AI Bot Earns $2.2M on Polymarket in 60 Days by Exploiting Mispriced Prediction Markets
An account using the pseudonym ilovecircle reportedly earned about $2.2 million on Polymarket in roughly 60 days by running an AI-driven, probability-based trading system. The trader treated Polymarket like a quantitative venue: Claude (Anthropic) generated Python code to connect to the Polymarket API, handle authentication, pull price data and execute trades. The bot aggregated multiple data sources — news feeds, social sentiment, on-chain whale activity, legislative trackers and sports feeds — and compared real-world probabilities to market-implied odds. When the AI model’s probability diverged materially from market prices, the system executed trades; the operator reported approximately 74% accuracy across markets including sports, crypto events and political outcomes. The build included dashboards to monitor large accounts and automated debugging/iteration, allowing one person to run what previously required an engineering team. The case highlights how automation, speed and data-driven probability models can outperform intuition in prediction markets and lowers barriers to entry for quantitative trading strategies.
Neutral
The news is neutral for broad crypto market direction but significant for prediction markets and trading behavior. Positives: it demonstrates that AI-driven automation and probability models can extract consistent profits from mispriced prediction markets, encouraging more algorithmic participation, higher liquidity and narrower spreads on platforms like Polymarket. That may increase competition and efficiency within prediction markets and attract capital into derivatives and event-driven trading. Negatives/risks: widespread adoption of similar bots can concentrate power with technically skilled operators, accelerate front-running, and reduce edge for casual traders, potentially harming smaller participants and increasing volatility within individual markets. For major crypto assets (BTC, ETH) the direct price impact is limited — this story affects niche markets and trading infrastructure more than spot crypto demand. Short-term: expect increased attention to prediction-market prices, possible volatility in specific event markets as bots reprice odds quickly. Long-term: greater automation will professionalize prediction markets, compress inefficiencies and raise technical barriers to profitable manual trading. The net market-level impact remains neutral because the story drives structural change in trading methods rather than altering macro liquidity or fundamental crypto demand.