Most AI investors have their eyes fixed on chip stocks and cloud infrastructure. But Franklin Templeton’s Sandy Kaul thinks they’re looking at the wrong part of the map. Her argument: the AI impact on blockchain could define the next major investment theme — and the window to position ahead of it is already opening.
Summary
Key takeaways
- Franklin Templeton’s Sandy Kaul argues that blockchain networks and crypto assets could be the next big AI investment theme, beyond semiconductors and cloud.
- Agentic AI — which acts autonomously without human input — generates microtransactions that traditional payment systems cannot efficiently handle.
- Public blockchain networks offer programmable transactions, cryptographic identity, and near-instant settlement suited for machine-to-machine payments.
- Robinhood launched AI-powered investing tools in May that let agents trade stocks and make purchases for users.
- Circle CEO Jeremy Allaire frames AI and blockchain as a single converging technological shift, not two separate trends.
Why the AI Investment Story May Have a Second Chapter
The first wave of AI investment was straightforward: bet on the picks and shovels. Semiconductor companies, hyperscale cloud providers, and data center operators absorbed enormous institutional capital as generative AI scaled up. That logic worked. But Kaul, head of digital assets and innovation at Franklin Templeton, warns that investors anchored to that narrative may be missing what comes next.
Her thesis centers on a specific evolution in AI architecture. Generative AI responds to prompts by producing content. Agentic AI operates differently — it completes tasks autonomously with minimal human involvement. An AI agent could book travel, compare prices, purchase computing capacity, retrieve data, or manage entire software workflows on a user’s behalf, all without waiting for a human to press confirm.
That distinction matters enormously for the payments world.
Agentic AI and the Microtransaction Problem
When millions of AI agents start transacting at scale — paying for an API call, a fraction of a second of computing power, or access to a specific dataset — the transactions involved are often worth fractions of a cent. That’s where existing financial infrastructure breaks down. Traditional payment networks become economically irrational when processing fees exceed the value being transferred. Credit card rails, bank transfers, and legacy settlement systems were never designed for this kind of high-frequency, sub-cent commerce.
This is the structural gap Kaul believes blockchain was built to fill.
Blockchain as Infrastructure for AI Agent Payments
Public blockchain networks solve the microtransaction problem through a combination of properties that legacy systems lack: programmable transactions, cryptographic identity verification, and near-instant settlement — without requiring a bank or card network as intermediary. AI agents holding digital assets could pay one another directly over blockchain rails, settling value in real time at minimal cost.
The investment implication is direct. If autonomous agents adopt blockchain-based payments at scale, demand for the underlying networks rises — and so does demand for the native cryptocurrencies that pay network fees. Rising transaction volumes would generate more revenue for developer incentives, network security, and decentralized applications built on top of those networks. That’s a feedback loop investors focused purely on AI hardware aren’t positioned to capture.
The Shift From Theory to Practice
The thesis is already beginning to take shape. Robinhood launched AI-powered investing tools in May that let agents trade stocks and make purchases on behalf of users. CEO Vlad Tenev has stated that AI agents will eventually match the capabilities of human traders — a claim that reframes blockchain infrastructure from a speculative asset class into potential operating plumbing for autonomous finance.
OpenAI and Anthropic are already racing to build increasingly autonomous systems capable of navigating software and completing complex tasks independently.
Circle and the Unified Technology Shift
Kaul’s thesis finds a powerful echo in the thinking of Jeremy Allaire, CEO of Circle. In a recent paper, Allaire argued that the rise of agentic AI and blockchain is not two parallel trends but one converging technological force. His framing: AI is driving the cost of knowledge work toward zero, while blockchain and programmable digital money are simultaneously compressing the cost of payments, settlement, and economic coordination.
The implication is that AI-native companies of the future may not operate through traditional corporate structures at all. Allaire envisions businesses running on-chain, with tokens representing ownership and governance, while software pricing evolves from monthly subscriptions toward pay-per-task models — where AI agents are both the buyers and the sellers of digital services. That’s a fundamentally different economic architecture than what exists today.
What This Means for Investors Watching AI
The analytical weight of this argument is worth sitting with. Investors who built exposure to AI through chipmakers like Nvidia captured real value as model training and inference scaled. But the next phase of AI adoption — agents transacting autonomously across digital economies — requires a different layer of infrastructure. Blockchain networks, programmable money, and native cryptocurrencies are that layer.
As autonomous AI systems become more capable, the volume of machine-to-machine transactions will grow in ways that legacy payment systems structurally cannot absorb. That’s not a speculative scenario — it’s an engineering constraint.
For investors, the question Kaul is raising isn’t whether to believe in AI. It’s whether their AI exposure actually covers the full stack of what AI adoption requires — or whether it stops at the hardware and cloud layer while the payment and settlement infrastructure beneath it goes unnoticed.
Franklin Templeton itself is already acting on the broader blockchain thesis. Its Benji tokenization platform has distributed approximately $2.44 billion in assets across multiple public blockchains as of mid-July, according to RWA.xyz data — a sign that one of the world’s largest asset managers is treating public blockchain infrastructure as a serious operational environment, not a side experiment.
FAQ
What is agentic AI and how does it differ from generative AI?
Agentic AI completes tasks autonomously with little human input — for example, booking travel, managing software workflows, or purchasing computing power. Generative AI, by contrast, responds to prompts by producing content such as text, images, or code. The key difference is initiative: agentic AI acts, while generative AI responds.
Why are traditional payment systems inefficient for microtransactions between AI agents?
Many transactions between AI agents involve very small amounts — sometimes fractions of a cent — such as paying for an API call or a second of compute time. Traditional payment networks charge fees that can exceed the transaction value itself, making them economically impractical for this type of high-frequency, low-value commerce.
How can blockchain networks benefit machine-to-machine payments in the AI economy?
Public blockchain networks provide programmable transactions, cryptographic identity, and near-instant settlement without requiring banks or card networks as intermediaries. This allows AI agents to hold digital assets and pay one another directly over blockchain rails at minimal cost — a structure designed for exactly the kind of automated, high-volume commerce agentic AI generates.
What investment opportunities arise from the convergence of AI and blockchain?
As autonomous AI agents adopt blockchain-based payments, demand for underlying networks and their native cryptocurrencies could grow alongside AI adoption. Investors currently focused only on AI hardware and cloud providers may be overlooking blockchain infrastructure as an additional — and potentially significant — way to gain exposure to AI’s next stage of growth.
Article produced with the assistance of artificial intelligence and reviewed by the editorial team.
