BlackRock believes widespread use of artificial intelligence could create new demand for stablecoins, blockchains and tokenised assets. In a new report, the firm highlights payments between AI agents and trading in computing power.
Photo: Larry Fink is chief executive of BlackRock.
When the first bitcoin ETFs were approved in the US in January 2024, it was in many ways BlackRock, led by chief executive Larry Fink, that set the wave in motion. Fink predicted early on that all types of assets will eventually be tokenised, meaning represented as digital units on a blockchain. That process is now under way. The asset manager is now linking the same development to artificial intelligence.
In the report "The Machine-Native Economy: How digital assets connect intelligence, commerce, and compute", published on 22 September, BlackRock argues that broad adoption of AI is an underappreciated source of demand and new applications for digital assets. AI is described as machine-native intelligence, digital assets as machine-native money.
The report's first point is that large language models and blockchains rest on the same basic idea. A language model breaks text into small units, tokens, and converts them into numbers the machine can compute with. A blockchain does something similar with value: cash, fund units or securities are represented as standardised digital tokens that can be transferred and settled.
The functions differ, BlackRock notes, but both translate something from the real world into a format machines can use directly. According to the report, that makes it easier for AI agents – programs that can plan and carry out multistep tasks with little human intervention – to read blockchain data than to connect to many fragmented legacy systems.
The second point concerns payments. When AI agents act on their own, they need payment systems that are open around the clock and can handle a large number of very small transactions, according to BlackRock. Examples include paying per request to a service, for data, or for computing power by consumption.
Card networks and ACH, the US system for bank transfers, offer a great deal of automation, but BlackRock points to several obstacles: account setup and authorisation that may require a human, fees that make very small payments uneconomic, and settlement that can take a business day or longer.
The report singles out x402, an open payment protocol from Coinbase that uses the web's 402 "Payment Required" status code to let machines pay each other. Stablecoins such as Circle's USDC are an early primary use case. Other players are launching their own standards: Stripe and Tempo are behind the Machine Payments Protocol, Stripe and OpenAI behind the Agentic Commerce Protocol, while Google and Visa have their own protocols for documenting that an agent is acting on behalf of a user.
Among digital assets, BlackRock expects stablecoins to dominate everyday transactions. A stablecoin is a token designed to hold a fixed value against a currency, most often the US dollar, which gives predictable pricing and settlement.
According to figures in the report, the market value of stablecoins in circulation exceeded US$300 billion in September 2026. Adjusted transaction volume passed US$11 trillion in 2025, roughly in line with the annual volumes of Visa and Mastercard. That is still far below the US$93 trillion that went through ACH the same year, but volume grew 80 per cent a year from 2020 to 2025, compared with about 8.5 per cent for ACH.
The growth could also create demand for the blockchains stablecoins settle on, such as Ethereum or Circle's Arc, where USDC is used to pay transaction fees. How much value flows to the networks' own cryptocurrencies depends, according to BlackRock, on how each network has designed its fees.
BlackRock also points to a study from the Bitcoin Policy Institute in which AI models in simulations favoured stablecoins for everyday payments and bitcoin for preserving value over time. The report stresses that these are simulated responses, not observed agent behaviour.
The third point concerns computing power, the capacity needed to train and run AI models. BlackRock notes that investors have focused on the huge investments in AI infrastructure, with estimates of more than US$5 trillion from 2025 to 2030, but argues that operating costs deserve equal attention.
Analyst consensus estimates for the cloud businesses of Amazon, Microsoft and Google point to combined revenue of about US$1.1 trillion in 2030, annual growth of 29 per cent from 2025. BlackRock envisions standardised contracts on computing power, including exchange-traded futures, enabling better price discovery and hedging. Such claims on capacity could in turn be represented as tokens, transferred, pledged as collateral and settled on-chain.
In such a market, AI agents could compare price, performance and location across providers, buy the capacity they need and pay per job. BlackRock sees Stripe's August 2026 agreement to acquire OpenRouter, which distributes workloads across more than 400 AI models from over 80 providers, as an early signal that compute procurement, billing and programmable payments may converge.
BlackRock adds its own caveats. Payments between AI agents are still at an early stage, and the market for computing power has little liquidity. Before standardised compute contracts can scale, the report says, challenges must be solved, including differences in performance between chip generations and large regional differences in electricity prices.
The report is written by Will Su, head of digital assets research, Robert Mitchnick, head of digital assets, Jay Jacobs, US head of equity ETFs, and William Helm, head of US iShares product innovation. Their conclusion is that digital assets could become an increasingly important part of AI's economic infrastructure as adoption spreads and agents become more capable.
Sources: BlackRock
You have read what you can read for free. With an account you get unlimited access to all news and articles on kaupr.io — and become part of the network of investors, professionals and builders following onchain finance.
