Mantic has raised $25 million to scale its AI-powered forecasting technology, with the financing led by Radical Ventures and participation from Balderton, Thinking Machines Lab, DRW, FT Ventures from the Financial Times, M12, Microsoft’s Venture Fund, Episode 1 Ventures and Charlie Songhurst.
Co-founder and CEO Toby Shevlane said he and fellow founder Ben started Mantic to solve forecasting, developing AI systems intended to predict future events and help organizations make better decisions.
The company is already targeting financial markets as an important commercial use case. Shevlane said some of the world’s leading hedge funds are using Mantic to support trading decisions.
Mantic is also pointing to its performance in competitive forecasting as evidence of the platform’s capabilities. According to Shevlane’s announcement, Mantic out-predicted all 676 human participants in a major forecasting competition during the summer of 2026. The accompanying announcement identifies the competition as the Metaculus Cup.
Forecasting systems attempt to estimate the probability of future outcomes by continuously evaluating data, new information and changing conditions. For institutional investors, that capability could potentially be applied to economic, geopolitical, corporate and market events where probabilities change rapidly.
Mantic plans to use the new funding to expand its technology and organization as it moves from demonstrating forecasting performance toward broader commercial adoption.
The company is actively hiring across AI research, engineering, product development, sales and operations. It is also onboarding new customers, with traders highlighted as a particular area of focus.
KEY QUOTES:
“This summer, for the first time, Mantic out-predicted all 676 humans in the preeminent forecasting tournament.”
“We want to scale this technology to help people make better decisions. Already, some of the world’s leading hedge funds are using Mantic for trading in financial markets.”
Toby Shevlane, Co-Founder and CEO of Mantic
