Franklin Templeton: Blockchain Backbone for Agentic AI, Crypto’s Killer Use Case

Par DCN Editorial Team
Publié le 22 July 2026

Analysis

Franklin Templeton Bets on Blockchain as the Backbone of Agentic AI: The True “Killer Use Case” for Crypto

One of the world’s largest asset managers has just taken a stand. Franklin Templeton, which manages nearly $1.8 trillion in assets, published a report on Tuesday, July 22, 2026, signed by Sandy Kaul, its head of digital assets and innovation, which states without hesitation that agentic AI is the “killer use case” for blockchain. This is no minor declaration: when an actor of this scale publicly validates the crypto investment thesis, the market listens. Especially since Kaul adds an even more direct corollary for investors: to capture the value of the AI revolution, you’ll need to own the cryptocurrencies themselves, not just shares of the tech giants.

1. The Franklin Templeton Thesis: Why Agentic AI Needs Blockchain

Let’s start by understanding the reasoning. Sandy Kaul distinguishes between “classic” AI (chatbots, content generation, analysis) and agentic AI: software capable of acting, paying, and deciding autonomously, without asking for permission at every step. Concretely, an AI agent can book a flight, purchase data, subscribe to a cloud service, or negotiate micro-financial transactions, all without human intervention.

The problem, Kaul explains, is that traditional payment infrastructures are not designed for this type of traffic. The Visa network, for instance, processes between 1,700 and 10,000 transactions per second (TPS). These numbers are comparable to those of high-performance blockchains: Solana achieves 6,284 TPS, Aptos 12,933 TPS, and BNB Chain 3,252 TPS. But the catch is that Visa only records the transaction within its TPS window — settlement takes 1 to 3 business days. A blockchain, on the other hand, records AND settles simultaneously. This is what Kaul calls the “speed gap”: a structural difference that becomes prohibitive when software agents perform thousands of micropayments per hour.

To illustrate: imagine an AI agent tasked with managing the SaaS subscriptions of a company with 5,000 employees. Each monthly subscription, each overage, each additional license triggers a payment. On a traditional banking system, these hundreds of transactions pile up, take 48 hours to settle, and require manual reconciliation. On a blockchain, settlement is instantaneous, auditability is native, and the cost per transaction is measured in fractions of a cent. The difference is not incremental — it is structural.

2. A Market of $3 to $5 Trillion at Stake

This is not a pipe dream. McKinsey & Company projects that agentic commerce — autonomous transactions carried out by AIs — will represent between $3 trillion and $5 trillion by 2030. That’s equivalent to the GDP of Germany or Japan. And every transaction, every micropayment, every contract will have to travel over a payment rail. The question is not whether this market will exist, but which infrastructure will capture it.

To give a sense of scale, the Visa network processed about $12 trillion in transactions in 2025, the vast majority from human transactions. If only 10% of the projected agentic commerce volume flows through blockchain, that would represent $300 to $500 billion in additional on-chain volume per year. That is enough volume to saturate several networks and generate considerable transaction fees for validators, stakers, and token holders.

Franklin Templeton answers unambiguously: blockchain is the best-suited infrastructure. Decentralized networks offer a unique combination that traditional systems cannot replicate: instant settlement, no intermediary, low transaction costs, and programmable interoperability. AI agents can literally “speak” the language of smart contracts — a technological leap that traditional banking APIs struggle to match.

