BlackRock, the world's largest asset manager, said the widespread adoption of artificial intelligence could represent an underappreciated source of demand for digital assets.

In a research paper titled “The Machine-Native Economy”, BlackRock said the rise of AI and machine-to-machine payments could boost demand for blockchains and other programmable payment infrastructure, including stablecoins and other on-chain assets. The firm also identified a potential opportunity for digital assets to support the compute market, in which claims on computing capacity could be tokenized, traded and used as collateral.

“Together, these developments position AI as a structural catalyst for digital asset adoption and digital assets as a potential facilitator of the AI economy,” the paper's authors — Will Su, Robert Mitchnick, Jay Jacobs and William Helm — wrote. “This relationship remains underappreciated and could expand the role of digital assets as core infrastructure for an increasingly autonomous digital economy.”

The crypto industry has long argued for a link between AI and digital assets, but BlackRock's research could bring that thesis to a broader audience of institutional investors. The firm's stance carries particular weight in traditional finance: through its iShares business, BlackRock already offers spot bitcoin and ether exchange-traded funds in the U.S., putting its arguments in front of allocators who may not otherwise encounter crypto-native research.

AI and machine-native payment rails

One of BlackRock's arguments is that the rise of agentic AI — AI systems built to pursue goals and complete tasks with limited human supervision — could increase demand for machine-native payment instruments. While existing payment rails can support some degree of automation, account setup, credentialing and authorization may still require human involvement. Merchant fees can also make low-value transactions uneconomic, while settlement and finality times can vary across providers.

BlackRock said stablecoins — tokens designed to track the value of fiat currencies such as the U.S. dollar — native cryptocurrencies and tokenized real-world assets are well suited to high-frequency, sub-cent, machine-to-machine transactions that take place around the clock.

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

Compute as a new market for crypto

The authors also saw an opportunity for digital assets in the growing market for compute — the processing power needed to train and run AI systems. With AI demand surging, AI companies could seek to lock in costs and secure providers to manage risk. Claims on that capacity could then be represented as tokens that can be transferred, pledged as collateral or traded.

“This could in turn broaden institutional investor participation and establish compute as a new opportunity for the broader digital asset ecosystem,” the authors said, adding that AI agents could use such markets to automatically purchase resources as needed. The paper presents tokenized compute as a potential avenue rather than an established market, leaving both its design and its adoption as open questions.

Industry voices echo BlackRock's thesis

BlackRock's thesis echoes arguments made by crypto executives. In July, Coinbase CEO Brian Armstrong pushed back against calls for crypto to pivot to AI, arguing that AI agents could stoke demand for crypto-based financial services. “AI being a megatrend takes nothing away from crypto,” Armstrong wrote, because AI agents will need programmable money rather than traditional banking rails. “If anything, it makes crypto more important,” he added.

Crypto companies are already building tools to support that activity. Coinbase's x402 protocol and Tempo's Machine Payments Protocol have both been designed to let AI agents automatically pay for online services. In May, Circle introduced agent wallets and payment tools for USDC, the dollar-backed stablecoin it issues, while OKX's Agent Payments Protocol is designed to support recurring payments and escrow arrangements in which funds are released after a task's completion. Whether those rails attract sustained machine-to-machine volume — and whether tokenized compute moves beyond a research proposal — are likely to serve as early gauges of the demand thesis BlackRock describes.