AI crypto convergence is becoming a clearer investment theme, according to 21Shares, which argues that September marked a shift in how markets viewed the relationship between artificial intelligence and digital assets.
In its October 6 market outlook, the crypto investment firm said investors were beginning to recognize the technologies as complementary. Its reasoning centres on autonomous software: AI agents need ways to hold funds, pay for services and access computing resources.
“September was the first month the market priced convergence rather than competition,” 21Shares wrote. That statement represents the firm’s market interpretation, rather than a universally established milestone.
AI crypto convergence starts with a payment problem
An AI assistant can recommend a service. An agent designed to complete a task may also need to purchase it.
Consider software instructed to compare shipping prices, buy a dataset or run a calculation using rented computing capacity. Completing that job requires more than a capable model. It requires a payment method, a spending budget and permission to transact.
21Shares explored this connection in a September 2 research paper, covering applications including payments, automated trading, storage and decentralized computing. The paper presents crypto infrastructure as a potential financial layer for increasingly autonomous software.
A wallet can provide an agent with access to funds under rules set by its operator. That does not make the software an independent legal person, and it does not remove the operator’s responsibilities. It changes how an authorized payment can be executed.
x402 shows how software can pay for a service
One practical example is x402, an open payment standard that connects payments to ordinary web requests.
A client requests access to a paid resource. The server responds with HTTP status 402 — Payment Required. The client can then submit a payment and retry the request to obtain access.
The official website describes applications including API payments, paid content and agentic commerce. Stablecoins are its primary payment use case, although the standard is designed to support multiple networks and payment approaches.
That makes the connection tangible: software can purchase a digital service as part of completing a task, rather than stopping each time to ask a person to arrange a subscription.
Official payment figures provide context
The x402 website displayed the following rolling figures when checked on October 7, 2026:
Metric | Displayed last-30-days figure |
Transactions | 75.41 million |
Payment volume | $24.24 million |
Buyers | 94,060 |
Sellers | 22,000 |
These are platform-displayed figures for a rolling period, not September calendar-month totals. Buyer counts should not be interpreted as unique people or independently verified AI agents.
The distinction matters. Payment activity demonstrates use of the infrastructure; it does not, by itself, measure the size of an autonomous AI economy.
The investment question is who captures the value
21Shares’ argument puts greater attention on infrastructure that software can use: settlement networks, payment tools and computing services. Its October outlook describes AI agents’ need for spendable funds and accessible computing as a reason the two sectors could develop together.
For investors, however, useful infrastructure and a successful token investment are different outcomes.
A service may attract customers while charging very low fees. Revenue may accrue to a company rather than a token. Activity may depend on incentives that disappear later. These possibilities make the business model as important as the adoption story.
Questions worth tracking include whether customers return, whether they pay without subsidies, and whether revenue reaches the asset investors actually hold.
Bitnxt view: follow paid usage beyond the AI label
Bitnxt’s editorial view: The strongest part of this story is the practical payment requirement. If software increasingly buys services on users’ behalf, developers will need reliable ways to fund, authorize and audit those purchases.
The next test is durability. Repeat customers, clearly measured payments and sustainable revenue would provide stronger evidence than transaction counts alone.
Readers can explore the broader argument in Bitnxt’s related coverage of Franklin Templeton’s view of agentic AI as a potential crypto use case.
