Why Micron Is Passing on 4F² DRAM

Micron will not pursue 4F² DRAM, Micron executive Scott DeBoer said at a company event on October 21, according to an account published by Tom’s Hardware on October 22. DeBoer said the company had evaluated the design for roughly two decades but cited added manufacturing complexity, cost and fundamental scaling problems as reasons to reject it.

Why a smaller cell does not guarantee denser DRAM

The 4F² label describes a cell’s nominal area relative to F, the minimum lithographic feature size. At identical feature dimensions, a 4F² cell has about 33% less theoretical area than a 6F² cell. An idealized layout could fit 50% more cells per wafer, but that geometric comparison does not establish a matching gain in practical bit density.

DeBoer said practical DRAM density depends on capacitor pitch, size and dimensions, not just the nominal transistor layout. Capacitor pitch is the spacing between neighboring capacitors; the transistor also has to fit beneath the capacitor arrangement. Those constraints help explain why a smaller cell-area figure alone does not settle how many usable bits a manufacturing process can produce.

Micron plans to keep scaling 6F² and develop 3D DRAM

Micron’s reported direction is to continue scaling conventional 6F² DRAM and develop 3D DRAM, which stacks memory-cell layers. That is a different approach from the vertical-channel transistor reported as part of YMTC’s separate 4F² development target.

Tom’s Hardware reported that Micron’s 1δ process was scheduled for volume production in the second half of 2027, with the company expecting a bit-density improvement in the mid-20% range over its predecessor. The projected gain is tied to that process, not to a move to 4F².

CXMT and YMTC have separate reported 4F² targets

Fanáticos del Hardware reported that CXMT targeted a 4F² DDR5 RDIMM server module before the end of 2026. The outlet also reported that YMTC’s separate 4F² vertical-channel-transistor design targeted 9,600 MT/s. Those targets describe separate development efforts, not Micron’s roadmap.