TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Micron says it is abandoning plans to bring 4F² DRAM to market after more than two decades of research, and will continue scaling its 6F² cell design. The decision contrasts with reported 4F² efforts by Samsung, SK hynix and China’s CXMT; whether those rivals can produce the design economically at scale remains uncertain.
Micron says it will not bring 4F² DRAM to market, ending a research effort that the company says lasted more than two decades. The decision, presented as part of its future memory roadmap, leaves Micron committed to scaling 6F² DRAM as Samsung, SK hynix and China’s CXMT pursue 4F² designs.
Micron President and Chief Technology Officer Scott DeBoer said the smaller cell geometry does not deliver a substantial density improvement in practice, because the required spacing between DRAM capacitors limits how closely cells can be packed. He said moving from 6F² to 4F² would add cost and manufacturing complexity without a commensurate density gain. The account of the webinar, reported by HotHardware, does not provide a quantified estimate of the gains or costs.
Micron’s stated alternative is to keep advancing its 6F²-based roadmap. Its 1δ (1-delta) DRAM is ramping yields, while 1ε (1-epsilon), which involves advanced wafer bonding, is in development. The roadmap also lists later planar memory generations as ζÅ (zeta-Angstrom) and ηÅ (eta-Angstrom). These are Micron’s generation labels, not direct equivalents to logic-chip process-node names.
The company also identified 3D DRAM as a longer-term direction. That approach would stack layers of DRAM storage cells vertically, rather than simply bonding memory to logic. It is a more fundamental architectural change: DRAM cells combine a transistor with a capacitor, making vertical stacking more complicated than in flash memory. Micron’s presentation, as described in the report, did not give a commercialization date for 3D DRAM.
Micron Bets on Capacitor Limits
The decision sets up a strategic difference in how memory makers are pursuing density. 4F² promises a theoretical 33% density improvement over 6F² based on cell geometry, but that theoretical figure does not establish the practical gain in manufactured products. Micron’s position is that capacitor spacing, rather than the nominal cell footprint alone, constrains useful density.
If Micron’s assessment proves right, continuing to refine 6F² could avoid the expense and complexity of a 4F² transition while the company develops other scaling methods. If rivals can make 4F² products at competitive cost and volume, however, Micron could face a near-term disadvantage. The available information does not establish which path will deliver better commercial results.
The stakes extend beyond a naming dispute. DRAM density affects how much memory can be made in a given area, while manufacturing costs and yields influence the economics of supplying memory for computers, data centers and other devices. The roadmap points to Micron’s priorities, but does not disclose product pricing, output targets or a schedule for any 4F² or 3D DRAM products from competitors.
Popular right now
Roborock Q7 M5+ Robot Vacuum and Mop
Philips 5500 Series Fully Automatic Espresso Machine with LatteGo
Dyson Cyclone V10 Origin Cordless Vacuum
As an affiliate, we earn on qualifying purchases.
From 6F² Cells to 3D Memory
F² describes DRAM cell geometry, not a chip’s physical size or a process node. The notation refers to the area relationship between a cell’s layout and a characteristic feature size, represented by F. Actual density also depends on the manufacturing process and the design of the cell’s transistor and capacitor. DRAM’s mixed structure means its scaling challenges differ from those of logic chips, which are built around switching transistors.
For years, 6F² has been the established geometry, while 4F² has been treated as a possible way to increase density. HotHardware’s account says Micron had researched 4F² for more than two decades but concluded it would not commercialize the design. The company’s latest roadmap instead describes continuing 6F² generations and eventual work toward vertically stacked DRAM cells.
According to the report, Samsung and SK hynix are among the companies pursuing 4F², along with CXMT. CXMT has claimed it will have 4F² memory in server DDR5 registered DIMMs by the end of the year. That is a company claim reported by HotHardware, not confirmation that such products have shipped or reached high-volume production.
“DRAM density is ultimately set by the spacing required by the capacitors, not by the nominal cell footprint.”
— Scott DeBoer, as summarized by HotHardware
Rivals’ Production Plans Remain Unproven
It is not yet clear whether Samsung, SK hynix or CXMT can manufacture 4F² DRAM at competitive cost, yield and volume. The supplied report describes their efforts but does not provide independently verified production data, product specifications or comparisons with Micron’s current and planned memory.
Micron has not disclosed a detailed timetable for its later 6F² generations or for commercial 3D DRAM in the material provided. Nor does the report quantify how much density 4F² would add in a real product after capacitor-spacing constraints are accounted for. DeBoer’s explanation is Micron’s technical rationale, not a published head-to-head manufacturing comparison.
Watch for Yield and Product Evidence
The next signals will come from Micron’s 1δ yield ramp and progress on 1ε, as well as any detailed schedules or products tied to its later roadmap generations. Evidence of manufacturing yields, capacity and actual product density would help clarify how much improvement Micron can achieve while staying with 6F².
For its rivals, the key test is whether 4F² moves from development and company projections into commercially produced memory. CXMT’s stated end-of-year target is a claim to watch, not a confirmed delivery milestone. Further technical disclosures or product announcements will show whether Micron’s decision leaves it behind on density—or avoids a costly step that competitors may struggle to scale.
Key Questions
What has Micron decided about 4F² DRAM?
Micron will not commercialize 4F² DRAM, according to its roadmap announcement as described by HotHardware. It plans to continue scaling 6F² memory instead.
Why is Micron rejecting 4F²?
Micron CTO Scott DeBoer said capacitor spacing limits density, reducing the practical benefit of a smaller cell footprint. He also said the 4F² approach would add cost and complexity.
Are Samsung, SK hynix and CXMT making 4F² DRAM?
The report says those companies are pursuing 4F² designs. It does not confirm that they have achieved economical, high-volume production. CXMT has claimed it plans server DDR5 RDIMMs using 4F² by the end of the year.
What is Micron pursuing instead?
Micron is advancing 6F² DRAM generations, including 1δ, which is ramping yields, and 1ε, which is in development with advanced wafer bonding. It also sees vertically stacked 3D DRAM as a longer-term direction.
Source: hothardware.com
