Agentic AI — systems that monitor environments, make decisions and act autonomously — is moving from back-office automation into live financial markets. According to a Fortune Business Insights report, the market was valued at $7.29 billion in 2025 and is projected to reach $139.19 billion by 2034, a compound annual growth rate of 40.5%. MarketsandMarkets further estimates agentic AI SaaS could grow at 46.8% CAGR through 2032, echoing how cloud reshaped IT budgets a decade ago.
Gemini Integrates LLMs Into Exchange Execution
Centralized exchange Gemini launched Agentic Trading by wiring large language models like Claude and ChatGPT into its order system via the Model Context Protocol (MCP). Users give plain-English instructions — for example, “rebalance into majors when volatility spikes” — and the AI agent calls Gemini APIs directly to place and manage orders within preset limits. The feature was reported by Decrypt.
Bitget Expands Agent Hub With Mulerun
Bitget followed a similar logic. Its Agent Hub, expanded through a partnership with automation platform Mulerun, lets users deploy and share AI-driven strategies that execute on Bitget’s own infrastructure. Bitcoin.com noted this blurs the line between copy-trading, bots and fully autonomous agents. Separately, derivatives venue Hyperliquid’s deep liquidity and composable perps are increasingly seen as raw material for such automated strategies, as a crypto.news report highlighted.
Neyro Stays Non-Custodial Inside DeFi
Neyro takes a different path: building non-custodial, on-chain agentic trading within the Aurum ecosystem. It combines AI-driven trading agents with on-chain execution so “users retain control over their assets as if they never left the wallet,” avoiding the centralized black box. Andrew Isaacs, recently appointed COO of Aurum Foundation — with nearly two decades at Galaxy Digital and Morgan Stanley, executing over $23 billion in transactions — said at a community event: “Our real strength is the trading intelligence behind the agents — combining scalable AI-driven execution with non-custodial infrastructure.” In a post cited by Cointelegraph, he argued that AI agents will drive the next stablecoin wave: “managing deposits and interacting directly with non-custodial wallets, turning bots into primary users of on-chain dollars.”
Where AI and Crypto Cycles Converge
The common thread: AI agents sit one layer above today’s trading front ends. Users set strategy parameters and risk limits; agents watch order books, funding rates and on-chain flows 24/7, executing only when conditions match. In Neyro’s design, execution happens via smart contracts on DEXes and derivatives protocols like Hyperliquid, preserving self-custody — a trade-off that avoids the centralization drift of early “AI trading” products. Visually, two curves are intersecting: traditional crypto cap growth flattens into cyclical patterns, while agentic AI’s 40.5% CAGR climbs from $7.29B toward $139.19B. AI-linked tokens and infrastructure are capturing an increasing share of risk capital. If the trend holds, agents will not just sit alongside crypto but live inside it — becoming the dominant marginal users of block space, stablecoins and derivatives liquidity. In that world, Gemini and Bitget compete on safely exposing order books to agents, while projects like Neyro test whether DeFi can absorb autonomous trading without sacrificing user sovereignty.