ByteDance to Launch Real-Time Spatial Video AI Model, Led by Zhang Yiming

ByteDance is preparing to launch a real-time spatial video generation AI model, with founder Zhang Yiming personally overseeing development. The model could be released as early as next month, according to Bloomberg. It builds on the company's existing Seedance video model and aims to support live streaming, short dramas, and games in interactive virtual worlds, though the timeline might shift.

This new model is designed to create virtual worlds that respond to voice or actions from users of ByteDance's Pico headsets. It generates video streams at 20 frames per second with a super-low delay of about 50 milliseconds, making it suitable for spatial computing and human-computer interaction.

Zhang sees this spatial video model as a key driver for the whole ecosystem, connecting ByteDance's AI tech, cloud computing, content platforms, and Pico hardware. By moving heavy tasks to the cloud, it could lower the cost of VR devices and shift industry competition from hardware specs to AI models, computing power, and content distribution.