Following NVIDIA, Samsung Develops "GAIA" AI PC Chip, Samples Sent to Lenovo and HP for Testing

Samsung Electronics is quietly laying out its strategy in the AI PC chip market. The System LSI division of the South Korean tech giant is developing a generative AI acceleration chip specifically designed for AI PCs, codenamed "GAIA." Prototype samples have already been provided to major PC manufacturers such as Lenovo and HP for performance verification, with mass production expected to begin as early as next year.

According to industry insiders, GAIA adopts a 4nm manufacturing process and is positioned as a "memory-centric AI accelerator," with a core design philosophy that places computing functions as close to memory as possible. Samsung is also advancing the integration of this chip with next-generation DRAM technology known as Processor-in-Memory (PIM), which has the capability to perform computations directly while storing data.

Unlike acceleration chips equipped with GPUs that are primarily used for AI training and inference, GAIA is specifically optimized for Neural Processing Unit (NPU) architecture and designed for generative AI tasks on PCs. This move marks Samsung's official entry into the emerging AI PC market.

Transplanting NPU Chips to PCs

GAIA is not an entirely new product line for Samsung but rather an extension of its mobile NPU technology to the PC scenario.

Samsung's System LSI division has long been deeply involved in the mobile application processor (AP) sector, with its Exynos series being the representative product. Essentially, GAIA involves re-adapting NPU chips originally designed for mobile devices for use in PC scenarios.

Samsung is not without precedent in the PC chip sector. In 2012, Samsung introduced Exynos processors into Chromebooks, but the project was discontinued two years later. More than a decade later, Samsung has chosen to re-enter this market with AI PCs as the entry point, adopting a more focused approach.

GAIA differs fundamentally from traditional PC processors in its positioning—the latter serves as the "brain" of the PC, handling general-purpose computing tasks, while GAIA focuses on AI operations, with the core objective of efficiently processing generative AI workloads.

Memory-Compute Fusion: A Differentiated Technical Route

In terms of technical route, a significant feature of GAIA is its "memory-centric" architectural concept, which deeply integrates computing units with memory to reduce latency and power consumption caused by frequent data transmission between the processor and memory.

Samsung is promoting the combination of GAIA with PIM technology. PIM is a next-generation DRAM technology that allows computational operations to be performed directly within memory chips, eliminating the need to repeatedly transfer data to a separate processor. If this integration is achieved, GAIA could possess potential efficiency advantages in local AI inference scenarios.

As the world's largest memory chip manufacturer, Samsung has natural integration advantages in PIM technology. The architectural design of GAIA also reflects Samsung's strategic intent to connect its semiconductor memory business with logic chip capabilities to create synergy.

Mass Production Timeline and Market Window

GAIA has currently entered the critical stage of supplying samples to customers. Lenovo and HP are verifying the performance of the prototype chips. As both companies are leading global PC manufacturers, their procurement decisions will be decisive in determining whether the chip can enter the mainstream market.

Samsung expects GAIA to achieve mass production as early as next year. This timeline largely aligns with the cycle of the global PC industry accelerating its transformation toward the AI PC form factor.

Currently, the AI PC market is seeing layout efforts from multiple parties. NVIDIA recently made a high-profile announcement of its entry into the Windows PC processor market, Qualcomm's Snapdragon X series has already secured a foothold, and Intel and AMD are continuously enhancing NPU capabilities on their respective platforms. By entering with a standalone AI acceleration chip, Samsung has chosen a differentiated path—not seeking to replace the main processor, but rather working in collaboration with existing PC platforms as a specialized AI computing module.

Whether GAIA can smoothly transition from performance verification to mass production implementation and secure formal adoption by Lenovo and HP will be the key node testing whether this strategy can be realized.