Franklin Templeton: Agentic AI Will Run on Crypto, Not Wall Street Rails

Sandy Kaul, Head of Digital Assets and Innovation at Franklin Templeton, has a message for anyone who thinks buying Nvidia stock covers their AI exposure: It doesn't.

Kaul published a paper Tuesday arguing that agentic AI—software that acts, pays, and decides on your behalf without checking with you at every step—will run on crypto rails, not Wall Street ones. Franklin Templeton manages nearly $1.8 trillion in assets.

"Blockchain will be pivotal in allowing agentic AI to realize its potential for consumer transactions, and the growth of agentic AI is likely to become the 'killer' use case that drives blockchain adoption," the asset manager said in its post.

The case turns on what AI agents actually do, Franking Templeton argues. AI agents go beyond what a simple chatbot can do—agents can shop, book, and pay for things autonomously, once given permissions by the user. A report by the AI advisory firm Capgemini cited in the paper describes the shift as moving AI "from a reactive, conversational chatbot to an autonomous system that can perceive its environment, devise a plan, and execute multi-step tasks to achieve high-level goals without constant human supervision."

That’s what agentic AI does. Execute things autonomously instead of just spitting out information.

Per a Bain & Company forecast, also cited by Kaul, "AI agents are expected to account for 15% to 25% of all U.S. e-commerce sales by 2030."

Traditional payment networks—built for human-scale, human-speed commerce—can't handle the math required to power a world in which agentic AI becomes a social phenomenon. There would be too many transactions for banks to process on time.

Blockchain networks, being distributed ledgers that record and settle transactions simultaneously without a bank in the middle, can.

“While bitcoin only processes ~7 transactions per second and Ethereum ~75 transactions, newer high-speed chains are recording maximum speeds ranging from 12,933 transactions per second (TPS) on the Aptos chain, 6,284 TPS on Solana and 3,252 TPS on BNB Chain,” the post reads. “These transaction speeds are on par with the Visa network that processes 1,700 to 10,000 transactions per second in normal operations.”

“Yet, even this comparison is misleading. Blockchains both record and settle their transactions in that TPS window whereas the Visa network only records a transaction. Settlement on the Visa network takes 1-3 business days.”

That speed gap matters when your software is making thousands of micropayments per hour. Coinbase already launched tools that let AI agents trade and pay autonomously. Google unveiled a payment protocol for agents in 2025, backed by the Ethereum Foundation.

The x402 Foundation—40 organizations including Visa, Mastercard, and AWS building open payment rails for AI—launched formally on July 14. The protocol revives a forgotten HTTP status code from 1991, originally reserved for web payments that never materialized, and puts it to work letting software pay software directly over the internet.

If each agent buys data, compute, and API access using a blockchain's native token (the cryptocurrency that powers transactions on that specific network), demand for coins like Solana and Ethereum could follow volume.

Kaul's conclusion is direct: "I believe what will become increasingly clear in coming years is that in order to capture the value of decentralized networks and businesses, investors will need to buy the cryptocurrencies and alt coins being issued by those entities."