BlackRock says AI agents will need stablecoins and blockchain payments

BlackRock argues that autonomous AI agents will need blockchain-based payment rails and stablecoins to transact without human oversight, positioning digital assets as core infrastructure for an increasingly machine-driven economy. The world’s largest asset manager’s framing adds a new institutional voice to a debate already shaped publicly by Binance founder Changpeng Zhao and Bitmex co-founder Arthur Hayes.

• BlackRock published a paper on September 22, describing AI as “machine-native intelligence” and crypto as “machine-native money.”
• The firm identified three areas of overlap: tokenization, agentic commerce, and computing capacity.
• BlackRock cited analyst estimates that hyperscaler cloud revenue could exceed $1 trillion annually by 2030.

• $1T hyperscaler cloud revenue seen exceeding this annually by 2030
• Sep 22 date BlackRock published its paper linking AI agents to crypto demand

BlackRock published a paper on September 22, arguing that wider adoption of artificial intelligence could generate significant new demand for stablecoins and blockchain-based payment systems, according to reporting by CryptoPotato. The asset manager contends that autonomous AI systems acting with limited human involvement will need financial infrastructure designed for machines rather than people. It also suggests blockchains could become the mechanism AI systems use to pay for the computing resources they consume.

BlackRock frames AI as “Machine-native intelligence,” crypto as “Machine-native money”

The paper, posted alongside a summary on X, describes AI and digital assets as technologies that developed along separate tracks but are now converging. BlackRock says AI systems are gaining the ability to interact directly with financial networks and carry out transactions on their own.

The firm identifies three points of overlap: tokenization, agentic commerce, and computing capacity. Each, it argues, reflects a structural link between how machines process information and how blockchains represent value.

On tokenization, BlackRock draws a parallel between large language models, which break text into tokens for numerical processing, and blockchains, which represent ownership claims as digital tokens. The two functions differ, but both systems convert information into standardized formats that machines can read.

Agentic commerce could outgrow card networks and ACH, BlackRock argues

BlackRock’s second area, agentic commerce, centers on AI agents making financial transactions directly. The firm argues this could raise demand for programmable payment infrastructure, with stablecoins and other cryptocurrencies serving as settlement instruments.

Card networks and the Automated Clearing House already support automated payments. BlackRock’s paper argues their onboarding requirements and settlement economics make them poorly suited to continuous, very low-value transactions that require programmable execution.

The third area, computing capacity, ties directly to the $1 trillion projection. BlackRock cited analyst estimates that hyperscaler cloud revenue could exceed that figure annually by 2030, and argued that standardized, transferable claims on computing capacity could become a digital-asset use case for financing and programmable settlement as AI agents operate for longer periods and need guaranteed access to compute.

CZ and arthur Hayes have floated the same compute-token idea

BlackRock is not the first to link AI compute demand to crypto. Arthur Hayes has argued that AI agents consume floating-point operations, not groceries, and may eventually want a token redeemable for compute, as CryptoPotato has previously reported.

Changpeng Zhao told Galaxy Research in June that agentic trading and payments would arrive in months rather than years, and that they would run on crypto because blockchains already communicate through APIs. BlackRock’s paper does not name either figure, but its case for programmable settlement rails runs parallel to both.

The paper stops short of guaranteeing the shift will happen. BlackRock frames AI as a potential structural catalyst for digital-asset adoption and digital assets as possible infrastructure for an autonomous economy, but the argument depends on whether AI agents generate enough transaction volume to justify moving off existing payment rails.

The BlockWest read. BlackRock’s paper reads less like a forecast and more like positioning: a firm that already runs a tokenized money market fund and a spot bitcoin ETF is laying rhetorical groundwork for settlement products built around machine-driven transactions. If allocators start pricing compute access as a financeable asset, custody and fund infrastructure providers stand to benefit before stablecoin issuers do.

BlackRock has not detailed specific products tied to the paper’s thesis, leaving open whether the firm will pursue tokenized compute claims or agentic payment rails through existing vehicles. The next signal to watch is whether other large asset managers echo the argument, or whether AI agent transaction volume on public blockchains actually materializes in the months ahead.

BlockWest is a news publication. Nothing here is investment advice. Read our disclaimer and editorial policy.