Samsung Takes On Nvidia, Qualcomm in AI PC Chip Push

Built on a 4-Nanometer Process, Mass Production Could Begin as Early as Next Year

Samsung Electronics is moving into the market for AI accelerators used in AI PCs. The company is expanding beyond general-purpose chips into high-performance semiconductor design for physical AI products such as AI PCs and robots as it seeks a new growth engine. The move sets up a direct contest with Nvidia and Qualcomm, which have already entered the PC accelerator market.

Industry officials said on July 9 that Samsung Electronics' System LSI Business, part of its Device Solutions division, is developing an AI accelerator for AI PCs called GAIA. The company has already supplied prototypes to major customers including China's Lenovo and US-based HP for performance testing. Mass production could begin as early as next year. The chip will be made using a 4-nanometer process. Samsung is also pursuing integration with processing-in-memory, or PIM, a next-generation DRAM technology that can process data where it is stored.

Unlike conventional processors that serve as the brain of a PC, GAIA is specialized for AI computing. It is designed with an optimized neural processing unit, or NPU, architecture to handle generative AI tasks on PCs more efficiently.

The industry views AI PCs as the next major battleground for AI semiconductors after AI data centers. PCs are expected to be the first key platform to bring about the era of AI agents.

Nvidia, Qualcomm and Huawei have also entered the AI PC accelerator market. Samsung is set to compete fiercely with them for an early lead.

Kang Hae-ryeong / Kim Chae-yeon, Korea Economic Daily reporters hr.kang@hankyung.com