According to a report published by Fidelity Digital Assets on August 19, the firm has identified centralized systems operated by big tech and fintech firms as the primary risk to the crypto AI thesis. The report warns that AI-driven digital activity might largely occur within centralized systems, potentially limiting demand for public blockchains. This represents a significant shift from the traditional expectation that autonomous AI agents would increasingly use blockchains for payments and settlements. As reported by Fidelity, the central question remains whether networks and applications can capture meaningful economic value from AI activity, with technology companies, banks, payment networks and fintech platforms building infrastructure that allows agents to transact through controlled environments.
According to Fidelity's analysis, higher transaction counts may not produce proportionate returns for native blockchain tokens. As reported, agent payments could generate substantial volume while producing limited fee revenue for underlying networks. Recent activity on the XRP Ledger demonstrated this distinction, with AI agents completing 1.4 million payments for approximately $280 in network fees. In contrast, Fidelity found that trading produced 49 times more Ethereum base layer revenue per dollar of volume than payments during the previous 180 days. The firm expects that stablecoin issuers and payment service providers may capture more value than base blockchains due to low fees and strong competition. This aligns with Grayscale's identification of agentic finance, verifiable record-keeping, and decentralized AI as key demand areas where traditional systems are not built for AI generation.
As reported by Fidelity, AI tools can help developers write, test and deploy blockchain applications faster, with research involving more than 100,000 GitHub developers finding that coding agents increased commits by as much as 180% and production releases by 30%. However, the firm cautioned that more software does not automatically create useful products, as applications still require distribution, liquidity, regulatory compliance and sustained user demand. According to the report, cheaper development could make blockchain features easier to reproduce, with networks finding it harder to distinguish themselves through technology when competitors can quickly copy or modify similar tools. Fidelity suggested that durable advantages could shift toward liquidity, distribution, security and trust, while Grayscale's Pandl noted that traditional systems were not built for what AI will generate.
According to Fidelity's assessment, AI lowers the cost of building software while making it cheaper to identify vulnerabilities and conduct attacks, creating pressure that could turn security from a basic requirement into a central competitive advantage. The firm cited evidence supporting both sides of this assessment, noting that researchers found AI agents identified genuine vulnerabilities in Ethereum-related software, including a flaw later disclosed as CVE-2026-34219. Regulatory requirements create another barrier, with institutions favoring systems offering clear identity controls, permissioning and legal accountability. As reported, fully permissionless networks could face difficulty connecting autonomous agents with regulated financial services. Fidelity did not forecast these outcomes but framed each as risks that could reshape how much value public chains capture.
