Mantic AI Secures $25M, Redefines Future Forecasting
September 21, 2026, 3:48 pm
Mantic, a British AI startup, secured $25 million in seed funding. This investment follows a significant victory in the Metaculus Cup. Its advanced AI forecasting system bested all human competitors. The system predicts complex, real-world events. It avoids common human biases. This achievement signals a major shift in predictive analytics.
Mantic, a British AI startup, just secured $25 million in seed funding. This funding round follows a pivotal win. The company's AI forecasting system outperformed every human participant in the Summer 2026 Metaculus Cup. This marked a significant milestone. Machines showed dominance in complex predictions.
Radical Ventures led the investment. Microsoft’s M12 venture fund also joined. Thinking Machines Lab and Balderton Capital contributed. Aaron Rosenberg, a partner at Radical Ventures, now sits on Mantic’s board. The valuation remains undisclosed.
Mantic operates from London. It focuses on judgmental forecasting. This branch of artificial intelligence receives less public attention. Chatbots and image generators often capture headlines. Mantic tackles "messy questions." These include business trends, technological shifts, geopolitical outcomes, policy impacts, and cultural developments. Forecasts span weeks or months. Its systems thrive where context, judgment, and research are crucial. Historical data alone is not enough.
Mantic's founders are Toby Shevlane and Ben Day. They established the company in 2024. Shevlane previously worked as a research scientist at Google DeepMind. His experience there sparked the idea. He recognized a critical need for better forecasts. Global developments profoundly affect AI's trajectory.
The startup does not build new frontier foundation models. Instead, it adapts existing models. These models come from other leading AI labs. Mantic tailors them for specialized forecasting tasks. The system learns from historical events. It refines its predictions. It continuously improves its accuracy.
The Metaculus Cup offered a public proving ground. Participants assigned probabilities to diverse outcomes. These included political, economic, and cultural events. Mantic finished ahead of every human contestant. Only one other AI system, a bot named laertes, surpassed it. Reuters reported this was the first time AI systems broadly dominated the competition. This showcased a clear technological leap.
Mantic's system demonstrated a key advantage. It overcomes a long-standing human forecasting flaw: crowd following. Human forecasters often align with consensus. This can lead to significant errors. Mantic avoids this herd mentality.
Consider one example. Participants predicted whether Shakira’s “Dai Dai” would surpass “Waka Waka” on the Billboard Hot 100. Most human forecasters predicted it would not. This proved incorrect. Mantic showed less commitment to the consensus view. It avoided a major miss.
Another instance involved the Colombian presidential race. Mantic assigned Abelardo De La Espriella approximately a 40% probability of winning. The human consensus estimate hovered around 30%. De La Espriella ultimately won the election. Mantic’s deviation from the crowd proved accurate.
One tournament cannot definitively establish universal reliability. Forecasting performance varies. It depends on the question's nature. It depends on the time horizon. Available information also plays a role. Yet, Mantic's results are significant. They confirm AI systems are serious competitors. This field was once exclusively human.
The commercial appeal of such precise AI forecasting is immense. Aaron Rosenberg confirmed widespread interest. Companies and government agencies globally are inquiring. Some already utilize Mantic's technology. Customer names remain undisclosed. Hedge funds and trading firms show particular enthusiasm. A small forecasting advantage translates directly into substantial financial returns. Mantic’s “superhuman” capabilities offer immediate value.
AI has spent recent years mastering generation. It creates text, images, and software. Mantic is betting on the next valuable capability. It is something far harder to fake. It is being right about what happens next. This shifts the paradigm. AI moves beyond content creation. It enters the realm of predictive intelligence.
This $25 million investment fuels that ambition. Mantic plans to expand its predictive AI platform. The goal is to upgrade human understanding of the future. It provides decision-makers with superior insights. These insights span various complex domains. This technology offers a strategic edge. Organizations can anticipate market shifts. Governments can foresee geopolitical events. Businesses can manage risks more effectively.
