Robinhood's Vlad Tenev launches AI trading agents, 24/7 stock trading, and perps at Hood Summit

Vlad Tenev used Robinhood's Hood Summit event in Houston to roll out 16 new products, positioning the company as the infrastructure layer for AI-powered retail trading. Three launches stand out.

24/7 stock trading makes Robinhood the first US broker to offer round-the-clock equities trading. Perpetual futures brings eight contracts to US retail investors for the first time. And Robinhood Agents is the most strategically significant of the three.

“We rolled out 16 products. Twenty-four-seven trading of stocks will be the first place in the US that allows that. Perps, the first true perps offering in the US with eight perpetual futures. And then Robinhood Agents — an evolution of our agentic financial superintelligence products... It has over 150,000 users who have actually connected external agents to it.”

Robinhood Agents

The product has three layers. The core agent brings within Robinhood what the company's MCP integration previously allowed externally — connecting AI agents like Claude or Codex to execute trades on a user's behalf. That MCP product launched a few months ago and already has 150,000 users who connected external agents. Robinhood Agents puts that capability inside the app, with trade approvals on by default and trading scoped to a dedicated account separate from retirement funds.

Agent Apps is an app store for proprietary data — unusual options flow, satellite imagery, congressional trading disclosures — that third-party providers plug into so Robinhood customers can feed that data into their strategies. OpenAI is one partner, offering GPT-6 Luna free through year-end. Agent Loops is the execution layer: autonomous, deterministic strategy runs while the user sleeps.

Tenev frames the pitch as giving retail traders the toolkit of a quant hedge fund, in the same way Cursor and Codex made programming accessible to non-engineers. Given that Robinhood has 28 million accounts, the distribution argument is straightforward even if the analogy is aspirational.

Prediction markets on earnings

Robinhood is also launching prediction market contracts on company-level EPS, revenue, and product delivery metrics — iPhone shipments, Tesla deliveries. The use case Tenev highlights is surgical hedging: if you're long a stock into earnings and want downside protection without selling, you trade the earnings contract rather than the underlying. That's a meaningfully different product from the event-driven prediction markets that have run hot recently.

Risk architecture

The agentic account is sandboxed by design. Money has to be moved in deliberately, trade approvals default to on, and the agent cannot touch retirement accounts without explicit action. Tenev floats the idea of a separate "financial controller" agent that could allocate across accounts and serve as a meta-risk layer, but says that's not the current product.

Market read

On broader conditions, Tenev is deliberately non-committal. He notes the Figma IPO felt like peak animal spirits, that crypto has cooled and then partially recovered, and that prediction markets have been "going gangbusters." His clearest conviction is narrower: among the various AI finance use cases being tried, agentic trading is the one with durable, retentive product-market fit right now. One-time tasks like moving money to a higher-yield account don't require the same infrastructure. That framing is both a product thesis and an implicit argument for why Robinhood is building where it is.