AI Pioneers Back Sooth Labs, a Startup Predicting Geopolitical and Market Events

Sooth Labs, a freshly minted AI startup founded by former Meta employees, just landed $50 million in Series A funding to build predictive models for businesses trying to forecast geopolitical and market events. The round values the company at roughly $335 million and is led by Felicis Ventures, with some serious firepower backing the bet: Yann LeCun (the AI legend and ex-Meta executive), Google Chief Scientist Jeff Dean, and Meta CTO Andrew Bosworth advising the operation.

The founding team reads like a who's who of AI talent. Chuck Hoover, Yaser Sheikh, and Ruslan Salakhutdinov—researchers with deep roots in computer vision, graphics, and machine learning—are building models designed to tackle one of the hardest prediction problems in tech: anticipating specific events that could move markets or reshape geopolitical landscapes. This is well beyond conventional forecasting; it's about training AI systems to recognize patterns and signals that humans and traditional models miss.

What makes this compelling is the combination of timing and talent. We're in an era where AI capabilities have expanded dramatically, but most commercial applications still focus on text, images, or narrow domains. Event prediction at scale—especially for volatile domains like geopolitics and finance—remains largely unsolved. The caliber of investors and advisors betting on Sooth suggests they see a genuine gap in the market. Whether the startup can actually deliver on this ambitious vision will depend on whether their models can beat human forecasters and survive real-world deployment in high-stakes environments.