Kioxia Discontinues 2D NAND Products, Last Shipments to Be Made in 2028 — 1980s Planar NAND Memory Reaches End of Life

Why It Matters

The discontinuation forces OEMs and industrial users to migrate to newer 3D NAND, accelerating the industry’s shift toward higher‑density, AI‑driven storage solutions and reshaping legacy supply contracts.

Key Takeaways

• Kioxia ends 2D NAND production by end of 2028
• Legacy planar NAND spans 32nm, 24nm, and 15nm nodes
• Last-time-buy orders close September 30, 2026
• Shift accelerates focus on advanced 3D NAND for AI workloads

Pulse Analysis

The retirement of planar NAND signals the close of a technology lineage that began at Toshiba in the late 1980s. For four decades, 2D flash served everything from consumer smartphones to rugged automotive controllers, leveraging relatively simple floating‑gate architectures. Over time, scaling limits and rising power demands made the planar approach untenable, prompting the industry to adopt stacked 3D structures that offer greater capacity per wafer and improved endurance. Kioxia’s decision reflects the inevitable culmination of that evolutionary arc, as the last viable 2D process nodes—32nm, 24nm, and 15nm—have become economically marginal.

For customers still reliant on legacy NAND, the announcement creates a narrow window to secure inventory. The September 30 2026 cutoff for last‑time‑buy orders means manufacturers of embedded systems, industrial equipment, and long‑life automotive modules must re‑engineer bill‑of‑materials or qualify alternative suppliers. This transition may temporarily tighten component availability, especially for niche formats like eMMC and UFS that historically depended on planar dies. However, the shift also encourages a broader industry move toward standardized 3D NAND platforms, which can deliver higher data rates and lower cost per gigabyte, aligning with the escalating storage demands of AI inference and training workloads.

Looking ahead, Kioxia’s focus will be on expanding its 3D NAND portfolio, including higher‑layer BiCS4 and beyond, to meet the explosive growth in data‑intensive applications. As AI models consume ever‑larger datasets, the need for dense, high‑throughput flash becomes a strategic priority for cloud providers, hyperscalers, and edge devices alike. The phase‑out of planar NAND thus not only cleans up legacy fabs but also frees capital for investment in next‑generation lithography and advanced packaging. In the near term, the market may see modest price adjustments for remaining 2D stock, but the longer‑term trajectory points to a unified, high‑density flash ecosystem that underpins the next wave of computing performance.