Memory Giants Split Over 4F² DRAM Strategy
Micron (MU) has abandoned the 4F² DRAM cell structure, focusing instead on 3D DRAM and improving the 6F² structure. Samsung (005930.KS) and SK hynix (000660.KS) continue developing 4F². Micron's 1-delta (1δ) DRAM is set for mass production in 2H 2027. The companies' differing strategies may impact the future DRAM market.
How this was made
The 30-second read
Why it matters
Micron’s explicit non-adoption of 4F² is a strategic roadmap change that can reprice technology optionality and execution risk for next-generation DRAM, while Samsung and SK hynix maintain 4F² development as a candidate.
Market read
Traders may adjust DRAM technology-risk positioning as Micron diverges from 4F² while peers keep it, with timelines pointing to 1δ in 2H 2027 and longer-term 3D stacking.
What to watch
The article lacks quantified cost/performance outcomes, yields, and customer qualification status; investors may overreact to the 4F² abandonment without evidence that 3D/wafer bonding will outperform on a timeline that matters for revenue.
Background
4F² is a proposed DRAM cell structure intended to reduce cell area versus 6F², potentially increasing integration density, but Micron argues it does not translate cleanly into manufacturing density due to capacitor spacing and complexity.
Ticker impact
Micron says it will not pursue commercialization of the 4F² DRAM cell structure and will instead prioritize 3D DRAM and wafer bonding.
Pressure on MU’s near-term 4F² optionality narrative, with investors refocusing on 3D DRAM progress and manufacturing economics.
The article is a direct roadmap policy change from Micron, but it does not provide financial guidance or immediate production numbers, limiting magnitude.
Market effects
Shifts the DRAM technology debate from pure cell miniaturization toward integration-density economics, potentially affecting how investors price 4F² versus 3D DRAM execution risk across the memory complex.
Could influence sentiment across US-listed DRAM exposure and Korean memory supply-chain expectations as the competitive roadmap diverges.
May affect global DDR5/DRAM supply expectations and technology leadership narratives, especially if 4F² economics prove less favorable than 3D stacking.
Counterpoint
Micron’s decision may not be bearish for near-term competitiveness if 4F² is not the binding constraint; wafer bonding and 3D DRAM could deliver better real integration density and cost sooner than the market assumes.
Key entities
- companyMicron Technology
Says it will not commercialize 4F², focusing on existing process improvements, wafer bonding, and 3D DRAM; targets 1δ mass production in 2H 2027.
- companySamsung Electronics
Confirms continued 4F² development, using Vertical Channel Transistors (VCT) to reduce cell area.
- companySK hynix
Promotes next-generation cell structure based on Vertical Gates (VG), sustaining 4F² as a candidate.
- companyChangXin Memory Technologies (CXMT)
Previously planned 4F²-based server DDR5 modules, but performance, yields, and cost competitiveness remain unconfirmed.
