Former Meta Executives Launch Sooth Labs, AI Event Prediction Startup Raises $50M at $335M Valuation

Gate News message, April 23 — Sooth Labs, an AI research lab founded by former Meta employees, is raising approximately $50 million at a valuation of around $335 million, with Felicis Ventures leading the round. Yann LeCun and Jeff Dean, Google's chief scientist, are participating investors, while Meta CTO Andrew Bosworth serves as an advisor.

Based in Pittsburgh, Sooth Labs aims to train AI models to help enterprises forecast the probability of geopolitical and market events. Unlike traditional forecasting algorithms, Sooth trains models on large-scale cross-industry datasets and supports multimodal inputs including video, audio, and text. In product demonstrations, the system predicted a 16% probability of the WHO declaring another pandemic before 2028 and a 33% probability of Anthropic going public this year. Target customers span finance, defense, insurance, and real estate sectors.

The founding team comes from Meta AI and Carnegie Mellon University. CEO Yaser Sheikh is a CMU visiting professor and former Meta vice president. Co-founder Ruslan Salakhutdinov is a CMU professor and Geoffrey Hinton's student who previously served as Apple's first AI research director before conducting AI research at Meta.