London-based artificial intelligence startup Mantic has raised $25 million in seed funding to advance a forecasting technology that recently outperformed human experts. The investment follows the company's standout performance in the summer 2026 Metaculus Cup, an online tournament where it assigned probabilities to political, economic, and cultural developments more accurately than human competitors. Radical Ventures led the round, with participation from Microsoft's venture fund M12, Thinking Machines Lab, Balderton Capital, and other investors.
Investor Backing and Strategic Support
The financing was led by Radical Ventures at an undisclosed valuation and included backing from Microsoft's venture fund M12, Thinking Machines Lab, Balderton Capital, DRW, FT Ventures from the Financial Times, Episode 1 Ventures, and Charlie Songhurst. Radical Ventures partner Aaron Rosenberg has joined Mantic's board as part of the investment. The investor group reflects confidence in a new category of artificial intelligence aimed at improving decision making through more accurate forecasts.
Outperforming Human Forecasters
During the summer 2026 Metaculus Cup, an online forecasting tournament that concluded this month, Mantic outperformed every human contestant and all but one automated entrant, a bot called laertes. Its high score, along with those of other AI participants, marked the first time that technology dominated the competition. The result gave Mantic a timely validation as it prepares to scale its predictive capabilities for commercial clients.
How the Forecasting System Works
Mantic specializes frontier AI models from other laboratories to improve their ability to predict future events. The system is tested on historical data, graded on its performance, and then refined through iterative improvements. CEO Toby Shevlane, who co-founded the company with Ben Day in 2024, said the idea originated during his time as a research scientist at Alphabet's Google DeepMind, where he needed better tools for predicting global events relevant to artificial intelligence.
Independent Thinking in High-Stakes Forecasts
Shevlane said one advantage over human forecasters was avoiding herd mentality during the tournament. In one example, humans overwhelmingly bet that Shakira's song Dai Dai would not overtake Waka Waka on the Billboard Hot 100, a consensus that proved wrong while Mantic did not bet heavily on it. Mantic also assigned roughly a 40 percent chance to Abelardo De La Espriella winning Colombia's presidential election, above the 30 percent consensus forecast, and he ultimately won.
Commercial Demand from Hedge Funds and Governments
According to Aaron Rosenberg, a partner at Radical Ventures and a new Mantic board member, the startup has drawn interest from companies and government agencies worldwide, some of which have already integrated its AI. Hedge funds and trading firms are particularly eager to use Mantic's forecasts because they can translate accurate predictions into immediate financial gains. Mantic declined to name its customers, but it is actively onboarding new clients, especially traders, as part of its early commercial expansion.
The $25 million seed round positions Mantic to build on a performance that reached superhuman levels in a major forecasting tournament. The company is actively hiring across AI research, engineering, product development, sales, and operations. By giving decision makers a sharper view of the future, Mantic aims to turn advanced forecasting into a practical tool for finance, government, and other industries.