Recently, Sophgo Artificial Intelligence has launched the "Sophgo Sparse Computing Industry-Academia-Research Alliance," ushering in a fresh chapter in its technological approach centered on open collaboration and joint ecosystem development. With the widespread integration of large-scale AI models, the average daily Token processing volume for AI services in China has skyrocketed from 100 billion in early 2024 to an anticipated 100 trillion by 2026. This surge has fueled a dramatic rise in demand for computational power, compelling the industry to pivot from a focus on scale expansion to a competitive emphasis on efficiency. Sparse computing, which optimizes computational power usage and lowers overall costs by capitalizing on the sparsity inherent in models and data, has garnered widespread industry agreement. Nevertheless, hurdles remain, including formidable technical challenges, substantial development expenses, and a fragmented ecosystem.

Leveraging its proprietary dual-sparsity core technology and a holistic AI computational power hardware-software solution, Sophgo has surmounted obstacles in algorithm design, compilation, and chip architecture, paving the way for commercial deployment. The formation of the alliance is aimed at dismantling industrialization barriers through collaborative efforts among industry, academia, and research institutions, propelling sparse computing from a "niche innovator" to a "mainstream preference."

The alliance is structured around the "6S Technology Architecture" and concentrates on three pivotal areas: breakthroughs in energy efficiency, acceleration of research and development transformation, and empowerment of the ecosystem. Its vision is to establish sparse computing as the cornerstone of AI computational infrastructure, offering an efficient and all-encompassing computational foundation for diverse industries.

At present, the alliance has successfully united esteemed universities such as Tsinghua University and Fudan University, institutions like the Guangdong Semiconductor Industry Association, and industry collaborators including China Mobile and BGI Group, all working in tandem to advance the sparse computing ecosystem.