The collaboration signals Samsung’s bid to lead in on-device AI chips by leveraging Arm’s architecture and the possibility of OpenAI involvement, highlighting a broader move toward edge AI hardware.

The project aims to enable on-device AI computations on terminal devices like smartphones without routing data to the cloud, reducing cloud compute demand, shrinking latency, and boosting data privacy.

Samsung’s System LSI unit will design the SoC around Arm’s AI architecture, with Samsung Foundry enabling mass production on a 2nm process, while Arm coordinates requirements and progress without acting as the chip vendor.

Arm Holdings has approved non-recurring engineering funding to develop a next-generation on-device AI SoC and has launched a joint R&D project with Samsung to pursue 2nm mass production.

The endeavor is portrayed as a strategic design-and-manufacturing alliance for Samsung, potentially setting a reference for future AI accelerators and expanding the 2nm edge AI chip ecosystem.

OpenAI is discussed as a potential end customer for the on-device chip, underscoring rising interest from AI firms in edge hardware alongside cloud development.

The collaboration operates under a Limited Use License, with Arm providing core technology (AI accelerator architecture and RTL), and Samsung handling full SoC design, procurement, mass production, and end-customer delivery and settlements.