Samsung Electronics is participating in “Semicon Taiwan 2026,” held in Taipei, Taiwan, from the 2nd to the 4th.

On the 1st, Samsung Electronics announced that Choi Jang-seok, Senior Vice President and Head of Product Planning Team, Memory Business, DS Division at Samsung Electronics, presented the “CUBE” strategy to overcome the limits of memory technology at the “Memory Executive Summit,” held ahead of the main Semicon Taiwan event. The strategy takes its name from the initials of △Capacity △Utilization (optimization) △Bandwidth △Efficiency (power and thermal efficiency).

Choi stated, “Memory capacity will be increased by expanding vertically rather than being constrained by PCB (printed circuit board) area.” He added, “In 3D, the key is not simply stacking, but how each layer is utilized,” and said, “Samsung is optimizing the placement of logic and memory to eliminate data processing latency.”

In line with the CUBE strategy, Samsung recently unveiled its “zHBM” and “zNAND-O (zNAND-O)” technologies. zHBM moves beyond the two-dimensional XY axes by vertically stacking HBM on top of an AI accelerator (xPU) along the Z axis. Whereas previously HBM was stacked next to the xPU, stacking HBM directly on top of the xPU reduces the physical distance between memories. This can increase performance by up to eight times compared with HBM5 and reduce thermal resistance by up to 90%.

zNAND-O is a product in which NAND flash is stacked in eight layers; although it is slower than HBM, its strength lies in storing larger amounts of data with lower power consumption. The company plans to supply samples of zNAND-O products in 2028.

Popular News