On June 24, media reports announced that TSMC, renowned for its exclusive manufacturing partnerships with tech giants like NVIDIA and AMD, has now, in collaboration with its subsidiary Global Unichip Corporation (GUC), secured a significant order for next-generation HBM4 key interface chips.
Industry sources revealed that GUC has been awarded a design order for HBM4 chips from a leading memory chip manufacturer. The finalized design is anticipated as early as next year, employing TSMC's advanced 12nm and 5nm processes to cater to high-performance and low-power consumption needs, respectively. This new project is expected to substantially boost revenue in the latter half of the year, solidifying GUC's role in the HBM supply chain.
The HBM4's most notable innovation is the increased stack height to 16 layers of DRAM. To enhance bandwidth transmission speeds, a logic IC must be integrated into the base of the HBM, marking a significant advancement in the next-generation HBM4 technology. This critical logic IC, or key interface chip, is likely one reason JEDEC has relaxed stack height restrictions.
GUC, an ASIC design service and IP company, has released its financial results for the first quarter of 2024. The company reported consolidated revenue of NT$5.69 billion for the quarter, reflecting a year-on-year decrease of 13% and a quarter-on-quarter decrease of 9.8%. The gross margin stood at 29.7%, down 2.2% from the previous year. Net profit after tax was NT$663 million, a year-on-year decrease of 29%, with an EPS of NT$4.94.
The financial report indicates that GUC’s first-quarter revenue from design services (NRE) was NT$1.386 billion, a 7% year-on-year decline. Revenue from wafer products was NT$4.164 billion, down 16% year-on-year. Additionally, AI and networking communication application chips collectively contributed 39% of GUC's first-quarter revenue, equaling the share from consumer electronics applications. Industrial applications accounted for 14%, while other applications comprised 8%, showcasing GUC's growing expertise in AI and networking technologies.