Gemini, the US-based cryptocurrency exchange founded by Tyler and Cameron Winklevoss, announced on April 27 that it has enabled AI models to directly manage user trading accounts through a feature called Agentic Trading. The system uses the Model Context Protocol (MCP), an open communication standard developed by Anthropic, to connect AI assistants including Claude and ChatGPT to Gemini's full trading API. The exchange describes it as the first agentic trading tool offered by a regulated US-based exchange.

The practical meaning: a user can authorize an AI model to query market data, analyze price patterns, and execute trades on their behalf without manual intervention at each step. Gemini has packaged this into prebuilt modules called Trading Skills. Two are available at launch. "Find the Spread" lets an AI query the bid-ask spread on any trading pair, and "Retrieve Candles" pulls historical price data for pattern recognition and backtesting. MCP functions as the connector layer, similar to a universal plug between the AI model and the exchange's trading API.

Gemini's "first regulated US exchange" framing is accurate but narrow. By the time of this announcement, at least two other exchanges had already shipped comparable products targeting international markets. Bybit, the world's second-largest exchange by volume with 80 million global users, launched its AI Trading Skills Hub on March 13, offering 253 API endpoints and zero-setup compatibility with ChatGPT, Claude, Google's Gemini model, Cursor, and Windsurf across six functional modules. VALR, a South African crypto exchange, followed on April 10 with an AI Service built on the open Agent Skills Standard, supporting agents including OpenClaw, Claude Code, and OpenAI's Codex.

The timing of Gemini's announcement also coincides with a period of significant strategic contraction for the company. In February 2026, Gemini announced it would exit the United Kingdom, the European Union, and Australia, completing those withdrawals by April 6. The company cut its workforce by 25 percent and, according to comments from the Winklevoss twins, is now focused on the US and Singapore. Gemini's African footprint covers only Egypt, Nigeria, and South Africa. India is not a supported jurisdiction. For most readers outside North America, the product's practical availability depends on whether their country falls within Gemini's narrowed footprint. Outside the US, Singapore, Egypt, Nigeria, and South Africa, Agentic Trading is largely a product announcement to observe from a distance rather than use.

The more actionable story for African users sits with VALR. The South Africa-based exchange holds a license from South Africa's Financial Sector Conduct Authority (FSCA) and also operates under European regulatory approvals, a relevant credential for users who transact with European counterparties or who are evaluating the exchange's international compliance standing. VALR has integrated its infrastructure with Onafriq, a payments network reaching nearly one billion mobile money wallets across 43 African markets. That includes M-Pesa in East Africa and MTN MoMo across West and Central Africa. An AI agent operating through VALR could, in principle, interact with trading infrastructure in economies where traditional banking penetration is limited. VALR serves 1.8 million registered users and 2,000 corporate clients. Developers on the continent can access a public repository of agent skills at github.com/valrdotcom/valr-agent-skills.

For South Asian users, Bybit's Dubai-based operations and broad regional presence make its AI Trading Skills Hub the more relevant entry point. India represents the world's largest crypto-curious market, and its regulatory landscape continues to evolve under the Prevention of Money Laundering Act (PMLA) framework. Bybit also serves a significant user base in Pakistan. The MCP standard underlying Gemini's system is open and platform-agnostic, meaning any exchange can implement it as an open, non-proprietary protocol. That openness is central to why the competition is moving so quickly.

The market context is significant. AI bots already account for an estimated 58 percent of all crypto trading volume, according to industry data from Finance Magnates. The crypto trading bot market was valued at $47.43 billion in 2025 and is projected to reach $54.07 billion in 2026, with a compound annual growth rate of 13.9 percent through 2035. North America holds the largest regional share at 41 percent, followed by Asia-Pacific at 37 percent.

The governance picture has not kept pace. A McKinsey survey cited in a recent MetaComp governance report found fewer than one in three organizations have adequate oversight frameworks for AI agents they have already deployed in financial contexts. The governance conversation has been developing for some time: Singapore's Infocomm Media Development Authority (IMDA) published the world's first cross-sector governance framework for agentic AI in January 2026, establishing a foundation that sector-specific efforts have since built on. In April 2026, Singapore's MetaComp addressed the financial sector directly with the StableX Know Your Agent (KYA) framework, the financial sector's first governance standard for AI agents. It is organized around four pillars: agent identity and registration, authority and permission controls, behavior monitoring, and ecosystem governance. MetaComp Co-President Tin Pei Ling identified a core structural problem in the current setup: "When an AI agent completes a transaction, its identity and permissions do not automatically expire," leaving open questions about continuity, accountability, and intervention.

Gemini's IPO status adds another layer to read into this announcement. The company went public in 2025 at a valuation of roughly $2.2 to $3 billion, and reports in early April indicated potential buyers were evaluating the exchange. The launch may serve a dual audience: active traders and potential investors or acquirers looking for evidence of a credible growth story. For readers tracking where agentic trading develops next, the signals worth watching include the pace of MCP adoption among regional exchanges, evolving regulatory frameworks in India and other emerging markets, and whether governance standards like KYA gain traction beyond Singapore.