Shengshu Technology

Shengshu Technology is focused on developing multimodal large models that integrate text, images, 3D, and video. The firm gained significant attention with the launch of its text-to-video model, Vidu, which is positioned as a competitor to OpenAI's Sora. Vidu is built on a proprietary visual transformation model architecture called Universal Vision Transformer (U-ViT), which combines Diffusion and Transformer models. The initial version of Vidu could generate 1080p video clips up to 16 seconds long from a single prompt. A subsequent version, Vidu 1.5, introduced the capability to generate videos from both images and text, along with a feature for ensuring consistency of subjects and styles across multiple clips. The Vidu platform is now publicly available worldwide, supporting both Chinese and English text prompts and offering both free and subscription-based premium services. The company's revenue model is primarily based on subscriptions, with a significant portion coming from corporate clients.

The company's primary product, Vidu, is designed for a range of applications, including design, game production, post-production, and creating anime-style videos. Shengshu Technology offers its services through both a SaaS model and a Vidu API for enterprises and developers, targeting sectors such as interactive entertainment, advertising, film, animation, and cultural tourism. For instance, the Vidu tool has been utilized in the production of what is reported to be China's first fully AI-generated movie and for the digital restoration of cultural relics. The company has secured substantial funding from a roster of high-profile investors. Initial angel funding rounds in 2023 raised nearly $14 million, with backers including Ant Group, Baidu Venture, and Z&Y Capital. Subsequent funding rounds have attracted investment from Qiming Venture Partners, Delta Capital, and the Beijing Artificial Intelligence Industry Investment Fund, among others.