Fidelity Digital Assets said six risks could prevent AI agents from creating lasting value for public blockchains or their native tokens. In a report published Aug. 19, Senior research analyst Max Wadington said closed systems operated by technology companies, banks, payment networks and fintech platforms may offer stronger performance, lower costs, broader distribution, established identity systems, credit and clearer compliance. The report also warned that payments could increase blockchain activity while directing more value to stablecoin issuers and payment providers. Fidelity cited 1.4 million AI-agent payments on the XRP Ledger that generated about $280 in network fees, and found that trading produced 49 times more Ethereum base-layer revenue per dollar of volume than payments over the previous 180 days. Other risks include limited economic value from faster software production, weaker technical differentiation, security vulnerabilities and regulatory constraints. Fidelity presented the risks as possible outcomes rather than forecasts, saying investors should focus on where agents deploy capital and which networks can combine liquidity, distribution, security and regulatory integration to capture durable demand. The analysis contrasts with Grayscale Head of Research Zach Pandl's view that AI adoption could create demand for Ethereum, Solana, Worldcoin and Bittensor through agentic finance, verifiable record-keeping and decentralized AI.