Key Takeaways
- Franklin Templeton estimates agentic commerce could reach $3-5 trillion by 2030, creating significant demand for automated machine-to-machine payments.
- Blockchains could outperform credit card networks for AI micropayments by offering programmable transactions, verifiable identities, and near-instant settlement.
- AI agents may increase demand for native tokens, such as SOL, by paying network fees with them.
Sandy Kaul, the asset manager’s head of Digital Assets and Innovation, argued that investors buying shares in chipmakers, data centers, and other AI-aligned companies may be overlooking how autonomous agents will conduct transactions.
Unlike generative AI tools that respond to human prompts, agentic systems can independently develop plans, interact with software, and execute multi-step tasks.
That development could produce trillions of dollars in machine-led commerce, and require payment infrastructure capable of processing tiny transactions without constant human approval.
Kaul believes blockchains are better suited to that role than legacy financial rails, potentially transforming crypto networks into the settlement layer for an economy where “software can pay software.”
Agentic Commerce Could Reach $5 Trillion
Institutional investors have already concentrated heavily on conventional AI exposure. The 10 largest S&P 500 companies, all aligned with the AI theme, now represent almost 40% of the index’s total market capitalization, according to Franklin Templeton.
However, agentic AI could shift value away from the companies building models and infrastructure toward the networks used by autonomous systems.
Estimates cited by Kaul place agentic commerce between $3 trillion and $5 trillion by 2030. By 2028, 33% of enterprise software could include agentic AI, while autonomous systems may handle as many as 15% of everyday business decisions.
Such agents could purchase computing power, make API calls, license data, and pay for digital services in real time. Each transaction may cost only a fraction of a cent, creating a machine-to-machine economy based on continuous micropayments.
Protocols are already emerging to support these transactions. Coinbase created x402, reviving the internet’s long-dormant “402 Payment Required” status code to let software request and complete payments automatically.
Coinbase later transferred the protocol’s intellectual property to the Linux Foundation.
Credit card networks and technology companies, including Stripe, Shopify, Google, and Amazon Web Services, have reportedly supported the broader standard.
AI agents could also reshape retail spending. Franklin Templeton cited projections suggesting they may influence between 15% and 25% of US e-commerce sales by 2030.
Why AI Agents May Need Blockchains
Traditional payment systems are poorly designed for transactions worth fractions of a cent.
Credit card payments typically carry fees of 2%-3% plus a fixed charge of around $0.30, making them impractical when an AI agent pays $0.001 for a second of computing power or a single data query.
Blockchains can process payments without the same minimum fee structure while supporting programmable transaction rules.
An agent could generate a single-use payment token specifying which merchant can accept it, the amount that can be spent, and when the authorization expires. A blockchain could verify those conditions and invalidate the token after the purchase.
Cryptographic identities could also help distinguish legitimate agents from unauthorized software. Each agent could receive credentials allowing it to sign transactions, while the network records an auditable history of its payments and decisions.
That transparency could prove particularly important when autonomous systems operate without real-time human supervision.
Companies would need to understand what an agent purchased, which rules it followed, and who authorized its activity.
Newer blockchains may also provide the throughput required for machine commerce. Kaul cited maximum transaction rates of 12,933 per second for Aptos, 6,284 for Solana, and 3,252 for BNB Chain.
Unlike card networks, which initially authorize payments before completing settlement later, blockchains can record and settle transfers in a single process. However, real-world performance, fees, and network reliability may differ substantially from advertised maximum throughput.
AI Payments Could Fuel Altcoin Demand
If AI agents use public blockchains, they will need to pay transaction fees in native tokens. An agent operating on Solana, for example, needs SOL to submit transactions.
Franklin Templeton expects this mechanism to create demand for the cryptocurrencies underlying the networks that capture agentic activity.
Higher transaction volumes could also produce more revenue for blockchain ecosystems. Foundations and decentralized organizations could use those funds to finance development grants, security audits, validator incentives and new applications.
Kaul described a potential flywheel: more AI activity increases token demand and network revenue, which attracts developers and produces additional applications, users and transactions.
Autonomous agents could simultaneously remove one of Web3’s largest barriers by managing wallets, token conversions, and payments in the background.
Consumers could use decentralized applications without having to buy crypto or understand blockchain infrastructure.
Still, greater network usage does not guarantee that every associated altcoin will appreciate. Token supply, fee-burning mechanisms, value distribution, and competition between blockchains will determine whether activity benefits holders.
Franklin Templeton nevertheless believes investors may eventually need crypto exposure to capture the value created by decentralized AI commerce.
The next stage of the AI investment cycle, Kaul argued, may not belong only to semiconductor companies and cloud providers. It could also reward the blockchain networks that allow autonomous agents to transact.