BlackRock's Digital Assets Research team published a paper during the week of September 22, 2026 arguing that stablecoins will become the payment infrastructure for autonomous AI systems conducting machine-to-machine commerce.

The paper, titled "The Machine-Native Economy: How digital assets connect intelligence, commerce, and compute," makes a structural case: AI agents require programmable settlement, legacy payment rails cannot deliver it, and dollar-pegged tokens are best positioned to fill the gap. The argument is targeted at institutional readers evaluating where to deploy capital as agentic AI moves from research labs into production environments.

Why ACH and card networks fall short

BlackRock's researchers identified traditional payment systems — including ACH transfers and credit card networks — as fundamentally misaligned with the low-cost, high-frequency transactions autonomous AI systems will produce. Settlement times, fees, and intermediary requirements that work for human commerce become structural bottlenecks when an algorithm purchases GPU compute time or data access in real time.

The scale behind the thesis

The data cited in the report anchors the magnitude of the shift. Stablecoin circulating supply crossed $300 billion by September 2026. Adjusted transaction volume topped $11 trillion in 2025 alone. Since 2020, stablecoin transaction volume has compounded at roughly 80% annually. Over the same period, ACH volume grew at about 8.5% per year — a roughly tenfold gap in growth rates that BlackRock's team frames as evidence of structural demand, not cyclical momentum.

What share of payment activity is AI-driven today?

A TRM Labs analysis cited in the paper estimates AI agents currently account for somewhere between 0.6% and 7.5% of activity across leading payment protocols. The wide range points to an early but measurable footprint, not yet a mature market. BlackRock frames the lower bound as a baseline that will expand as agent frameworks reach production scale.

The report also quantifies the underlying compute economy that will drive that expansion. Cumulative investment in AI infrastructure between 2025 and 2030 could exceed $5 trillion. Analysts project cloud industry revenues from Amazon, Microsoft, and Google to reach roughly $1.1 trillion by 2030.

Which stablecoins could capture the AI use case?

Tether's USDT and Circle's USDC dominate stablecoin volume today, but BlackRock's paper suggests the AI-payments use case could create demand for tokens optimized for specific characteristics: programmability depth, cross-chain interoperability, and compliance features tailored to enterprise AI deployments. The paper stops short of naming specific issuers but signals that market segmentation, not market consolidation, is the likely next phase.

The report also floats the possibility of an entirely new digital asset class — assets tied to tokenized computing capacity. Regulations and market standards for that segment remain unformed.

What is the regulatory pathway?

The paper acknowledges that US stablecoin legislation has been a slow-moving process and that standards for tokenized compute markets are still taking shape. That delay matters for institutional adoption, since enterprise AI deployments will require compliance features and legal certainty current frameworks do not yet provide.

The institutional consequence runs in both directions. Asset managers, custodians, and payment processors that build stablecoin infrastructure now position themselves for a transaction category BlackRock's analysts believe will compound at multiples of legacy rates. Firms that wait for clearer rules risk ceding the machine-to-machine settlement layer to early movers.

The next test arrives as US stablecoin legislation moves through Congress and as agent-payment standards — including emerging specifications for machine-initiated transactions — shift into enterprise deployment. BlackRock's research effectively converts a market forecast into a strategic brief for the world's largest pool of capital.