ShengShu Technology
Category: Foundation Models / LLMs
ShengShu Technology is a Beijing-based multimodal generative AI company developing video generation models (Vidu) and embodied foundation models (Motus/MotuBrain), building toward a general world model that bridges digital and physical domains. ShengShu Technology was founded in 2023. The company is led by Luo Yihang (罗义行). Based in Beijing, China. Team size: 51-200. Total funding raised: $400.0M. Latest round: Series B. Key investors include Alibaba Cloud; Ant Group; Baidu Ventures; Qiming Venture Partners; Zhongguancun Science City; LINK-X Capital (星连资本); Delta Capital (达泰资本); TAL Education (好未来); Zhuoyuan Capital (卓源亚洲); C&D Emerging Investment (建发新兴投资).
AMW Analysis
ShengShu Technology is a Beijing-based multimodal generative AI company developing video generation models (Vidu) and embodied foundation models (Motus/MotuBrain), with the stated goal of building a general world model that bridges digital content generation and physical world simulation. The company recently closed a Series B round of approximately $275-293 million, led by Alibaba Cloud with participation from state-backed funds and strategic investors including Baidu Ventures, TAL Education, and others. This funding, the company’s largest to date, is intended to accelerate development of its general world model—a multimodal AI that perceives video, audio, and text to simulate human interaction. The news flow indicates ShengShu is positioning Vidu as a competitor to OpenAI’s Sora and expanding into robotics with Motus, while pursuing real-time, high-quality video AI that could reduce production costs. The round signals continued investment from major Chinese technology firms and state-linked capital in the company’s foundation model and world model ambitions.
AMW analysis, generated from 3 tracked news signals.
- Founded
- 2023
- Headquarters
- Beijing, China
- Team size
- 51-200
- Total funding
- $400.0M
Value proposition
Building the world's first general world model that bridges digital content generation (video/3D) with physical world simulation (robotics/autonomous driving), powered by proprietary U-ViT diffusion transformer architecture.
Products and solutions
Vidu — AI video generation platform (text-to-video, image-to-video, reference-to-video); Vidu MaaS Platform (API for developers/enterprises); Vidu Q1/Q3 Pro models (top-ranked on VBench); Motus — embodied foundation model for robotics; MotuBrain — robot general-purpose brain combining world model prediction with action output; Vidu4D — single video to 4D reconstruction
Unique value
First Chinese company to release a video generation model rivaling OpenAI's Sora (Vidu, April 2024); pioneered U-ViT architecture (Diffusion Transformer) before OpenAI/Google; unique dual-track strategy covering both digital content (Vidu) and embodied AI (Motus/MotuBrain) under a unified world model framework.
Target customer
Global individual creators, professional content producers, enterprises in interactive entertainment, advertising/marketing, film/animation, cultural tourism, and smart hardware; embodied AI/robotics companies for industrial, commercial, and home applications
Industries served
Interactive Entertainment; Advertising & Marketing; Film & Animation; Cultural Tourism & Creative Industries; Smart Hardware; Embodied AI / Robotics; Autonomous Driving
Technology advantage
Proprietary U-ViT (Vision Transformer backbone for diffusion models) architecture published at CVPR 2023 — predating OpenAI's Sora architecture; UniDiffuser (ICML 2023) — first unified transformer for multimodal diffusion; DPM-Solver (NeurIPS 2022) — fast ODE solver for diffusion sampling; strong academic lineage from Tsinghua AI Institute; first-mover advantage in China's AI video space
How they differentiate
Unique dual-track strategy: simultaneously pursuing digital content generation (Vidu) and embodied AI (Motus/MotuBrain) under a unified general world model architecture, unlike competitors focused solely on video generation. Strongest academic AI research team in China's generative AI space (Tsinghua AI Institute). Pioneered U-ViT architecture before OpenAI.
Main competitors
Kuaishou (Kling AI — generated $150M revenue in 2025); ByteDance (Jimeng AI video); PixVerse (backed by Alibaba, $60M); OpenAI (Sora); SenseTime
Key partnerships
Alibaba Cloud (strategic full-stack AI partnership + lead investor); Baidu AI Cloud (Vidu API on Qianfan platform); Strategic partnerships with embodied AI companies for industrial/commercial/home robotics; Honor (smartphone AI video partnership)
Notable customers
ByteDance; Samsung; Wondershare; TAL Education Group; Alipay; Honor; Served 200+ countries/regions via Vidu platform; thousands of enterprise/developer teams via Vidu MaaS
Major milestones
March 2023: Company founded; June 2023: Angel round led by Ant Group ($14M); April 2024: Launched Vidu — China's first Sora-class video generation model; CVPR 2023: U-ViT paper published (Diffusion Transformer backbone); ICML 2023: UniDiffuser published; March 2025: Luo Yihang (ex-ByteDance) appointed CEO; 2025: 10x growth in users and revenue; February 2026: Series A+ of ¥600M ($82M); April 2026: Series B of ¥2B ($293M) led by Alibaba Cloud; Vidu Q1/Q3 Pro models top-ranked on VBench; Launched Motus embodied foundation model
Growth metrics
10x growth in both users and revenue in 2025; Vidu platform serves 200+ countries/regions; thousands of enterprise/developer teams on Vidu MaaS
Market positioning
Leading Chinese AI video generation startup with strong academic pedigree from Tsinghua; positioned as a world model company bridging digital content and physical robotics — differentiating from pure video generation competitors; top-ranked on VBench benchmarks; first-mover advantage over Sora in China
Geographic focus
China (primary), global markets (200+ countries via Vidu platform)
Patents and IP
U-ViT backbone for diffusion models (CVPR 2023); UniDiffuser — one transformer for all multimodal diffusion (ICML 2023); Analytic-DPM (ICLR 2022); DPM-Solver (NeurIPS 2022); ProlificDreamer text-to-3D (NeurIPS 2023); Vidu4D reconstruction (NeurIPS 2024)
About Luo Yihang (罗义行)
Former Head of AI Solutions at ByteDance's Volcano Engine (cloud unit); Tsinghua University alumnus
Latest news about ShengShu Technology
- ShengShu Technology closes ~$275M Series B for general-purpose world model development
- ShengShu Technology secured a 2 billion yuan ($293 M) Series B led by Alibaba Cloud. The round will accelerate its "general world model"—a multimodal AI that perceives video, audio and text to simulat
- Alibaba Cloud led a 2 billion yuan ($293 M) round for ShengShu, the Beijing AI video generator behind Vidu and TurboDiffusion. The cash, joined by Baidu Ventures and Luminous, fuels real‑time, high‑qu
- ShengShu Technology raised over 600 million RMB in Series A+ funding to scale Vidu, the top video generator in China. By open-sourcing TurboDiffusion, efficiency increased 200x, rendering 5-second vid
More Foundation Models / LLMs companies
Official website: https://www.shengshu.com
