Franklin Templeton’s Sandy Kaul is making a bold bet: agentic AI, not tokenization or DeFi, will be the use case that finally pushes blockchain into the mainstream. In remarks this week, the firm’s head of digital assets argued that autonomous AI agents — programs that can plan, execute, and adapt tasks without human intervention — require the trustless settlement and programmability that only crypto rails can provide.
Why agents need blockchains
Kaul’s argument hinges on a simple problem. As AI agents start handling real-world transactions — booking travel, managing supply chains, executing trades — they need a way to pay for services and verify outcomes without a human in the loop every time. Traditional payment rails are too slow and too permissioned. Crypto rails, by contrast, let agents hold wallets, sign transactions, and settle instantly.
“Agentic AI is the killer use case for blockchain,” Kaul said. She pointed to the growing number of experiments where AI agents use crypto wallets to pay for compute, storage, or API calls autonomously. The logic is straightforward: if an agent needs to buy cloud credits at 3 a.m., it can’t wait for a bank to clear a wire.
What Franklin Templeton is watching
Franklin Templeton has been in the crypto space longer than most traditional asset managers. It launched a tokenized money market fund in 2021 and has been building out its digital assets team. Kaul said the firm is now tracking how agentic AI frameworks — like those from startups and open-source projects — are integrating blockchain for settlement.
She didn’t name specific projects or predict a timeline. But the message was clear: the intersection of AI and crypto isn’t just about trading bots or meme coins. It’s about infrastructure that lets software act independently in the economy.
The catch
None of this is guaranteed. Agentic AI is still early, and most agents today run in sandboxed environments. Security is a major question — if an agent’s private key gets compromised, who’s liable? Regulators haven’t touched this yet. Kaul acknowledged the challenges but argued the direction is inevitable.
“We’re at the very beginning,” she said. “But the pieces are coming together.”
What’s next
Franklin Templeton plans to release more research on the topic later this year. For now, the industry is watching whether the agentic AI wave will actually drive on-chain activity — or remain a theoretical pitch. Kaul’s bet is that by 2027, the question won’t be whether agents use crypto, but how much of the economy runs through them.
