Samsung, one of the largest semiconductor companies in the world, has announced a partnership with NAVER to develop chips optimized for hyper-scale AI models. It is a comprehensive partnership with NAVER, the world’s leading internet company with AI expertise.
The collaboration will leverage Samsung’s cutting-edge products, such as computational storage (SmartSSD), CXL (Compute Express Link), PIM (Processor In Memory), and PNM (Processor Near Memory), and pool their hardware and software expertise to significantly improve performance during massive AI workloads.
According to the companies, recent AI developments have led to an exponential increase in data volumes. However, the current computational technologies have performance and efficiency limitations that prevent them from effectively managing these enormous workloads. Therefore, NAVER and Samsung have decided to collaborate on the development of AI-optimized semiconductor chips and additional solutions.
Samsung Electronics will contribute its hardware and manufacturing expertise, while NAVER will contribute its expertise in AI-based algorithm development and verification. NAVER will optimize HyperCLOVA, a hyper-scale model with over 200 billion parameters, and its compression algorithms.