CoinDesk
Fidelity Digital Assets warned that AI agents could swell on-chain activity while leaving the underlying tokens behind. Analyst Max Wadington argued in a Wednesday report that cheaper AI development risks flooding crypto with applications that don't generate lasting demand or durable token value. AI agents settled more than $73 million across roughly 176 million blockchain transactions in the year through April, per a Keyrock report, yet most of that economic upside may bypass public chains entirely.
Why it matters
The crypto-AI narrative has run on the assumption that autonomous agents would drive sustained demand for programmable financial infrastructure. Alchemy CEO Nikil Viswanathan framed it bluntly, saying crypto was built for AI agents, not humans. But as AI lowers the cost of building, Wadington argues, competitive advantage shifts away from the technology itself toward liquidity, distribution, security, trust and regulatory integration. Banks, fintechs and tech companies can offer lower costs, better performance, regulatory clarity and established distribution, potentially siphoning agent traffic before it ever touches a public chain. Coinbase, Stripe and Visa are already building competing machine-to-machine payment systems that could keep much of that flow inside incumbent rails.
Market impact
The token-capture problem is the core of the bear case. Micropayments, the canonical AI-agent use case, generate relatively low fees and can route to Layer 2 networks or settle off-chain entirely. Beneficiaries are likely to be stablecoin issuers and service providers rather than base-layer tokens: Coinbase's x402 already settles predominantly in Circle's USDC. On the security side, Anthropic's Mythos model has forced the industry to reckon with AI-accelerated vulnerability discovery across smart contracts, bridges, key management and oracle systems. Kraken parent Payward joined Anthropic's Glasswing project to get restricted access to Mythos 5 for defensive use. The risk, per Fidelity, isn't that AI fails to drive activity. It's that AI succeeds, but base-layer tokens capture only a sliver of the upside.
Frequently asked questions
-
What did Fidelity's report say about AI and crypto?
Fidelity Digital Assets analyst Max Wadington argued that AI agents could swell crypto activity while bypassing base-layer tokens, with value accruing to stablecoin issuers, L2s and incumbent payment rails.
-
How much have AI agents transacted on blockchains?
AI agents settled more than $73 million across roughly 176 million blockchain transactions in the year through April, according to a Keyrock report cited by Fidelity.
-
Why might AI agent activity bypass public blockchains?
Banks, fintechs and tech companies can offer lower costs, better performance, regulatory clarity and established distribution, potentially siphoning agent traffic before it hits public chains.
-
What is the token-capture problem?
It is the risk that more on-chain activity does not translate into higher token value, because micropayments generate low fees and route to Layer 2 networks or off-chain settlement. Stablecoin issuers like Circle may capture the upside instead.
-
How is AI changing crypto security risks?
Advanced AI systems like Anthropic's Mythos model can identify weaknesses in smart contracts, bridges, key management and oracle systems faster. Kraken parent Payward joined Anthropic's Glasswing project to find and fix vulnerabilities before attackers exploit them.
Crypto News
CryptoSlate
TheBlock
CoinTelegraph
WatcherGuru