Chinese startup Shengshu AI aims to rival OpenAI’s Sora – meet Vidu

Shengshu AI, a Chinese start-up, has launched its new text-to-video tool, Vidu, for global users. The tool supports both Chinese and English text prompts, making it accessible to a wider audience. Users can generate clips of four or eight seconds through the official website. This development places Shengshu among other firms offering similar services, such as Zhipu AI and Kuaishou Technology.

Vidu, first unveiled in April, can generate a four-second clip in just 30 seconds, making it one of the fastest tools available. The technology is based on Shengshu’s self-developed architecture, U-ViT, which was detailed in a research paper by a team led by Zhu Jun, the company’s chief scientist and a professor at Tsinghua University. Shengshu’s leadership team includes several Tsinghua alumni, highlighting the university’s significant role in China’s AI ambitions.

The tool also features a new character-to-video function, allowing users to animate real or fictional characters using simple text prompts. This capability lays the groundwork for potential commercial applications in the animation and content industries. Zhang Xudong, Shengshu’s product director, envisions future developments where users can animate multiple characters and scenes, integrating AI tools with traditional sectors.

Shengshu has attracted significant investment, securing tens of millions of US dollars from backers like Qiming Venture Partners, Baidu, Alibaba’s Ant Group, and the Beijing AI Industry Investment Fund. This financial support underscores the confidence in Shengshu’s potential to lead in AI video generation, positioning it as a strong competitor to OpenAI’s Sora.