Coinbase Ventures Maps Four Layers for Autonomous Trading Agents
Future investment priorities include agent frameworks and terminals, proprietary market data and reasoning models, and infrastructure for agent compliance audits.

Coinbase Ventures has outlined four components for autonomous trading software and named investment priorities as AI agents take on market research, strategy development, position management and risk control.
The design groups the software into intelligence, agent frameworks, execution, and compliance and observability. Coinbase Ventures placed x402, a pay-per-use data protocol, in the intelligence layer and named Senpi and Supertake as examples of trading-agent frameworks.
The company cited improved multi-step reasoning in frontier models and wider adoption of MCP and A2A connectivity standards as developments behind agentic trading. It also pointed to open-source projects OpenClaw and Nous Hermes, along with closer connections between tokenized stocks and perpetual contracts.
The firm identified agent frameworks and terminals, proprietary market data and reasoning models, and infrastructure for agent compliance audits as future investment priorities.
Coinbase CEO Brian Armstrong previously urged developers exploring agentic finance to reach out to Coinbase Ventures.