Zhang Yiming returns to oversee ByteDance’s “World Model” project: integrates with Pico headset to instantly generate video footage, and personally directs the efforts.

On September 8, 2026, multiple media reports indicated that ByteDance was developing an AI model for real-time spatial video generation based on the Seedance model. Founder Zhang Yiming personally supervised the project, coordinating cross-departmental teams and computing resources. The launch was scheduled for October. This model aims to enable synchronized visual changes when users move, turn around, or speak, providing an immersive experience with Pico headsets. Its applications include live streaming, short videos, and games. ByteDance ranked world model development as its top AI priority in 2026, allocating eight-digit RMB in budget for related research, with the goal of launching the first world model by the end of the year.