Crypto exchange Gemini has rolled out Agentic Trading, a new system that allows users to connect AI models such as ChatGPT and Claude directly to their exchange accounts—effectively handing autonomous systems control over market monitoring, trade execution, and risk management.

The platform, which Gemini describes as the “first agentic trading tool available through a regulated US-based exchange,” integrates its full trading API with the MCP open standard developed by Anthropic. The setup enables AI agents to interact directly with exchange infrastructure and execute trading strategies based on user-defined parameters in real time.

Humans Set the Rules — AI Runs the Market

“We believe we’re at the beginning of a fundamental shift in how people interact with financial markets,” Gemini stated. “Agentic trading isn’t just a feature. It’s a new paradigm where AI handles execution, patterns, and discipline, while users focus on strategy and goals.”

In practice, the system shifts traders away from direct market participation. Instead of placing orders themselves, users define strategy constraints while AI agents handle execution decisions across live markets.

At the core of the rollout are Trading Skills, modular AI functions designed to replicate professional trading workflows. These include tools such as Find the Spread, which scans real-time bid-ask spreads across trading pairs, and Retrieve Candles, which pulls historical price data for pattern recognition and backtesting. Gemini said additional features—including portfolio analytics and advanced execution logic—are already in development.

The launch signals a deeper transformation in financial market structure, where users are increasingly positioned as supervisors rather than decision-makers.

This shift reflects the rapid rise of “agentic AI,” where autonomous systems are no longer limited to analysis or recommendations but are granted direct access to financial infrastructure and execution layers.

Across the crypto sector, competing frameworks are emerging. Coinbase’s x402 protocol, now under the Linux Foundation, enables AI agents to access wallets and digital payment systems, while Tempo’s Machine Payments Protocol focuses on machine-to-machine financial transactions. However, Gemini’s implementation is the first explicitly designed for regulated exchange-level trade execution.

The Beginning of Fully Delegated Trading

With Agentic Trading, Gemini is effectively compressing the role of the human trader into a supervisory layer—where investment logic is set in advance, but execution is continuously delegated to AI systems operating in real time.

The shift raises a structural question for markets: if AI is responsible for timing, execution, and risk responses, what role does the human trader ultimately retain beyond setting parameters?

While proponents argue this could improve discipline, reduce emotional trading, and enhance efficiency, it also introduces a new dependency—where trading performance is increasingly shaped by autonomous systems operating at machine speed, outside of direct human intervention.

The rollout comes as exchanges and fintech platforms race to build AI-native financial infrastructure, embedding machine agents directly into trading, payments, and liquidity systems.

Gemini’s model positions it at the center of this transition—where trading systems are no longer just automated tools, but autonomous actors executing strategies continuously across live markets.

In this emerging structure, the boundary between human intent and machine execution becomes increasingly blurred, marking a shift toward financial systems where AI does not just assist trading—but performs it.