BlackRock says the convergence of artificial intelligence and digital assets could turn blockchains into financial infrastructure for autonomous AI agents, creating networks where machines can pay for data, services and other resources without relying on conventional human-centered payment systems.
The asset manager outlined the thesis in its 2026 Thematic Mid-Year Update, published August 24, arguing that AI and digital assets could reinforce each other as both technologies move deeper into everyday economic activity.
BlackRock’s framework separates their roles: AI provides what it calls “machine-native intelligence,” crypto can function as “machine-native money,” and blockchains provide programmable infrastructure connecting intelligence with economic activity.
The argument extends the investment case for blockchain beyond cryptocurrency trading. If autonomous software increasingly makes decisions and purchases services independently, BlackRock suggests blockchain infrastructure could provide a settlement layer capable of executing those transactions programmatically and nearly instantaneously.
AI Agents Could Pay Other Machines Directly
BlackRock illustrates the concept through an autonomous travel agent.
An AI system could access a user’s schedule and travel preferences, research flights and hotels and independently assemble an itinerary. To complete that task, however, the agent might need to purchase information from third-party providers, including airfare, room-rate or availability data.
BlackRock envisions those exchanges involving micropayments from a user-controlled wallet that settle over blockchain rails in near real time, producing a verifiable record of the transaction.
The distinction matters because today’s financial infrastructure was predominantly designed around people and businesses initiating transactions through banks, cards and payment processors.
Autonomous agents could generate a different transaction pattern: extremely small payments occurring at high frequency between software systems. Traditional card economics can become inefficient for micropayments because fixed processing costs can represent a substantial portion of very small transactions.
Stablecoins and blockchain wallets potentially provide an alternative by allowing software to hold programmable value and execute payments automatically according to predefined permissions.
The concept is already moving beyond research. Circle launched its Arc blockchain mainnet on September 16 with infrastructure explicitly designed for real-time value movement and “agentic economic activity.” BlackRock is one of Arc’s founding validators alongside Mastercard, Visa, DTCC, ICE and Standard Chartered.
Tokenization Expands the Machine Economy
BlackRock sees the AI-blockchain convergence extending beyond payments. Its report describes tokenization as a potential next evolution of financial-market infrastructure, moving traditional assets including funds, credit and securities onto programmable digital ledgers. Ethereum currently hosts the largest share of tokenized real-world assets, BlackRock noted.
Combining tokenized assets with autonomous agents could eventually allow software not only to spend money but also to interact programmatically with financial markets.
The concept nevertheless remains early. Agentic financial systems must still address identity, authorization, cybersecurity, consumer protection and the question of who is legally responsible when autonomous software makes an incorrect or unauthorized transaction.
BlackRock itself is treating AI security as a material operational issue. A September job posting for an AI and Security Architecture director specifically lists responsibility for agent-based workflows, LLM applications and tool-driven integrations.
The firm’s thesis therefore does not predict that blockchains will replace existing payment networks. Instead, it identifies a potential new source of demand for digital-asset infrastructure as AI becomes increasingly autonomous.
Crypto’s next major user, under that scenario, may not necessarily be a person actively opening a wallet or trading a token. It could be software — continuously buying information, accessing services and settling transactions with other machines through programmable blockchain networks.