Odyssey Raises $310M at $1.45B Valuation to Transform AI Model Simulation

Odyssey, an artificial intelligence laboratory dedicated to developing advanced 'world models,' has successfully closed a substantial Series B funding round, raising $310 million. This significant investment now values the company at $1.45 billion. The funding initiative was spearheaded by Natural Capital, and saw contributions from a diverse group of investors including Amazon.com Inc., AMD Ventures, GV, EQT, and IQT, alongside continued support from its existing investor base such as Elad Gil, Google LLC's chief scientist Jeff Dean, and Y Combinator president Garry Tan. The fresh capital is earmarked to propel Odyssey's mission to create AI systems capable of understanding and simulating the complexities of both the physical and digital realms. Unlike conventional large language models that primarily focus on pattern recognition or predicting subsequent words, Odyssey's world models aim to learn how the world operates. This approach is intended to push the boundaries of AI, moving towards more general and comprehensive intelligence. In conjunction with the funding announcement, Odyssey also revealed a strategic collaboration with Amazon Web Services (AWS). Under this agreement, AWS will become Odyssey's preferred cloud provider, granting the AI lab access to AWS's specialized Tranium custom silicon. Tranium is specifically engineered by AWS to facilitate the training and deployment of AI models at scale, providing Odyssey with the robust computational infrastructure necessary for its demanding world model workloads. According to Oliver Cameron, co-founder and CEO of Odyssey, world models represent a new class of foundational AI, capable of simulating the world itself. He emphasized the need for industry partners to advance the frontier of general world models, aiming for a 'GPT-3 moment' in this field. Ron Diamant, Vice President and Distinguished Engineer at Amazon, highlighted that world models are among the most compute-intensive AI workloads, and AWS's Trainium is purpose-built to meet such scale requirements. This funding and partnership underscore the growing industry focus on developing more sophisticated and comprehensive AI capabilities.