All News

Odyssey Launches AI Model for Real-Time Interactive Video Worlds

Odyssey, founded by self-driving pioneers, has created an AI model that streams interactive video frames every 40 milliseconds, allowing users to explore video content like a 3D game. This new world model predicts future states to generate coherent, spatially consistent video streams, promising to revolutionize entertainment, education, and more by making video content interactive and dynamic.

Published May 28, 2025 at 03:08 PM EDT in Artificial Intelligence (AI)

Imagine watching a video where you don’t just passively observe but actively explore the scene, moving around as if inside a 3D-rendered world. This is the groundbreaking promise of Odyssey, a startup founded by self-driving technology veterans Oliver Cameron and Jeff Hawke. Their new AI model streams video frames every 40 milliseconds, enabling viewers to interact with video content in real time without relying on traditional game engines.

Odyssey’s model is powered by a novel world model that predicts the next state of the environment based on current conditions and user actions. This allows it to generate pixels that feel realistic, maintain spatial consistency, and output coherent video streams lasting five minutes or more. The result is an interactive video experience that blurs the line between traditional video and immersive 3D environments.

While the current demo is rough around the edges—environments can appear blurry or shift unexpectedly—the technology is rapidly improving. Odyssey streams video at up to 30 frames per second using clusters of Nvidia H100 GPUs, with a cost of $1 to $2 per user-hour. The company is actively researching richer world representations to enhance temporal stability and expand the range of user interactions beyond simple motion.

What sets Odyssey apart is its approach to data collection. Instead of relying solely on publicly available datasets, the startup has developed a 360-degree backpack-mounted camera system to capture real-world landscapes. This unique data source aims to create higher-quality, more realistic models that better represent the complexity of real environments.

Odyssey’s vision extends beyond entertainment. They foresee interactive video evolving across ads, education, training, travel, and more—transforming every form of video content into an on-demand, explorable experience. Importantly, Odyssey pledges collaboration with creative professionals rather than replacement, offering tools to integrate AI-generated scenes into popular creative software like Unreal Engine, Blender, and Adobe After Effects.

The startup has attracted $27 million in funding from investors including EQT Ventures, GV, and Air Street Capital, with Pixar co-founder Ed Catmull serving on its board. As the technology matures, Odyssey could redefine how we consume and create video content, making it a dynamic, interactive medium rather than a static one.

However, this innovation arrives amid industry concerns about AI’s impact on creative jobs. While some studios use AI to cut costs amid layoffs, Odyssey emphasizes partnership with creatives to unlock new storytelling possibilities rather than replace human talent.

In summary, Odyssey’s AI-driven interactive video technology is a bold step toward immersive media where viewers become participants. By blending real-world data capture with advanced world modeling, it promises to open new frontiers in entertainment, education, and beyond—ushering in a future where video is not just watched but explored.

Keep Reading

View All
The Future of Business is AI

AI Tools Built for Agencies That Move Fast.

QuarkyByte’s AI insights can help developers and creators harness Odyssey’s interactive video technology to build immersive applications and next-gen media experiences. Explore how our solutions optimize AI model deployment and integration for real-time interactive content that transforms user engagement and storytelling.