On October 21st, Wangsu Science & Technology inked a strategic cooperation agreement with AI video generation firm Sand.ai, centering on audio-video inference and low-latency edge inference. Drawing on its comprehensive full-stack AI infrastructure capabilities, Wangsu Science & Technology will join forces with Sand.ai's cutting-edge video generation technology. This collaboration aims to drive joint technology development, market expansion, and ecosystem collaboration, thereby fostering technological innovation and the widespread adoption of video generation. The core team at Sand.ai comprises experts from esteemed institutions such as Microsoft Research Asia and stands at the forefront of the global auto-regressive video generation field. Their Magi-1 represents the world's inaugural large-scale auto-regressive video generation model, while their latest innovation, MAGI-2 Preview, is the globe's first open-source video generation model boasting a hundred billion parameters and utilizing a Mixture of Experts (MoE) architecture. This partnership is set to elevate audio-video inference efficiency, enhance generation throughput speed, and bolster large-scale concurrent delivery capabilities, thereby expediting the integration of video generation technology across a broader spectrum of industries.