The BlackRock AI agents thesis puts a practical question at the center of the crypto discussion: how will software pay for the services it uses? An October 5 insight, published under Robert Mitchnick’s byline, explores that question through The Machine-Native Economy: AI and Digital Assets.
The accompanying paper lists Will Su, Robert Mitchnick, Jay Jacobs and William Helm as authors. It examines programmable payments and potential markets for claims on computing capacity.
For investors, the opportunity requires careful interpretation. A growing need for automated settlement does not establish which network will win—or how much value its token might capture.
BlackRock AI agents thesis connects software with economic activity
BlackRock argues that stablecoins could lead transactional use, while standardized compute claims could develop into another digital-asset market. It also acknowledges that agentic payment activity and compute-market liquidity remain limited.
That makes the thesis a framework for evaluating an emerging market. The commercial test is whether businesses and their software customers find these systems useful enough to adopt repeatedly.
An agent that can identify a suitable service still needs permission to spend, a way to pay and evidence that the service was delivered. Payment infrastructure becomes part of completing the task.
x402 shows how software payments can work
The x402 standard provides a concrete example. Its official website describes a payment flow built around HTTP requests: a client requests a resource, receives a 402 Payment Required response, pays and retries the request.
The standard is intended for uses such as paid APIs, digital content and agentic commerce. It is not tied to a single blockchain.
Consider an illustrative research service. An authorized agent requests a paid dataset, encounters a payment requirement and settles the charge before receiving access. The payment becomes part of the interaction between software systems.
That example shows the proposed commercial function. It does not mean every agent should have unrestricted wallet access or that a payment confirms the quality of the purchased information.
Part of the transaction | Question a business must resolve |
Authorization | What may the agent purchase, and within which budget? |
Payment | Which currency and settlement method does the seller accept? |
Delivery | Was the requested service provided successfully? |
Accountability | Who handles errors, disputes and unauthorized spending? |
These questions will influence adoption alongside speed and transaction costs.
Mitchnick has discussed the connection before
The October insight follows earlier comments from Mitchnick about AI and digital assets.
In BlackRock’s April 20, 2026 Aladdin podcast, he described crypto assets as:
“a much more natural instrument for an AI agent economy”
He also distinguished AI tokenization from financial tokenization. The processes are technically different, but both translate inputs into representations that machines can use.
In the same discussion, Coinbase executive Brian Foster explained why agents carrying out tasks may need wallets and access to funds. He described Coinbase’s work on tools that allow agents to use digital assets to complete those tasks.
The earlier discussion supplies context for the latest publication. It should not be mistaken for a new product announcement.
What does the paper say about XRP, XLM and HBAR?
The reviewed paper does not identify XRP, Stellar or HBAR as its selected beneficiaries.
The screenshot’s connection between those tokens and BlackRock’s thesis is the commentator’s investment interpretation. Mitchnick’s previous employment at Ripple does not convert that interpretation into BlackRock’s support for XRP.
Investors assessing any proposed beneficiary need additional evidence: actual integrations, paying customers, settlement activity and a clear mechanism connecting usage to token demand.
A network could process more payments while keeping fees extremely low. A service could also attract users who transact in stablecoins without holding substantial quantities of its native token. The relevant economics must be examined separately for each system.
Bitnxt view: watch the sellers, payments and repeat usage
Bitnxt sees an important shift in the conversation: AI systems are increasingly discussed as potential customers of digital services, with payments forming part of their workflow.
Our earlier analysis, Why AI Agents Could Become the Biggest Crypto Users, examines the broader adoption question. The distinction between transaction frequency and economic value remains useful: many tiny payments can create substantial activity without immediately creating substantial revenue.
The strongest evidence would be service providers accepting these payments, agents returning to purchase again and infrastructure earning sustainable income from that activity.
BlackRock’s publication adds institutional attention to the theme. Turning the thesis into an investment case still requires evidence of adoption and a defensible explanation of where the resulting value goes.