The ability to accurately predict uncertain events holds immense power. It informs strategic planning. It enhances operational efficiency. It enables proactive decision-making. Mantic's success in the Metaculus Cup is a powerful validation. It proves the practical application of its sophisticated AI. This British startup is not just building a product. It is shaping the future of foresight. This marks a new frontier for artificial intelligence.
Mantic, a British AI startup, just secured $25 million in seed funding. This funding round follows a pivotal win. The company's AI forecasting system outperformed every human participant in the Summer 2026 Metaculus Cup. This marked a significant milestone. Machines showed dominance in complex predictions.
Radical Ventures led the investment. Microsoft’s M12 venture fund also joined. Thinking Machines Lab and Balderton Capital contributed. Aaron Rosenberg, a partner at Radical Ventures, now sits on Mantic’s board. The valuation remains undisclosed.
Mantic operates from London. It focuses on judgmental forecasting. This branch of artificial intelligence receives less public attention. Chatbots and image generators often capture headlines. Mantic tackles "messy questions." These include business trends, technological shifts, geopolitical outcomes, policy impacts, and cultural developments. Forecasts span weeks or months. Its systems thrive where context, judgment, and research are crucial. Historical data alone is not enough.
Mantic's founders are Toby Shevlane and Ben Day. They established the company in 2024. Shevlane previously worked as a research scientist at Google DeepMind. His experience there sparked the idea. He recognized a critical need for better forecasts. Global developments profoundly affect AI's trajectory.
The startup does not build new frontier foundation models. Instead, it adapts existing models. These models come from other leading AI labs. Mantic tailors them for specialized forecasting tasks. The system learns from historical events. It refines its predictions. It continuously improves its accuracy.
The Metaculus Cup offered a public proving ground. Participants assigned probabilities to diverse outcomes. These included political, economic, and cultural events. Mantic finished ahead of every human contestant. Only one other AI system, a bot named laertes, surpassed it. Reuters reported this was the first time AI systems broadly dominated the competition. This showcased a clear technological leap.
Mantic's system demonstrated a key advantage. It overcomes a long-standing human forecasting flaw: crowd following. Human forecasters often align with consensus. This can lead to significant errors. Mantic avoids this herd mentality.
Consider one example. Participants predicted whether Shakira’s “Dai Dai” would surpass “Waka Waka” on the Billboard Hot 100. Most human forecasters predicted it would not. This proved incorrect. Mantic showed less commitment to the consensus view. It avoided a major miss.
Another instance involved the Colombian presidential race. Mantic assigned Abelardo De La Espriella approximately a 40% probability of winning. The human consensus estimate hovered around 30%. De La Espriella ultimately won the election. Mantic’s deviation from the crowd proved accurate.
One tournament cannot definitively establish universal reliability. Forecasting performance varies. It depends on the question's nature. It depends on the time horizon. Available information also plays a role. Yet, Mantic's results are significant. They confirm AI systems are serious competitors. This field was once exclusively human.
The commercial appeal of such precise AI forecasting is immense. Aaron Rosenberg confirmed widespread interest. Companies and government agencies globally are inquiring. Some already utilize Mantic's technology. Customer names remain undisclosed. Hedge funds and trading firms show particular enthusiasm. A small forecasting advantage translates directly into substantial financial returns. Mantic’s “superhuman” capabilities offer immediate value.
AI has spent recent years mastering generation. It creates text, images, and software. Mantic is betting on the next valuable capability. It is something far harder to fake. It is being right about what happens next. This shifts the paradigm. AI moves beyond content creation. It enters the realm of predictive intelligence.
This $25 million investment fuels that ambition. Mantic plans to expand its predictive AI platform. The goal is to upgrade human understanding of the future. It provides decision-makers with superior insights. These insights span various complex domains. This technology offers a strategic edge. Organizations can anticipate market shifts. Governments can foresee geopolitical events. Businesses can manage risks more effectively.
The ability to accurately predict uncertain events holds immense power. It informs strategic planning. It enhances operational efficiency. It enables proactive decision-making. Mantic's success in the Metaculus Cup is a powerful validation. It proves the practical application of its sophisticated AI. This British startup is not just building a product. It is shaping the future of foresight. This marks a new frontier for artificial intelligence.