3. Concrete Initiatives That Validate the Thesis

The Franklin Templeton report does not appear in a vacuum. Several major initiatives are already underway, and their convergence suggests the infrastructure is being built before our eyes:

• Coinbase has already launched tools allowing AI agents to trade and pay autonomously on its platform. Developers can deploy AI-managed wallets in a few clicks via Coinbase’s Developer Platform API, and agents can execute trades, stake tokens, or interact with DeFi protocols without human intervention.
• Google unveiled an agent payment protocol in 2025, backed by the Ethereum Foundation. The tech giant sees blockchain as the future machine-to-machine payment rail — a strong signal from a company that dominates global cloud infrastructure.
• The x402 Foundation, officially launched on July 14, 2026, already brings together 40 organizations, including Visa, Mastercard, and AWS. Their goal: to build open payment rails for AI, using HTTP status code 402 (originally reserved for web payments in 1991 but never implemented). The x402 protocol allows one piece of software to pay another directly over the internet, without a banking intermediary.
• The x402 protocol in practice: each AI agent purchases data, computing power, or API access using a blockchain’s native token. Transactions are cryptographically signed, settled in seconds, and publicly verifiable. No more invoices, SEPA transfers, or credit cards — the agent pays directly, in real time.

This is no longer speculation. The infrastructure is being built before our eyes, and the biggest names in finance and tech are actively participating. The question is no longer “if” but “how fast.”

4. The Direct Consequence for Investors: Buy the Tokens, Not Just the Stocks

This is perhaps the most important sentence in the Franklin Templeton report. Sandy Kaul writes word for word: “I believe it will become increasingly clear in the years ahead that, in order to capture the value of decentralized networks and companies, investors will need to buy the cryptocurrencies and altcoins that underpin these ecosystems.”

In plain English: owning shares of Nvidia, Microsoft, or Google is not enough to capture the value of the agentic AI revolution. These companies sell tools (chips, cloud, software). But if the settlement layer of AI exchanges rests on blockchains, then it will be the native tokens — Solana, Ethereum, Aptos, BNB — that capture a significant portion of the value created. This is a powerful argument that upends the traditional AI investment thesis.

This analysis aligns with what we were already writing in June 2026 in our analysis of the GENIUS Act: tokenization is not a marginal phenomenon; it is the next phase of the internet. Agentic AI could be the catalyst crypto needed to move from a speculative asset to an indispensable infrastructure of global commerce.

5. Analysis of the Best-Positioned Blockchains

If the Franklin Templeton thesis is correct, not all blockchains are equal. Let’s examine the best-positioned contenders to become the payment rail for agentic AI.

Solana (SOL) — The Favorite Among AI Developers

With 6,284 TPS and transaction fees below a cent, Solana is a natural candidate for agent micropayments. The AI development ecosystem on Solana is already active, with projects like Grass (decentralized data network), Render (decentralized GPU rendering), and hundreds of DeFi protocols that could serve as building blocks for agents. Solana also has the advantage of architectural simplicity: a single runtime, no sharding, sequential execution that makes agent development easier. The price of SOL hovers around $145 at the time of writing, up 8% over the week. Its market cap of $68 billion offers sufficient liquidity for institutions.

Ethereum (ETH) — Security and Institutional Maturity

With “only” 75 TPS on Layer 1, Ethereum is not competitive on raw speed. But the game is played on its Layer 2 ecosystem (Arbitrum, Optimism, Base). These L2s cumulatively handle thousands of TPS and already host mature DeFi protocols (Aave, Uniswap, MakerDAO) that AI agents can use. Ethereum’s advantage is institutional trust: it is the most audited, most decentralized blockchain, and the one Franklin Templeton has already invested in through its funds. ETH at $1,927 is consolidating, but the agentic AI narrative could be the catalyst that propels it toward $2,500. Its market cap of $232 billion makes it the safest choice for institutional investors.

Aptos (APT) — Maximum Speed

With 12,933 TPS at peak, Aptos is technically the fastest blockchain cited by Franklin Templeton. Built by former Meta engineers (Diem project), it uses the Move language, which offers interesting security guarantees for AI-managed smart contracts. Aptos’s main advantage lies in its architecture: Block-STM (parallel execution) allows it to process transactions in parallel without conflict, ideal for the massive volume of micro-transactions generated by agents. However, its ecosystem is still young and its market cap (around $3.5...