AI agents could become a new source of demand for stablecoins as autonomous software begins paying for data, services and computing power on its own, according to a research paper published by BlackRock, the world's largest asset manager.

The paper, released in mid-September 2026, said the growth of agentic artificial intelligence — software systems that can act and transact with minimal human oversight — could create demand for digital assets by enabling payments to be made without direct human involvement. For stablecoins, whose use has so far been concentrated in crypto trading and transfers, the analysis frames autonomous machine activity as a potential new category of demand alongside their existing roles.

Stablecoins Seen as Leading Machine Payments

BlackRock said stablecoins are well suited to machine-to-machine payments because they can settle transactions around the clock and support small-value payments that can be difficult to process economically through traditional payment networks. Those two characteristics — continuous settlement and low-value transfers — are central to why the paper identifies stablecoins, rather than other digital assets, as the likeliest medium for payments initiated by software rather than people.

“Several types of digital assets may support agentic commerce, but stablecoins are likely to lead transactional use,” the paper states.

The thesis is consistent with findings reported by BitKE in March 2026, which observed that AI agents chose stablecoins for everyday transactions such as micropayments and cross-border transfers more often (53.2%) than Bitcoin (36%), while roughly 80% of agents favored Bitcoin as a long-term store of value:

REPORT 80% of AI Agents Choose Bitcoin as a Long-Term Store of Value

For everyday transactions such as #micropayments and cross-border transfers, #stablecoins were actually #chosen more often (53.2%) than Bitcoin (36%). #AIAgents pic.twitter.com/FziHClFoQf

— BitKE (@BitcoinKE) March 6, 2026

The BitKE data, published months before BlackRock's paper, offers one of the earliest published snapshots of how agents split their use of digital assets between day-to-day payments and long-term holding.

Stablecoins — blockchain-based tokens typically pegged to fiat currencies such as the US dollar — have so far been used mainly for crypto trading and transfers. BlackRock estimates that adjusted stablecoin transaction volume exceeded $11 trillion in 2025, while stablecoin market capitalization had risen above $300 billion by September 2026. The firm cautioned, however, that current AI-agent payment activity remains at an early stage, with autonomous agents still accounting for only a small share of observed machine-payment activity. That share is one of the clearest measures available for tracking whether agentic commerce grows beyond its early footprint.

Computing Power as a Potential Market

The paper also points to computing power as a potential market for digital assets. As demand for AI infrastructure grows, BlackRock said standardized claims on computing capacity could eventually be traded, used as collateral and settled on blockchain networks — effectively turning access to computing resources into a transferable claim that software itself could hold and exchange.

The firm cited third-party estimates that combined annual revenue from the cloud businesses of Amazon, Microsoft and Google could reach about $1.1 trillion by 2030 — a figure BlackRock highlighted to illustrate the scale of the computing market that AI agents could potentially access.

Early Tests Already Underway

The concept is already being tested through payment protocols such as Coinbase's x402, which allows software to pay for APIs, data and other services using stablecoins. Uptake of protocols like x402 is among the earliest observable evidence of whether machine-initiated payments can function in practice.

BlackRock's broader thesis remains forward-looking, however. If it plays out, the development could shift stablecoins beyond their current role in crypto trading and transfers toward becoming a payment layer for software that increasingly operates, purchases services and manages resources on its own.