Platforms bifurcate into trading and investing
Jon Ma, CEO of Artemis, on building 24/7 AI agents for trading & investing
This split means the broker interface is becoming the product, not just the account. As agents take over order entry, trading platforms win by maximizing speed, flow, and high frequency engagement, while long term platforms win by packaging research, portfolio construction, tax tools, and trusted custody around buy and hold behavior. The same execution rail can sit underneath both, but the surrounding workflow, acquisitions, and monetization start to diverge.
-
Robinhood is already leaning into the trading side. It opened Agentic Trading to third party AI agents through MCP, and its recent product mix includes social feeds and always on trading infrastructure, which fits a platform optimized for frequent decisions and execution volume rather than deep fundamental research.
-
Fidelity, Schwab, and Interactive Brokers point toward the long term side by surrounding the brokerage account with research libraries, screeners, portfolio tools, tax workflows, and advisor style planning. In that model, the valuable surface is the place where an investor or agent forms conviction, not the place where the order gets routed.
-
This is why M&A should diverge. A trading platform would buy charting, social, signal, and automation products, similar to how SoFi bought Composer to turn investment ideas into automated strategies. A long term platform is more likely to buy research, data, planning, and portfolio intelligence products.
Over time, brokers that only provide custody and execution look more interchangeable, while brokers that own either the trader workflow or the conviction workflow get stronger. The next wave of winners will bundle AI on top of accounts in two distinct ways, one built for constant action, and one built for long duration capital allocation.