Starting this week, Robinhood’s approximately 29 million customers will have access to trading agents powered by OpenAI and Anthropic. Customers can choose between different OpenAI and Anthropic models and instruct these agents to execute trades, conduct research, and build sophisticated investment strategies.
This launch coincides with Robinhood’s annual HOOD conference. Previously, the company introduced its MCP tool in May, enabling technical users to connect their own agents to its trading platform.
Although artificial intelligence is gradually permeating every corner of the financial industry, Robinhood is the first company to launch a non-technical trading assistant for a large customer base—a move that could transform how Americans invest and potentially trigger ripple effects across broader markets.
A demonstration seen by Fortune showed that Robinhood users are first asked to name their agent, then choose between OpenAI’s GPT-6 Luna, GPT-6 Sol, or Anthropic’s Opus 4.8.
After configuration, users can give instructions to the agent in plain English. The agent can perform a range of tasks, from simple trades to more complex operations, such as “Buy $200 worth of Ford stock,” or more advanced “Loops.” Robinhood explains, “You can set up a Loop to check the market every morning and execute trades when certain conditions are met, or run a strategy that operates overnight, seeking opportunities while you sleep.”
Robinhood’s new service also includes a set of safeguards, which the company says are designed to prevent agents from taking unintended actions. These include providing agents with dedicated trading accounts and allowing users to set limits on the amount per transaction. Users can also opt for a confirmation process that requires agents to obtain final approval before executing any trades.
Robinhood Agents will also provide users with free access to a range of data providers, such as Unusual Whales and the crypto-focused Token Terminal, for a limited time.
The new frontier of investing
Robinhood executives told Fortune that easy-to-use agents and a financial data repository will equip users with investment capabilities comparable to the tools used on Wall Street. If true, the arrival of Robinhood Agents would mark another milestone in the company’s stated mission to “democratize finance.”
Robinhood CEO Vlad Tenev said in a statement: “Without a market, ownership cannot function; without traders, a market cannot function. We are transforming Robinhood into the world’s best platform for active traders by providing tools that were previously available only to hedge funds, major banks, and quantitative firms.”
At the same time, the large-scale rollout of trading agents could alter investment patterns in unpredictable ways. This may include a significant increase in trading volume on exchanges or the emergence of new trading strategies.
You can also imagine less optimistic scenarios. For example, what if trading agents begin communicating with each other and collectively buy or sell a particular asset on a large scale? This could, in turn, increase market volatility and even trigger outright panic, especially if malicious actors are involved.
If issues arise with agent-based trading, it remains unclear where ultimate legal responsibility lies. Robinhood’s position is that a custodial agent is not equivalent to providing financial advice; any advice or actions taken by the agent are akin to a customer seeking guidance from the internet or a friend. However, as with many AI-related issues, the legal landscape surrounding agent-based trading is still evolving.
Another issue is how much customers will ultimately pay to run their own trading agents. During the launch phase, Robinhood plans to offer the basic GPT-6 Luna model free until the end of this year, while using OpenAI and Anthropic agents will be charged at standard token rates. Company executives say that for most trades, the cost of using these agents will be negligible, but it remains to be seen whether this holds true if compute costs rise unexpectedly or if a large number of customers adopt research-intensive trading strategies.
Another question is how many investors will actually use these agents for trading. Nevertheless, early signs indicate they are being used. Robinhood reports that since launching a more technical version of the tool this spring, over 150,000 customers have opened agent accounts. As of late September, various agents on the platform were executing nearly 30 million trades per day on Robinhood.
Currently, Robinhood is the only brokerage offering non-technical agents at scale, but other fintech and crypto companies—including eToro, Public, and Coinbase—have already enabled users to connect their own agents via MCP tools. A likely prediction is that these companies will soon introduce trading agents directly within their platforms, and over time, traditional brokerages like Schwab and Fidelity will follow suit.
In the near future, it’s easy to imagine an environment where agents execute billions of trades daily, and retail investors deploy sophisticated strategies in new corners of the market. How this will impact market performance and wealth accumulation remains to be seen.