Community Trust ScoreVerified
Robinhood wants AI to trade crypto for you. The company said eligible U.S. customers can now connect artificial intelligence agents to their accounts to buy and sell cryptocurrencies on their behalf — a direct expansion of the agentic trading platform it first rolled out for equities back in May.
The equities version launched May 27 in beta. Crypto is the next layer. Users who want in will need to open a separate account specifically for agentic trading — it doesn’t fold into an existing one. From there, the AI agent handles trades autonomously, while the user gets real-time profit-and-loss tracking and notifications. The whole thing runs through Robinhood’s Model Context Protocol servers. And if a user gets cold feet, they can disconnect the agent at any time. No lock-in. That flexibility seems deliberate, probably to ease the obvious anxiety retail investors might have about handing over trade execution to a machine.
Seventy thousand.
That’s how many agentic accounts were opened in the first few weeks after the equities launch. Pretty fast uptake for a beta product. A Robinhood executive said the goal is to level the playing field — giving retail investors access to tools that institutions have basically monopolized for years. Whether AI agents actually close that gap is a different question, but the demand numbers at least suggest people want to try.
Coinbase Already In, Competition Heating Up
Robinhood isn’t alone here. In June, Coinbase rolled out its own agentic trading service, letting users automate trades and payments through AI. AI agents have also started showing up alongside tokenized stocks, which is a separate but related trend worth watching. The race is on, and it’s moving fast. Platforms that don’t move on this probably fall behind on user engagement, which is maybe why Robinhood pushed crypto into the mix so quickly after the equities beta.
The broader shift toward AI-driven trading isn’t just a product story — it’s a regulatory one too. And that part’s getting complicated.
Congress Presses the SEC on Market Risk
The House Financial Services Committee sent questions to the SEC about agentic trading. Specifically, Representatives Bill Foster and Brad Sherman raised concerns about market volatility and broker-dealer liability. Their worry isn’t abstract. If AI agents across multiple platforms are trained on similar datasets, there’s a real chance they start behaving in sync — buying or selling at the same moments, amplifying price swings rather than smoothing them out. That kind of coordinated behavior, even if unintentional, could create serious instability.
The liability question is thorny too. When an AI agent makes a bad trade — or a series of them that moves a market — who’s responsible? The broker-dealer? The developer who built the model? The user who clicked “connect”? Nobody’s answered that cleanly yet. The committee set a response deadline of July 31, so the SEC has until then to weigh in. What comes back from that process will probably shape how every platform in this space operates going forward.
Unclear yet whether the SEC will push for new rules, issue guidance, or simply punt the question back to existing frameworks. The industry’s waiting.
What Retail Crypto Traders Are Actually Getting
For the average Robinhood user, the pitch is simple: let the AI watch the market so you don’t have to. Crypto markets run 24/7, which is exhausting to track manually. An agent that can execute trades in the middle of the night, respond to price moves in real time, and send notifications when something happens — that’s genuinely useful for retail investors who can’t sit at a screen all day.
But “useful” and “safe” aren’t always the same thing. The concerns Foster and Sherman raised aren’t just regulatory box-checking. They’re pointing at a real structural risk. Retail investors plugging into AI agents en masse, all running similar logic, could make markets choppier rather than more efficient. That’s the scenario regulators seem most nervous about.
Robinhood’s decision to keep the agentic account separate from regular accounts is probably part of the answer to that. It creates a firewall of sorts, limits cross-contamination, and keeps users conscious that this is a distinct type of activity. The ability to disconnect at any time matters too — it means the user stays in control, at least in theory.
The 70,000 accounts figure from the equities beta already gives Robinhood data on how people actually behave with these tools. Whether crypto users behave differently — more impulsive, more reactive to volatility — is something the company will find out fast. Crypto isn’t equities. The swings are bigger, the hours are longer, and the regulatory floor is still being poured.
Coinbase’s June launch means there are already two major U.S. platforms running live agentic crypto services right now. The SEC response due July 31 could change the shape of both.
Frequently Asked Questions
How does Robinhood’s AI crypto trading actually work?
Users open a separate agentic trading account, connect an AI agent, and the agent executes cryptocurrency trades on their behalf through Robinhood’s Model Context Protocol servers, with real-time profit-and-loss tracking and the ability to disconnect at any time.
What concerns did Congress raise about AI trading platforms?
Representatives Bill Foster and Brad Sherman questioned the SEC about market volatility risks and broker-dealer liability, specifically whether AI agents trained on similar data could act in concert and amplify market swings, with a response deadline set for July 31.
