SK hynix is operating a customer exhibition booth at the Venetian Expo during CES 2026, held from January 6 to 9 local time, to display its next-generation AI memory solutions under the theme "Innovative AI, Sustainable tomorrow."
The company is focusing on the customer booth this year, after previously running both a SK Group joint exhibition and a customer booth, to increase interactions with key customers and discuss collaborations.
SK hynix is presenting its 16-layer HBM4 with 48 GB capacity for the first time at the event. This follows the 12-layer HBM4 with 36 GB, which achieved a speed of 11.7 Gbps, and is being developed in alignment with customer timelines.
The booth also features the 12-layer HBM3E with 36 GB, expected to drive the market in 2026, including a joint exhibit with a customer showing GPU modules incorporating HBM3E for AI servers.
Additional products on display include SOCAMM2, a low-power memory module for AI servers, and LPDDR6, which provides improved data processing speed and power efficiency for on-device AI compared to prior generations.
In NAND flash, SK hynix is exhibiting its 321-layer 2 Tb QLC product for ultra-high capacity enterprise SSDs, offering higher integration, power efficiency, and performance than previous QLC generations to address demand from expanding AI data centers.
An "AI System Demo Zone" demonstrates interconnections of future AI memory solutions, including custom HBM (cHBM) that integrates functions into the HBM base die per customer needs to improve GPU/ASIC performance and reduce data transfer power; PIM-based AiMX; CuD for in-memory computing; CMM-Ax adding computation to CXL memory; and data-aware CSD.
A large-scale mock-up of cHBM is available for viewing its structure, reflecting the shift in AI competition toward inference efficiency and optimization.
Justin Kim, President and Head of AI Infra at SK hynix, stated that as AI innovation accelerates, customer requirements evolve quickly, and the company will address them through differentiated memory solutions and close customer cooperation to advance the AI ecosystem.
