Fidelity Digital Assets is out with new research arguing that AI agents are about to stop being financial assistants and start being financial actors. They'll execute trades, arrange loans, manage portfolios, and handle payments — all with limited human oversight, the firm says. Trading, lending, and asset management are likely to feel it first. Payments, despite huge transaction volumes, run on thinner margins and may not deliver the same economic value to infrastructure providers.
But the research comes with a warning: more AI-driven trading doesn't mean every financial platform wins. As AI agents get better at shopping across platforms for lower costs and better execution, real competitive advantage will likely sit with platforms that have deep liquidity, reliable data, clearing capabilities, and user networks.
Fidelity also points out that as AI drags down the cost of financial software development, code itself may become commoditized. That could make network effects — liquidity, user base, trust, regulatory compliance — more important, not less. And when AI agents start trading at scale, they'll create a new wave of machine-to-machine financial activity: buying compute, data, and digital services. That puts new pressure on identity verification, asset permissions, payment settlement, and auditable execution.
Industry insiders argue financial institutions will need to treat AI agents not just as productivity tools but as actual customers. Infrastructure that can offer secure identity, real-time data, deep liquidity, and verifiable execution could become the key value-capture layer of the AI finance era.
Source: thestreet.com