China’s AI Video Generation Moves Toward Practical Use as ByteDance’s Seedance 2.0 Draws Attention

Chinese technology platforms including ByteDance and Xiaohongshu are moving AI video beyond technical demonstrations and into film, television and advertising production. Models capable of maintaining character consistency across shots and precisely controlling camera movement could streamline storyboarding, filming and editing. ByteDance’s Seed team released Seedance 2.0 on February 12, 2026. The model supports text, image, audio and video inputs, making it an important test of the commercial viability of generative video.

Volcano Engine said on March 5 that Seedance 2.0 would cost 28 yuan per million tokens with video input and 46 yuan without it. ByteDance paused the model’s global rollout in mid-March after Disney, Netflix and the Motion Picture Association alleged copyright infringement. It relaunched the service in Southeast Asia and Latin America on March 24, allowing users to generate videos of up to 15 seconds at 1080p. Uploads of real people’s faces are prohibited, and IP filters and C2PA labels have been added.