that can operate at high frequency and low value. BlackRock’s figures point to robust growth in stablecoin transaction volumes over recent years, and the report suggests AI-driven machine payments could be a significant new source of demand that further accelerates on-chain activity.

While the paper is forward-looking, it notes several adoption challenges. Regulatory clarity, interoperability between payment rails and blockchains, security of automated payment flows, and the distinction between simple automated scripts and truly agentic systems all affect the timeline and scale of adoption. Nonetheless, the report sees a credible pathway in which stablecoins and tokenized compute play central roles in enabling a machine-native financial layer. The central insight is that money and intelligence that are native to machines naturally complement each other.

In sum, BlackRock views the convergence of agentic AI and programmable money as an emerging structural trend. It does not promise immediate mass adoption, but it frames existing experiments and growing transaction volumes as preparatory steps toward a future where autonomous systems routinely handle procurement, settlement, and resource allocation without human intervention.

Key Insights Table