Mantic builds AI systems for predicting events in the messy world of human affairs, including geopolitics, business, policy, technology, and culture. In these domains, a purely data-driven modelling approach is insufficient. Flexible reasoning and research are required, which is why human superforecasters outperform automated methods. Mantic’s goal is to change that, delivering automated predictions at unprecedented accuracy and scale.
The company has come out of stealth with a team of AI technical staff drawn from Google DeepMind, Citadel, Cambridge, Oxford, and other AI startups. Mantic has raised £3m in a pre-seed round led by Episode 1, with backing from DRW and AI researchers at Anthropic and Google DeepMind.
Mantic is pushing the frontier of AI forecasting accuracy. It won the top prize-money in the Q1 2025 Metaculus AI Benchmark Tournament and its latest system sets a new state of the art when backtested on the 348 questions from Q2 2025.
The company aims to solve judgmental forecasting, where forecasters must understand the state of the world, reason about how events will play out, and assign probabilities over possible outcomes. Human superforecasters are currently the best judgmental forecasters, but statistical models are insufficient where data is scarce and strategic or political context is critical. Prediction markets like Polymarket and Kalshi exist, but they too have limitations and have not been integrated deeply into decision-making.
Mantic believes automation will transform judgmental forecasting in the same way it did weather forecasting. AI forecasting systems can be backtested, allowing models to be trained and evaluated at high speed. This reduces evaluation times from months to milliseconds and allows repeated simulations of world events. Reinforcement learning combined with large datasets of forecasting questions enables systems with a much greater volume of forecasting experience than any human could accumulate.
Unlike human forecasters or prediction markets, AI forecasting can scale in speed, scope, and depth. Predictions can be delivered rapidly, refreshed frequently, and tailored to client needs. Mantic’s system has already demonstrated success in the Metaculus Cup and has been used to track forecasts on energy markets in China, geopolitical risks in the Middle East, and corporate leadership changes in the German DAX index.
Mantic aims to provide industrial scale prediction, offering a radar-like system that scans the world and highlights what matters. The ambition is to move beyond individual predictions to build a richer picture of possible futures, enabling decision-makers to better navigate uncertainty.
