Micron has decided to abandon 4F² DRAM, one of the most anticipated memory architectures in the industry. After about two decades of work, the company concluded the design offers no meaningful density advantage, increases manufacturing complexity and costs, and has fundamental scaling limitations. Scott DeBoer, Micron's president and chief technology and products officer, said so at the company's event on Wednesday.

Psst!Add AlexTech.ai asPreferred Source on Google

The 4F² cell promises about 33% less area than the conventional 6F² design, which would theoretically allow around 50% more cells per wafer. The catch, according to DeBoer, is that density is set mainly by capacitor pitch rather than transistor area, since transistors must still fit beneath the capacitors. "It is a bit of a myth that 4F² actually provides anything versus 6F²," he said.

The choice puts Micron on a different path from its rivals. According to TechInsights, Samsung and SK hynix still have 4F² on their roadmaps. Micron will instead keep scaling 6F² using advanced lithography, new materials and wafer bonding. Its upcoming 1δ process, due in volume production in the second half of 2027, promises a mid-20% gain in bit density, and the company plans to adopt High-NA EUV lithography for future generations.

DeBoer acknowledged that 4F² is technically feasible and that Micron has built experimental devices with it. The problem is that complexity and cost become unsustainable as dimensions shrink. "When we look at scaling 4F² beyond where we are sitting today, there it is; it is physics," he said, pointing to 3D DRAM as the long-term successor. Micron is already developing several generations of 3D DRAM but has not said when it will reach commercial production.

Micron pulls the plug on future 4F² DRAM after two decades of research in dramatic shift official pr

Image: Micron — https://www.micron.com

The decision comes as memory is already under heavy pressure. As previously reported, Micron has warned that memory will remain scarce in 2027 and 2028, and its roadmap already centres on the 1δ node and 6F². Meanwhile, CXMT is moving into 3D NAND to diversify, while Samsung and SK hynix remain on the 4F² path.