Odyssey Announces Investment from NVentures and Samsung Next

Today, we’re excited to announce an investment from NVentures—NVIDIA’s venture capital arm—and Samsung Next to accelerate our research towards a general-purpose world simulator

Oliver Cameron

February 12th, 2026

Since founding Odyssey two years ago, our singular focus has been to research and deploy frontier world models. Our latest model, Odyssey-2 Pro, is a breakthrough general-purpose world model, which developers can now integrate with all kinds of applications, from robotics to gaming to education to defense and much more.

Today, we’re excited to announce an investment from NVentures—NVIDIA’s venture capital arm—and Samsung Next to accelerate our research towards a general-purpose world simulator—where from any starting point, a model can generate infinitely-long, interactive simulations of anything you can imagine. This adds to our funding from other world-class investors, such as GV, EQT, Air Street Capital, Elad Gil, Jeff Dean, Guillermo Rauch, Garry Tan, and Kyle Vogt.

This investment reflects shared conviction that world models have reached a critical inflection point, similar to where language models stood around the release of GPT-2.

Odyssey-2 Pro, a nascent world simulator

“We were impressed with the rapid technical advances demonstrated by Odyssey-2 Pro, showing promising progress towards interactive world simulation, and the early signs of teaching artificial intelligence true cause-and-effect.”

—Andy Duong, Investment Director from Samsung Next

With this new funding and our world-class research team—recruited from DeepMind, OpenAI, ByteDance, Tesla, Waymo, Meta, Wayve, and Luma—we’ll accelerate our research on increasing simulation length, quality, and interactivity, and scale the systems necessary to train and serve next-generation world models.

We’re on a journey to achieving the GPT-3 of world models—and this investment will accelerate that.

World Model

Odyssey-2

Our most powerful general purpose world model yet, materially advancing the state-of-the-art in physical accuracy of world models

World Model

Starchild-1

A step beyond world models that learn only from visual observation, toward systems that learn from richer multimodal interaction with the world

World Model

Agora-1

A multi-agent world model, enabling multiple participants—human or AI—to share and interact within the same world simulation in real-time

Reinforcement Learning

PROWL

A novel RL-driven adversarial framework where an RL agent explores game environments with the objective to improve world model performance

World Model

Odyssey-2

Our most powerful general purpose world model yet, materially advancing the state-of-the-art in physical accuracy of world models

World Model

Starchild-1

A step beyond world models that learn only from visual observation, toward systems that learn from richer multimodal interaction with the world

World Model

Agora-1

A multi-agent world model, enabling multiple participants—human or AI—to share and interact within the same world simulation in real-time

Reinforcement Learning

PROWL

A novel RL-driven adversarial framework where an RL agent explores game environments with the objective to improve world model performance

World Model

Odyssey-2

Our most powerful general purpose world model yet, materially advancing the state-of-the-art in physical accuracy of world models

World Model

Starchild-1

A step beyond world models that learn only from visual observation, toward systems that learn from richer multimodal interaction with the world

World Model

Agora-1

A multi-agent world model, enabling multiple participants—human or AI—to share and interact within the same world simulation in real-time

Reinforcement Learning

PROWL

A novel RL-driven adversarial framework where an RL agent explores game environments with the objective to improve world model performance

World Model

Odyssey-2

Our most powerful general purpose world model yet, materially advancing the state-of-the-art in physical accuracy of world models

World Model

Starchild-1

A step beyond world models that learn only from visual observation, toward systems that learn from richer multimodal interaction with the world

World Model

Agora-1

A multi-agent world model, enabling multiple participants—human or AI—to share and interact within the same world simulation in real-time

Reinforcement Learning

PROWL

A novel RL-driven adversarial framework where an RL agent explores game environments with the objective to improve world model performance