Agent Operating Systems for Investing

This points to brokerages becoming pipes, while the real value shifts to the software layer that turns an investment idea into a live, monitored position. In that setup, the user writes a thesis in plain language, the agent pulls data, builds a model, watches for conditions like price moves or earnings signals, and then sends orders through the brokerage account when the rules are met.

• The workflow Jon Ma describes is already showing up in products. Robinhood says third party AI agents can connect through its Trading MCP to read account data and place orders in a dedicated agentic account. Public says its Agents monitor conditions and automate investing plans inside its brokerage.
  1 robinhood 2 public 3 public
• The key split is between idea formation and order routing. Artemis is aiming at the top layer, where an investor compares tokens, stocks, private companies, and prediction markets in one place, forms conviction, and sets targets. The broker underneath handles custody, balances, order types, and execution.
  4 sacra 5 sacra
• Different products are choosing different levels of autonomy. SoFi says Composer helps investors turn a view into backtested, rules based automation, while Public emphasizes approved plans that run when preset conditions hit. That suggests the first winning agents may look more like smart execution software than fully autonomous portfolio managers.
  2 public 3 public 6 sofi 7 sofi

The next buildout is likely to center on agent operating systems for investing, not just better broker apps. The winning product will own the thesis, the monitoring loop, and the handoff into execution across stocks, crypto, and other assets, while brokerages compete to be the safest and easiest underlying venue for those agents to plug into.