Alibaba Cloud leads Shengshu's ~RMB 2B Series B

Shengshu Technology announced on April 10 that it had closed a Series B of nearly RMB 2 billion. Alibaba Cloud led the round. CNIC, Jiu'an Haitang, TAL Education and Guanghe Ventures joined as strategic investors, and existing backers including Xinglink Capital, Daitai Capital, CDFE Emerging Investment, Baidu Ventures and Zhuoyuan Asia also invested again.

The Beijing-based company presents its work as a single "Foundation World Model" with two branches. The World Generation Model covers digital content, mainly the Vidu video series. The World Action Model handles physical control.

Shengshu says Vidu Q3 ranked first in Artificial Analysis's video benchmark after launch. It also says Q3 can generate up to 16 seconds of synchronized audio and video, with multiple shots, camera control, music and sound effects, and dialogue in several languages. Vidu is sold to developers through an API platform and to users through subscription products. It is also available on Alibaba Cloud's Bailian platform.

On the robotics side, Shengshu open-sourced Motus in December 2025. It describes Motus as a world action model built on a large video-generation model using its UniDiffuser framework. The company claims Motus has about 40% higher multi-task success rates than Physical Intelligence's Pi0.5. It also claims Motus was the first model to show scaling laws for embodied foundation models. The release doesn't give details on how either result was measured.

"The core of world models is to equip AI with the ability to form unified representations and predictions of the real world," founder Zhu Jun said.

All performance claims come from the company's own press release. Shengshu didn't give a valuation or say how it will divide the new money between Vidu and Motus. It said only that it is hiring more staff worldwide.