Polystrat Agent Executes 4,200 Trades With 376% Peak Return
Since its launch in February 2026, the Polystrat AI agent developed by Valory AG has demonstrated significant traction on the Polymarket prediction platform. Operating on the Olas protocol, the agent executed over 4,200 trades within its first month, achieving peak returns of 376% on a single trade. This performance starkly contrasts with that of human traders, of whom only 7% to 13% typically achieve profitability. In comparison, over 37% of Polystrat AI agents have recorded a positive profit and loss, more than double the rate of their human counterparts.
The agent is designed to provide retail users with a 24/7, non-emotional trading tool that adheres to disciplined, data-driven strategies. By removing human error and fatigue from the equation, these autonomous systems are proving their ability to consistently identify and act on market opportunities around the clock.
AI Bots Now Control Over 30% of Polymarket Wallets
The rapid success of agents like Polystrat is part of a much larger trend. According to analytics platform LayerHub, more than 30% of wallets on Polymarket are already using AI agents to execute trades. This indicates a fundamental shift in market structure, where human traders are increasingly competing directly with automated systems. This development is occurring as the prediction market sector continues to expand, having exceeded $44 billion in total notional trading volume in 2025 across major platforms.
Polystrat AI agents already outperform human participants in Polymarket, with over 37% of them showing a positive P&L versus less than half that number for human participants.
— David Minarsch, CEO and Co-founder of Valory AG
The advantage of these machine-driven strategies lies in their ability to process vast datasets and execute trades without emotional bias. Projects like Walbi and exchange integrations from Bybit are also introducing no-code AI tools, making automated trading more accessible to retail users and accelerating the move away from manual execution.
Olas Aims to Build a User-Owned 'Agent Economy'
Beyond simple trading performance, the long-term vision for protocols like Olas is to create a user-owned "agent economy." This model envisions a decentralized ecosystem where autonomous AI agents, owned and controlled by individuals, perform a wide range of value-generating tasks across Web3. The primary goal is to empower users with sophisticated automation, preventing a future where AI-driven economic activity is controlled exclusively by large, centralized platforms.
For investors, the success of these early applications validates the AI-crypto narrative by providing concrete performance metrics. As these agents expand their capabilities beyond simple trades to analyze niche, "long-tail" prediction markets, they could become powerful tools for aggregating dispersed knowledge. This positions AI-centric protocols as critical infrastructure for the next wave of data analysis and on-chain automation.
