Kioxia recently detailed its ambitious plan to achieve a 1,000-layer 3D NAND technology by 2027. By extrapolating past trends and refining NAND cell technology, the company has observed a tenfold increase in the number of 3D NAND layers from 24 in 2014 to 238 in 2022. Kioxia asserts that, with an annual growth rate of 1.33 times, it will reach the 1,000-layer milestone, which will significantly enhance the NAND chip density.

In its quest to compete with other memory manufacturers and achieve flagship-level NAND density, Kioxia is aiming to solidify its market position. However, Western Digital, its partner, seems to have a different view on the trajectory of 3D NAND development.

At an investor meeting on June 10, Western Digital's Executive Vice President, Robert Soderbery, noted that the manufacturing cost of 3D NAND is higher than that of 2D NAND. He pointed out that the era of 3D NAND demands greater capital intensity, and the cost reduction is minimal as bit density increases. Although not directly addressing the cooperation with Kioxia, Soderbery referred to the situation as "the end of the layer race."

Soderbery stated that to optimize capital deployment, the growth rate of NAND layers will slow down. He emphasized, "We are no longer running in the layer migration like hamsters. 3D NAND layer nodes must be durable, feature-rich, and future-oriented."

In other words, the lifespan of any specific node will be extended, and Western Digital will seek to maximize the return on capital expenditure at the node level. Soderbery indicated that the strategy will involve leveraging Western Digital's stronger customer relationships to provide high-end nodes for premium use cases. Large customers will provide demand information to Western Digital, which will commit to meeting that demand through manufacturing arrangements.

Both Western Digital and Kioxia have announced their 218-layer BiCS 8th generation 3D NAND product technology. The BiCS 9 and BiCS 10 generations are expected to reach 300 and 400+ layers, respectively. These are still far from the 1,000-layer target. Considering Western Digital's stance on capital expenditure and investment returns, it is understandable that the company may be reluctant to join Kioxia in the multiple layer jumps needed to reach the 1,000-layer goal by 2027. Western Digital prefers to decelerate the layer growth rate rather than maintain or increase it.

The news, originally published on June 21, 2024, by the China Flash Market, highlights the strategic divergence between Kioxia and Western Digital in the pursuit of 3D NAND technology advancements, suggesting a potential shift in their collaborative approach to the future of memory storage solutions.