ByteDance is reportedly preparing an AI model for real-time spatial video generation that could launch as soon as next month, expanding its work in immersive computing and virtual environments for robotics and autonomous systems.

Founder Zhang Yiming is reportedly overseeing the project.

Built on ByteDance’s Seedance model, it is designed to generate interactive 3D environments for live streams, short-form dramas, games, and Pico headsets.

The system would reportedly shift heavy processing to the cloud to reduce computing demands on headsets.

One source said it can deliver content with about 0.05 seconds of latency and run at 20 frames per second.

The effort comes as Google DeepMind has publicly signaled work on world models, an area linked to broader industry interest in robotics and immersive systems.

🔗 Source: Bloomberg

Recent ByteDance developments