AMD Secures AI Pacts with Samsung and Naver

Seoul, South Korea – Advanced Micro Devices (AMD) has entered into memorandums of understanding with Samsung Electronics and Naver Corporation to strengthen memory supply and, by each, develop next-generation computing infrastructure for artificial intelligence (AI) systems.

At a signing ceremony held at Samsung’s semiconductor complex in Pyeongtaek, Samsung agreed to become the primary supplier of High Bandwidth Memory 4 (HBM4) for AMD’s Instinct MI455X GPUs. 

The two companies will also collaborate on DDR5 memory for AMD’s upcoming 6th Gen EPYC CPUs, codenamed “Venice,” which will underpin AMD’s Helios rack-scale architecture.

Samsung’s HBM4 utilizes a 6th-generation 10-nanometer-class DRAM process and a 4nm logic base die, delivering processing speeds of up to 13 gigabits per second and a maximum bandwidth of 3.3 terabytes per second. 

The companies will also explore potential foundry partnerships for future AMD products. Samsung already supplies HBM3E for AMD’s Instinct MI350X and MI355X accelerators.

In a separate agreement, AMD executives met with Naver leadership at Naver 1784 headquarters in Seongnam to establish a partnership to build a GPU computing environment for HyperCLOVA X, Naver’s large language model. 

The collaboration will integrate AMD infrastructure into Naver’s data centers, cloud services, and user applications. 

AMD and Naver also plan to provide AI computing resources for academic researchers and launch joint research initiatives to expand the AI ecosystem.