Sixty days after launch, roughly 100,000 retail investors have opened funded Agentic Trading accounts on Robinhood — accounts where an AI agent can buy and sell equities, options, and crypto on their behalf without a human clicking “confirm.” For anyone tracking the trajectory of the AI financial advisor space, the number is a useful real-world signal: consumers are willing to hand execution authority to AI faster than most incumbents expected.
What the Numbers Show
Robinhood posted record Q2 2026 results on July 29: total revenue of $1.308 billion, up 32% year-over-year, with net income of $561 million (+45%). Gold subscribers — the paid tier that unlocks the company’s AI tools — hit 4.8 million, up 39% from a year ago. Net deposits reached $21.7 billion for the quarter. According to Robinhood Investor Relations, Agentic Trading accounts now hold more than $100 million in assets under custody, less than two months after the feature went live.
What This Means for AI-Driven Advice
The adoption pace matters because agentic trading represents a qualitative step beyond robo-advisors. Classic robo-advisors rebalance a pre-set portfolio; Robinhood’s AI agents execute discretionary orders — options strategies included — based on natural-language instructions or external AI models connected via the Model Context Protocol. Early users are running complex multi-leg options trades, a workflow that would typically require a licensed broker or a sophisticated self-directed investor. The 100K figure suggests the friction of “letting an AI trade for me” is lower than the industry assumed, which has direct implications for how quickly AI-native advisory products can scale beyond budgeting and planning into full execution.
What Comes Next
Robinhood’s CEO Vladimir Tenev flagged on the earnings call that the current barrier is technical: deploying an AI agent still requires writing code or using a compatible third-party AI tool. The company is working to lower that bar so non-technical retail investors can configure and launch agents without help. When that happens, the 100K milestone may look like a pre-launch warm-up.