CoinDesk
Franklin Templeton's head of digital assets and innovation, Sandy Kaul, told investors that the next major AI trade may sit in blockchain networks rather than chipmakers or cloud providers. Her argument rests on agentic AI: autonomous software agents that book travel, buy compute, retrieve data and pay other agents without human input. Kaul said those machine-to-machine transactions will frequently be worth fractions of a cent, and that traditional card and bank rails break down once fees exceed the payment itself. Public blockchains, with programmable transactions, cryptographic identity and near-instant settlement, are better suited to carry that load, she argued.
Why it matters
The investment angle is that AI agents will need native crypto to pay network fees. Rising transaction volumes would translate into more demand for those tokens, more revenue for validator incentives and a new layer of demand for decentralized applications. Circle CEO Jeremy Allaire has put forward a parallel vision in a recent paper, framing agentic AI and blockchain as a single technological shift rather than two separate ones. He argues AI is collapsing the cost of knowledge work while blockchain and programmable money do the same for payments, settlement and coordination. Software pricing, in his framing, shifts from monthly subscriptions to pay-per-task models where agents are both buyers and sellers.
Market impact
Early infrastructure signals are already visible. Robinhood launched AI-powered investing tools in May that let agents trade stocks and make purchases for users, with CEO Vlad Tenev saying AI agents will eventually rival human traders. OpenAI and Anthropic are racing to ship increasingly autonomous systems that navigate software and complete complex tasks independently. If agentic commerce scales, the second-order beneficiaries sit below the model layer: the chains that settle agent payments, the stablecoins they transact in, and the tokens that secure those networks. Institutional AI portfolios have so far concentrated on chips, hyperscalers and data centres. Kaul's read is that a serious bid into blockchain rails is the next leg.
Frequently asked questions
-
Who is making the case that blockchain is the next AI trade?
Sandy Kaul, Franklin Templeton's head of digital assets and innovation, argued in a recent post that blockchain networks and crypto assets could benefit as autonomous AI agents begin transacting with one another.
-
Why would AI agents need blockchain payment rails?
Many machine-to-machine transactions would be worth fractions of a cent, for API calls, seconds of compute or dataset access. Kaul argues traditional card and bank rails become uneconomic once fees exceed the payment itself, while programmable blockchains can settle those micropayments cheaply and instantly.
-
What does Circle CEO Jeremy Allaire add to the thesis?
Allaire has published a paper framing agentic AI and blockchain as a single technological shift rather than two separate ones. He argues AI is collapsing the cost of knowledge work while blockchain and programmable money do the same for payments, settlement and coordination.
-
What real-world agentic AI activity is already happening?
Robinhood launched AI-powered investing tools in May that let agents trade stocks and make purchases for users. CEO Vlad Tenev has said AI agents will eventually rival the capabilities of human traders, while OpenAI and Anthropic are racing to build more autonomous systems.
-
How would AI agent adoption affect crypto token demand?
Kaul argues that if agents pay network fees in native crypto, rising machine-to-machine transaction volumes would drive more demand for those tokens, more revenue for validator incentives and stronger economics for decentralized applications running on those chains.
CryptoSlate
CoinTelegraph
WatcherGuru
WuBlockchain
TheBlock