Investinglive iconInvestingliveSep 23, 2026 ~6 min source read

BlackRock paper: AI agents could pay in stablecoins on Ethereum and Circle’s Arc

BlackRock’s digital assets team argues autonomous AI agents may need machine-native payment rails, naming stablecoins and settlement venues such as Ethereum and Circle’s Arc, but it presents this as a research thesis rather than a commitment of capital.

Share this story

Send the public story page.

Useful takeaways from this story.

BlackRock’s research paper, The Machine-Native Economy, proposes that autonomous AI agents could create new stablecoin demand for machine-to-machine microtransactions.

The paper highlights Ethereum and Circle’s Arc as potential settlement rails and cites Coinbase’s x402 protocol as an emerging machine-payment standard.

BlackRock frames tokenised compute as a potential investable resource, but notes agentic payment activity and compute-market liquidity remain limited.

# What BlackRock proposes

# Why stablecoins and blockchains

# Protocols and early standards

BlackRock calls out Coinbase's x402 protocol as an emerging open standard for machine-to-machine stablecoin payments. The mechanism described is straightforward: an agent executing a task that requires repeated micro-payments can transact autonomously using stablecoins on programmable rails, avoiding manual account setup and merchant fee structures that make tiny payments uneconomic.

# Market context and numbers cited

The paper uses public-scale data to support the thesis. It notes stablecoins had more than $300 billion in circulation as of September and that adjusted transaction volume exceeded $11 trillion in 2025, which it says grew at roughly an 80% CAGR since 2020. BlackRock contrasts that with ACH volume, which it places near $93 trillion in the referenced year, and cautions the figures are not directly comparable.

# Tokenised compute as an investable resource

BlackRock proposes a second, related idea: compute itself could become tokenised and tradable. The paper points to consensus estimates that combined hyperscaler cloud revenue could approach $1.1 trillion by 2030 and suggests standardised compute contracts might be tokenised, used as collateral, and ultimately traded as exchange-traded compute futures. The paper cites early signals such as Stripe's acquisition of OpenRouter to illustrate commercial momentum.

# Limits and near-term outlook

The authors explicitly treat the work as research, not a fund launch or capital commitment. They acknowledge that agentic payment activity and compute-market liquidity are currently limited and that the customers described have not yet shown up at scale. They also note that near-term price or ETH demand dynamics are more likely to be driven by ETF flows and macro data than by a gradual build of agentic payments.

# Practical implications for readers

If autonomous agents do scale, expect attention to settle on stablecoin issuers, settlement-layer tokens and protocols that support machine-to-machine standards. But timelines extend over years, and BlackRock's paper should be read as a forward-looking thesis that could influence sentiment rather than an immediate source of capital inflows.

More context around this story.

Cryptocurrency Functions!

Cryptocurrency Functions!

A cryptocurrency is a decentralized unit of communication that uses encryption to protect transactions and influence new modules’ development. Cryptocurrencies are a type of alternative currency, more principally a type of digital currency. Bitcoin was just the initial financial intermediary in 2009, but even now it sh

Loading more related stories...

Keep reading in the app

Open the app view to save this story, compare related coverage, and continue from the same source.

Open in app