Robinhood Rolls Out AI Trading Agents to 29 Million Customers
Robinhood has made AI-powered trading agents available to all of its roughly 29 million customers, letting users direct trades and build strategies in plain English using models from OpenAI and Anthropic. The move marks the first mass-market release of agentic trading and raises questions about market impact, liability, and cost.
Robinhood has begun giving all of its roughly 29 million customers access to artificial-intelligence trading agents, making it the first major brokerage to put agentic trading in the hands of a mass retail audience. The rollout, announced during the company's annual HOOD summit, lets users pick between models from OpenAI and Anthropic and then instruct the agents to execute trades, conduct research, and assemble complex investment strategies.
In a demonstration, a Robinhood user was prompted to name their agent and choose among OpenAI's GPT-6 Luna or GPT-6 Sol, or Anthropic's Opus 4.8. Once configured, the agent accepted plain-English commands ranging from simple orders like buying $200 of Ford stock to more elaborate assignments the company calls Loops. According to Robinhood, a Loop can be set to check the market every morning and execute a trade when certain conditions are met, or to run a continuous overnight strategy that looks for opportunities while the user sleeps.
The company says the service includes guardrails intended to keep agents from behaving unexpectedly. Those include a dedicated trading account for the agent, user-set limits on how much the agent can trade at one time, and an optional confirmation process that requires the agent to seek final approval before executing a transaction. Robinhood is also offering free access to data providers such as Unusual Whales and the crypto-focused Token Terminal for a limited period.
The launch follows a May release of a so-called MCP tool that let technical users connect their own agents to Robinhood's platform. That earlier version has already drawn more than 150,000 customers who opened agentic accounts, and as of late September, various agents were transacting on the platform nearly 30 million times a day, according to the company.
Robinhood executives argue that easy-to-use agents paired with libraries of financial data will give ordinary investors tools comparable to those used on Wall Street. CEO Vlad Tenev framed the move as an extension of the company's stated mission to democratize finance. «Ownership doesn't work without markets, and markets don't work without traders,» Tenev said in a statement. «We're making Robinhood the best place in the world for active traders by delivering tools once reserved for hedge funds, big banks, and quant firms.»
The mass release of trading agents could shift investment patterns in ways that are hard to predict. A major uptick in active trading volume is one possibility, along with the emergence of new strategies. Less rosy scenarios are also conceivable, such as agents conferring with one another and moving en masse into or out of a given asset, which could inject greater volatility into markets or even trigger panic if malicious actors are involved.
If something goes wrong with agent-based trading, it remains unclear where legal liability would fall. Robinhood's position is that hosting agents does not amount to providing financial advice, and that any suggestions or actions are akin to a customer asking the internet or a friend. The legal landscape around agentic trading, however, is still evolving.
Cost is another open question. For the rollout period, Robinhood plans to offer the lower-end GPT-6 Luna free until the end of the year and to charge the standard token rate for the OpenAI and Anthropic agents. Executives say the cost of using the agents for most transactions will be negligible, but that could change if compute prices rise unexpectedly or if large numbers of customers pursue research-intensive strategies.
It also remains to be seen how many investors will actually use the agents for trading. Robinhood is for now the only brokerage offering non-technical agents at scale, though other fintech and crypto firms, including eToro, Public, and Coinbase, already let users connect agents through MCP tools. Those firms are widely expected to roll out trading agents directly within their own platforms, and conventional brokerages such as Schwab and Fidelity could follow. In the near future, it is easy to imagine agents placing billions of trades a day while ordinary investors deploy elaborate strategies in new corners of the market. How that affects market performance and wealth accumulation is still unknown.
6
