Samsung Electronics is accelerating its layout in the AI PC-specific chip market, with its self-developed accelerator processor codenamed "GAIA" planned for mass production in 2027. The chip was developed by the System LSI division under Samsung's Device Solutions business unit, and has already completed prototype samples, which have been sent to major PC manufacturers such as Lenovo and HP for testing, officially entering the performance verification phase.

GAIA uses a 4nm manufacturing process, with its core architecture centered around a Neural Processing Unit (NPU). Unlike traditional CPUs and GPUs, this chip focuses on accelerating processing tasks related to generative AI. At the same time, Samsung is exploring collaborative solutions between GAIA and Processing-In-Memory (PIM) technology—PIM can perform data computation directly within storage devices, helping to reduce latency and energy consumption caused by data transfer, which is crucial for energy efficiency optimization in edge-side AI.

Beyond PC chips, it also targets the robotics market

In terms of application positioning, GAIA's ambitions go beyond edge AI computing scenarios on the PC side, and it also targets the physical AI application markets such as robotics. In terms of power consumption design, Samsung emphasizes optimizing performance per watt, but so far, specific performance metrics, interface specifications, and adaptation plans for existing PC platforms of GAIA remain undisclosed.

This is not Samsung's first entry into the PC processor field. In 2012, Samsung had developed Exynos application processors for Chromebooks, but the relevant business was terminated two years later. Now, as AI PCs have become the new competitive high ground in the semiconductor industry following data centers, companies such as NVIDIA, Qualcomm, and Huawei have already entered the market. If the project proceeds as scheduled, Samsung will officially start mass production of GAIA in 2027, adding another heavyweight player to the competition in AI PC chips.