Why Micron Is Suddenly Ditching 4F² DRAM as Rivals Double Down

Micron has abandoned plans to commercialize 4F² DRAM, despite spending more than two decades researching the cell design and despite Samsung, SK hynix, and China’s CXMT pursuing it. The geometry promises a theoretical 33% density gain over the established 6F² layout, but Micron CTO Scott DeBoer argues that capacitor spacing - not the nominal cell footprint - ultimately limits density. In Micron’s view, 4F² would deliver little practical improvement while adding substantial cost and manufacturing complexity. The decision could put Micron at a near-term disadvantage if competitors produce 4F² memory economically at scale, but the company is betting that the technology will prove an expensive detour.

Instead, Micron plans to keep improving 6F² DRAM through successive process generations, including 1δ, which is ramping yields, and 1ε, now in development with advanced wafer bonding. Later roadmap stages include zeta- and eta-Angstrom technologies. Micron’s longer-term goal is 3D DRAM: memory made by stacking multiple layers of storage cells, rather than simply placing memory above logic. That transition is difficult because DRAM cells depend on tall, charge-leaking capacitors and tightly paired transistors; stacking them raises challenges involving capacitor design, heat, and signal integrity. Micron expects conventional scaling to approach its physical limits, and is preserving investment and engineering resources for 3D DRAM while using advanced lithography, materials, and bonding to extend 6F².