Algorithmic Trading, LLM-Powered Strategies & the Rise of Agentic AI Trading
PepperstoneIn this episode, Chris Weston (Pepperstone) and Steven Hatzakis (ForexBrokers.com) break down three stages of automation reshaping trading today: classic algorithmic trading, LLM-powered strategies, and the newest frontier — fully agentic trading systems. They cover how algo trading evolved from MetaTrader and TradingView indicator scripts into today's AI-assisted strategies, what actually separates "AI-coded" from "AI-powered," and what it takes to build a genuine agentic trading desk — from sensory and research agents through to a "chief investment officer" agent making the final call. They also get into the practical side: MCP servers, token budgets, and how a retail trader can realistically start experimenting without over-engineering it. Topics covered: ✅ Why most algorithmic strategies get over-optimized ✅ The real difference between automated, LLM-powered, and agentic trading ✅ How an agentic trading desk is structured (sensory, research, risk & reflection agents) ✅ Token costs and budgeting for agentic strategies ✅ MCP servers and tool access — the technical backbone of agentic trading ✅ How to start experimenting safely (demo accounts, test capital, defining your goals) Got a strong view on agentic AI or automated trading? Drop it in the comments — we'd love to hear from you. This content is for general information and educational purposes only and does not constitute financial or trading advice. Trading leveraged products carries a high level of risk. #AlgorithmicTrading #AgenticAI #AITrading #Pepperstone Chapter Markers 00:00 Today's three themes: algo, LLM & agentic trading 00:33 Algorithmic trading: where it all started 02:30 MetaTrader, TradingView & the 700-hour backtest 04:14 From automated strategies to true LLM-powered trading 05:56 Trading is about probabilities, not certainty 07:26 Inside an agentic trading desk: sensory, research & CIO agents 09:54 The catch: token costs & budgets 11:39 What traders need before going agentic: MCP servers & tools 12:22 Why agentic AI is "curious" (and unpredictable) 14:03 Getting started: goals, demo accounts & where to learn 16:10 Freeing up screen time (and adrenaline) 16:43 Why retail traders should still master backtesting first 17:47 "In five years, this is all we'll be talking about" 18:42 The technical barrier — and the shortcut through connectors 19:10 Voice-native agents: OpenAI's Atlas demo 20:04 Chris's idea: make your AI interview you first 20:39 Wrap-up & sign-off This content is not intended for a UK audience. Please visit our Pepperstone U.K. channel here: https://www.youtube.com/channel/UCk5HwBByE4VRj03__6iHgCA 🤝 Switch to Pepperstone. Sign up for a live trading account: https://bit.ly/Switch2Pepperstone 🐦Follow us on X/Twitter to keep up to date with market news, access expert commentary, analysis and more from our global market analysts: https://twitter.com/PepperstoneFX 👉 Subscribe to Chris Weston's Daily Fix newsletter: https://cloud.go.pepperstone.com/join-daily-fix The material provided here has not been prepared in accordance with legal requirements designed to promote the independence of investment research and as such is considered to be a marketing communication. Whilst it is not subject to any prohibition on dealing ahead of the dissemination of investment research we will not seek to take any advantage before providing it to our clients. Pepperstone doesn’t represent that the material provided here is accurate, current or complete, and therefore shouldn’t be relied upon as such. The information, whether from a third party or not, isn’t to be considered as a recommendation; or an offer to buy or sell; or the solicitation of an offer to buy or sell any security, financial product or instrument; or to participate in any particular trading strategy. It does not take into account readers’ financial situation or investment objectives. We advise any readers of this content to seek their own advice. Without the approval of Pepperstone, reproduction or redistribution of this information isn’t permitted. #fx #trading #CFDs