Samsung unveils AI-focused memory expansion and partners with AMD at OCP summit

During the OCP Global Summit, Samsung Electronics presented a suite of next-generation memory and storage products designed for the growing demands of large AI models. The exhibit featured customized HBM, HBM5, and processing-in-memory chips that reduce data movement and improve energy efficiency. In partnership with AMD, Samsung demonstrated a Compute Express Link (CXL) memory expansion solution that enables flexible memory sharing across servers, easing AI inference workloads.

Additional highlights included 2-nanometer gate-all-around wafers, Groq language processing unit wafers, and Nvidia-related hardware, illustrating Samsung’s reach into advanced chip manufacturing. Enterprise SSDs with key-value cache optimization were also shown, and the Samsung MAX Lab was promoted as a testing ground for customers to evaluate these technologies under real AI workloads before commercial rollout.