Cohu, Inc., a global supplier of equipment and services optimizing semiconductor manufacturing yield and productivity, today announced that its Neon platform was selected by a leading U.S. memory and data storage technologies customer for inspection metrology of high bandwidth memory devices (HBM) used in high-performance computing (HPC) and generative artificial intelligence (AI) applications.
Neon offers a highly efficient vision system enabling full 6-sided optical inspection and measurement of micro-pillars, integration with factory automation and an efficient output sorting process that addresses current and roadmap HBM market needs. Cohu’s vision technology delivers accurate and reliable metrology of thousands of micro-pillars distributed over the HBM device. The accompanying AI inspection software utilizes proprietary deep learning and neural network-based pattern recognition that improves inspection yield. The HBM market is projected to be approximately $23 billion this year with a CAGR of 22% through 2029, representing an opportunity for more than $100M revenue for this class of inspection metrology systems.
“We are pleased to have received an initial order from one of the world’s leading semiconductor memory manufacturers to improve quality of HBM memory used in AI data centers,” commented Luis Müller, Cohu President and CEO. “This is a new market application for Cohu products, and it represents a strategic and substantial growth opportunity aligned with secular growth in artificial intelligence.”