1 min read
Add as a preferred source on Google

Samsung Electronics Shows zNAND-O Design for Edge AI Workloads

The memory design uses through-silicon vias to stack NAND in four- or eight-layer configurations, shortening data paths and improving speed and power efficiency.

Layered memory package resting on a semiconductor inspection platform / TokenPost.ai
Layered memory package resting on a semiconductor inspection platform / TokenPost.ai

Samsung Electronics has shown a zNAND-O memory design for edge artificial intelligence workloads as demand grows for customized storage in AI infrastructure.

The design uses through-silicon vias to stack NAND vertically in four- or eight-layer configurations. The structure is intended to shorten data paths while improving speed and power efficiency.

Samsung displayed the zNAND-O model at a memory-related event in the United States in August. The company’s approach focuses on adapting NAND storage for AI systems that require specialized performance characteristics.

The company is adapting NAND storage for AI applications, including edge systems that process data closer to where it is generated.

Samsung’s display highlights its development of storage technologies aimed at AI applications, including edge systems that process data closer to where it is generated.

Simon Yoon

Reporter

Simon Yoon reports on blockchain technology for TokenPost. Send corrections or tips to info@tokenpost.com.

Loading…