Gemini has launched Agentic Trading, making it the first regulated U.S. crypto exchange to offer direct AI agent integration for autonomous trade execution. The feature, announced on April 27, lets users connect any MCP-compatible AI model to their Gemini account, define strategies in natural language, and let the agent handle execution, according to Decrypt.
How It Works
The system is built on the Model Context Protocol (MCP), the open standard originally developed by Anthropic that gives AI agents structured API access. Gemini integrated its full trading API with MCP, meaning anything available through the traditional API is now accessible to AI agents, according to the Gemini blog post.
Users connect Claude, ChatGPT, or any other MCP-compatible model through Gemini’s Developer Platform. From there, they can instruct their agent in plain language: sell BTC when Bitcoin hits $100K, or buy when the bid-ask spread drops below 0.01%. The agent monitors conditions and executes trades within user-defined parameters.
At launch, Gemini shipped three modular “Trading Skills” that agents call as building blocks: Get Market Data (real-time prices and order book depth), Find the Spread (bid-ask analysis for any trading pair), and Retrieve Candles (historical OHLCV data for pattern recognition and backtesting). The company said it plans to expand the Skills library to enable fully agentic trading across a broader set of capabilities.
Regulatory Positioning
The launch is significant because Gemini operates under U.S. regulatory oversight. Permissionless blockchain agents already exist across DeFi, but they operate without the compliance guardrails required of regulated exchanges. Gemini’s approach validates a specific model: bounded agent autonomy within a regulated custodial framework, where the exchange controls the execution environment and users set the constraints.
This follows a pattern across crypto infrastructure. Decrypt noted that Coinbase’s x402 protocol (now under the Linux Foundation) and the Stripe-backed Tempo network’s Machine Payments Protocol are building similar bridges between AI agents and financial infrastructure. Binance launched its Agentic Wallet on April 24, enabling AI agents to execute on-chain transactions across BNB Chain, Solana, Base, and Ethereum.
The MCP Bet
The broader signal: MCP is expanding well beyond its origins in coding tooling. Gemini’s integration marks one of the first deployments of MCP in regulated financial services, a domain where the protocol’s structured access model (defined capabilities, explicit permissions, auditable actions) maps cleanly onto compliance requirements.
For agent builders, the question is whether MCP becomes the default integration standard for financial platforms, or whether exchanges and brokerages develop proprietary agent APIs. Gemini’s decision to build on an open standard rather than a locked-down integration suggests the exchange is betting on ecosystem effects over platform lock-in.
Gemini shares (GEMI) rose approximately 0.25% on the day of the announcement, per Decrypt. The stock remains down more than 55% year-to-date.