Livejust nowSingapore-Based ShengShu Technology Unveils Vidu S1 for Real-Time AI Video
← Back to stories

Singapore-Based ShengShu Technology Unveils Vidu S1 for Real-Time AI Video

Source: PR Newswire

ShengShu Technology, the Singapore-headquartered AI company behind the Vidu video generation platform, has unveiled Vidu S1 — a next-generation video foundation model capable of real-time interactive generation.

Singapore-Based ShengShu Technology Unveils Vidu S1 for Real-Time AI Video
SGAI Daily

ShengShu Technology, the Singapore-headquartered AI company behind the Vidu video generation platform, has unveiled Vidu S1 — a next-generation video foundation model capable of real-time interactive generation. The model transforms AI video from a one-shot clip generator into a live, conversational experience that responds to voice input in real time.

Unlike conventional video generation models that process a single prompt and produce a fixed clip, Vidu S1 adopts an autoregressive diffusion architecture that continuously predicts and generates subsequent frames based on prior output, live voice instructions, and conversational context. The system delivers 540P resolution at 25 frames per second (up to 42 FPS) and can run on consumer-grade GPUs — a notable engineering achievement for real-time video AI. Users can create interactive characters from a single uploaded image, whether a person, anime character, or pet, paired with a customisable voice.

The announcement, made at the 2026 Global Digital Economy Conference, positions Vidu S1 as a foundation for applications spanning AI companionship, interactive livestreaming, game NPCs, branded avatars, virtual customer service, and XR experiences. ShengShu achieved the real-time performance through a combination of inference acceleration techniques including TurboDiffusion, low-bit SageAttention, and TurboServe — an inference serving engine that dynamically allocates compute resources based on interaction state. This stack allows continuous real-time generation across extended conversations without degrading visual quality or responsiveness.

Why it matters for Singapore: ShengShu Technology represents a growing cohort of AI companies that maintain their global headquarters in Singapore while competing at the frontier of foundation model development. The company's ability to ship real-time interactive video generation — a category that companies like OpenAI, Runway, and Pika are also pursuing — demonstrates that Singapore-based AI firms can hold their own in the most technically demanding segments of the AI race. For Singapore's broader AI ecosystem, Vidu S1 shows that the city-state's role as a base for ambitious AI companies extends beyond enterprise SaaS into cutting-edge generative AI infrastructure.

Your daily AI edge in Singapore: in <5 minutes.

We do the reading so you don't have to. Get the essential TL;DR on local AI moves delivered to your inbox every morning.