Franklin Templeton’s Sandy Kaul ties the next AI trade to public blockchain payment rails
Circle CEO Jeremy Allaire echoes the thesis, arguing agentic AI and blockchain are converging into one economic system.
Franklin Templeton digital assets head Sandy Kaul is pitching public blockchains and crypto assets as the “next AI trade” as autonomous agents begin transacting with each other. Circle CEO Jeremy Allaire is pushing a parallel framework that treats agentic AI and programmable money as one converging economic stack.
Key Takeaways
- Franklin Templeton’s Sandy Kaul framed blockchain networks and crypto assets as a potential next-stage AI exposure as autonomous agents start transacting.
- The core mechanism is machine-to-machine micropayments that can be fractions of a cent, where legacy payment fees can exceed the transaction value.
- Kaul argued public blockchains fit the use case via programmable transactions, cryptographic identity, and near-instant settlement that lets agents pay each other directly.
- Circle CEO Jeremy Allaire described agentic AI and blockchain as converging into a single economic system, extending the thesis toward on-chain coordination and pay-per-task pricing.
Kaul’s ‘Next AI Trade’ Pitch: From Chips and Cloud to Crypto Rails
Sandy Kaul, Franklin Templeton’s head of digital assets and innovation, is explicitly trying to rotate the AI trade away from the crowded “picks-and-shovels” complex of semiconductors, hyperscale cloud, and data centers. Her claim is that as agentic AI becomes more common, the under-owned layer is the payment and settlement infrastructure that lets software transact with software.
The market context at publication time was risk assets leaning softer: bitcoin traded at $65,741.03 (down 1.58%) and ether at $1,929.96 (down 0.019%), with XRP at $1.14 (down 0.61%) and SOL at $77.78 (down 0.20%), according to CoinDesk market data. That price action matters mainly because Kaul’s pitch is not an immediate catalyst tied to a single token. It is a directional narrative about where future transaction demand could show up.
Why Agentic AI Breaks Traditional Payments at Fractions-of-a-Cent Scale
Kaul’s mechanism is straightforward: agentic AI is built to take actions with minimal human input, not just generate content. In her framing, agents will book travel, compare prices, buy computing power, retrieve data, and manage workflows.
That behavior implies a different payment profile. Many agent-to-agent transactions could be worth fractions of a cent, such as paying for an API call, a second of compute, or access to a dataset. In that regime, traditional payment networks become structurally mismatched because fees can cost more than the underlying transaction value.
This is why the setup reads as narrative-first. The “why” is coherent, but the packet provides no quantified timeline for when agentic commerce becomes large enough to matter for any chain’s fee market.
Public Blockchains as Machine-to-Machine Payment Infrastructure
Kaul’s argument for public blockchains is feature-based: programmable transactions, cryptographic identity, and near-instant settlement. The practical implication is that agents could hold digital assets and pay one another directly over blockchain rails rather than routing through banks or card networks.
For traders, the cleanest linkage she offers is fee demand. If agent payments scale, agents would need native cryptocurrencies to pay network fees. In that scenario, higher transaction volumes could translate into higher token demand and more network revenue that can fund developer incentives, network security, and decentralized applications. The conditional is doing real work here. The thesis does not name which networks win, and it does not provide on-chain indicators that would let the market price the adoption curve today.
Validation Checklist: Early Signals and the Data Traders Still Don’t Have
The near-term question is whether this stays a conference-room narrative or becomes measurable flow.
One early real-world datapoint cited is Robinhood’s launch of AI-powered investing tools in May that let agents trade stocks and make purchases for users, alongside CEO Vlad Tenev’s view that AI agents will eventually rival human traders. The packet also points to OpenAI and Anthropic racing to build more autonomous systems that can navigate software and complete complex tasks.
The missing pieces are the ones that would make this tradable with precision: any follow-up from Kaul or Franklin Templeton that names specific networks or token categories, and any measurable on-chain indicators tied to the AI-agent payment thesis. Allaire’s “recent paper” also needs publication details and concrete implementation pathways to move the convergence claim from conceptual to operational.
Marcus Hale Analysis: The AI-Agent Payment Narrative Is Tradable, but Still Unpriced-in by Metrics
I treat Kaul and Allaire’s framing as a plausible mechanism in search of a dataset. The threshold that matters is whether agentic AI produces sustained, observable fee demand on public chains, because that is the bridge from “AI story” to “token cash flow” via native-fee spend.
This looks more like a sentiment catalyst than a fundamental shift until someone names the rails and shows the volume. If agent-to-agent payments become a real line item in fee markets, the setup starts to look structural rather than narrative-driven, because the beneficiaries would be the networks that capture the transactions and the tokens that clear the fees.
