The brokerage says third-party AI agents can trade through separate accounts under user-set controls, as CEO Vlad Tenev signals a broader push into automated investing for eligible US customers.
Robinhood has introduced “Agentic Trading,” a feature that lets customers connect third-party AI agents to dedicated brokerage accounts using Model Context Protocol, or MCP. The company says those agents can assess market conditions, build portfolios, rebalance positions and execute stock trades on a user’s behalf, while operating inside a separate account that cannot access a user’s main holdings unless permission is granted. Users also receive real-time alerts and an instant shutdown option, with compatible models including Anthropic’s Claude and OpenAI’s ChatGPT. Robinhood also unveiled an “Agentic Credit Card” that allows AI agents to make purchases within user-defined limits. HOOD stock climbed as much as 10% in a single trading day in late May after the announcement. The rollout points to a broader attempt to bring institutional-style automation to retail users. CEO Vlad Tenev said in early July 2026 that AI agents would soon match human trading competencies and framed the shift as a way to democratize strategies previously reserved for institutional investors. He also said Robinhood plans to extend crypto trading support to AI agents for eligible US customers. The move comes as Coinbase is reportedly developing similar AI integration features, underscoring how agentic finance is emerging as a new competitive front. Regulatory oversight remains a key uncertainty, particularly as SEC (U.S. securities regulator) and FINRA (U.S. broker-dealer self-regulator) would face questions over liability, fiduciary responsibility and market manipulation if AI systems begin making autonomous decisions for retail investors. For crypto users, the planned expansion matters because digital-asset markets trade around the clock and are typically more volatile than equities, potentially boosting volume on Robinhood’s platform while adding risks that regulators have yet to fully address.