Micron CTO: 4F² Memory Architecture Is a "Smoke Screen," Firm Commitment to 6F² Roadmap
nashnova research
Micron's CTO Scott DeBoer publicly dismissed 4F² memory architecture as 'smoke and mirrors,' saying the company will stick with 6F² and next-gen 3D DRAM after evaluating the technology for over two decades — a direct rebuttal to the market panic triggered by CXMT's recent breakthrough claim.
Why does Micron call 4F² "smoke and mirrors"?
DeBoer stated bluntly at a webinar: the density advantage of 4F² over 6F² is a misconception.
His core argument: what actually determines DRAM performance is capacitor pitch — the physical distance between capacitors — not the cell-area number. This means → even if cell area shrinks from 6F² to 4F², unchanged capacitor pitch yields no real gain in storage capacity.
He added that 4F² is more complex and more expensive, with no material edge over 6F².
Why did CXMT's announcement rattle the market?
Earlier this month, China's CXMT — the world's fourth-largest DRAM maker — announced a key breakthrough in 4F² cell architecture.
Shares of Samsung, SK Hynix, and Micron all fell on the news, as the market feared Chinese chipmakers were closing the technology gap.
In plain terms = 4F² looks on paper like it is a generation ahead of 6F²; that is why the headline spooked investors. Micron is now saying that "paper" itself is misleading.
Why is Micron betting on 3D DRAM instead of 4F²?
DeBoer explained the underlying physics: pushing 4F² to more advanced nodes hits a wall because each capacitor still needs enough stored charge for a reliable read — shrinking pitch is far harder than shrinking transistors.
This means → 4F² will hit a physical ceiling within a few process generations, while 3D DRAM sidesteps it entirely by stacking structures vertically — capacitors run in a different direction, unlocking a new scaling path.
In plain terms = shrinking further on a flat plane is nearly maxed out; building upward is the way forward — that is the core reason Micron chose 3D DRAM.
What does the EUV question mean for CXMT?
DeBoer highlighted a critical fact: whether a chipmaker uses 4F² or 6F², the node at which EUV lithography — etching circuits with extreme-ultraviolet light — becomes necessary is exactly the same.
This means → the 4F² architecture by itself does not help any manufacturer bypass the need for EUV.
CXMT cannot access EUV equipment due to U.S. export controls — so even if 4F² technology works in the lab, whether it can be mass-produced at scale remains an open question.
市场有风险，内容仅供研究参考，不构成投资建议。