Samsung Electronics announced its next-generation memory and storage technology strategy on Tuesday, aimed at boosting the performance of artificial intelligence infrastructure.

The South Korean chipmaker unveiled the related proposals at the 2026 Open Compute Project Global Summit, which opened in San Francisco.

Samsung put forward three core technology directions for improving AI infrastructure performance: high-performance memory, CXL-based memory expansion technology, and data storage and management solutions equipped with next-generation solid-state drives (SSDs).

Samsung corporate vice president and head of its application engineering team, Shim Ho-jun, delivered the keynote address.

Shim said that as AI models continue to grow in size, the scale of data and key-value caches required for inference keeps expanding. To improve AI system performance, beyond memory bandwidth, the importance of memory expansion and data storage technology is steadily rising.

He added that Samsung is strengthening joint design and technical validation with global customers and partners from the early stages of research and development, carrying out full-stack optimization of the entire AI system.

Samsung also plans to jointly showcase a CXL-based memory expansion solution with AMD at this summit, providing a technical path to address the memory capacity bottlenecks of existing systems.