Robinhood announced on July 10 that eligible U.S. customers will soon be able to connect third-party AI agents to dedicated accounts for autonomous crypto trading, according to crypto.news. The expansion follows the company’s May 2026 beta launch of agentic trading for equities and options, which attracted 70,000 agentic accounts in roughly 60 days.
How It Works
Users connect AI agents to a dedicated Robinhood account and define trading strategies with predefined guardrails. The agents execute trades within those limits, with real-time profit-and-loss tracking and push notifications. A Robinhood executive said the system lets customers “build trading strategies with predefined guardrails instead of watching their accounts continuously,” per crypto.news.
Robinhood did not announce a launch date for the crypto version. U.K. customers will receive access after the U.S. rollout.
70,000 Accounts in 60 Days
The 70,000 agentic account figure is the most concrete adoption metric yet from a major retail brokerage. Robinhood disclosed the number during a Friday presentation, per crypto.news. For context, Robinhood reported roughly 25.8 million funded accounts in its most recent earnings. 70,000 agentic accounts represents a fraction of the total user base, but the pace of sign-ups across a two-month beta suggests demand extends beyond early adopters.
The equity/options beta launched in late May. Crypto access was announced on July 10 via Robinhood’s official X account, which posted: “Eligible US customers will soon be able to connect their AI agent to a dedicated Robinhood account to trade crypto on their behalf.”
Broader Exchange Race
Robinhood is not alone. Kraken announced the same week that it is rebuilding its entire app around AI agents for market monitoring and portfolio guidance, according to Benzinga. Coinbase introduced an AI investment advisor in June. The competitive dynamic among retail crypto platforms has shifted from fee competition and token listings to agent infrastructure and autonomous execution capabilities.
The Guardrail Question
Robinhood’s approach requires users to set limits before agents trade. That is a different model from Kraken’s approval-gated system, where users confirm each trade individually. The distinction matters: guardrail-based agents can act overnight while users sleep, while approval-gated agents require active confirmation. Both assume users can define risk parameters accurately before markets move. Neither model has been tested through a significant drawdown with thousands of autonomous agents executing simultaneously.