Wangsu Science & Technology and Sand.ai Partner for AI Video Generation Infrastructure
Wangsu Science & Technology has forged a strategic partnership with AI video generation firm Sand.ai to focus on audio-video reasoning and low-latency edge inference. The collaboration pairs Wangsu Science & Technology’s full-stack AI infrastructure platform with Sand.ai’s video generation models to drive technology co-building, market expansion, and ecosystem operations for enterprise clients.
Technical Integration of Edge Infrastructure and Autoregressive Models
Sand.ai develops foundational video models utilizing an autoregressive route where video generates block by block in a time series. This approach differs from the Diffusion architecture used by models such as Sora, Kling, and Seedance. Sand.ai’s Magi-1 model holds the top position on the Google DeepMind Physics IQ benchmark, while its newer MAGI-2 Preview ranks sixth globally on the Artificial Analysis Image to Video leaderboard. Wangsu Science & Technology supplies an edge-integrated platform combining network, intelligent computing, and security to deliver elastic compute resources, model tools, edge inference, and intelligent scheduling.
Deployment Plans and Market Projections for Video Generation
Sand.ai will prioritize deploying its newest models and technology through the Wangsu Science & Technology platform to grant industry clients early access. Both companies plan to build integrated AI video solutions for smart terminals and real-time interactive scenarios. Grand View Research estimates the global AI video market will expand from $4.55 billion in 2025 to $42.29 billion by 2030 at a compound annual growth rate of 32.2 percent.
Leadership Perspectives on the Collaboration
Wangsu Science & Technology Chairman Li Boyang stated that AI video is accelerating from technical innovation toward industrial application, with infrastructure acting as a crucial support for model value. Sand.ai CEO Cao Yue noted that combining Wangsu Science & Technology’s strengths in computing power, edge inference, and intelligent scheduling with Sand.ai’s model training expertise will improve inference efficiency, concurrent capacity, and delivery stability.