Chinese memory maker ChangXin Memory Technologies (CXMT) has rattled stock markets in South Korea and the United States with plans to roll out a next-generation DRAM product this year using what it calls 4F² technology. Blocked by U.S. export controls on advanced chipmaking equipment, China is making its bid by changing the structure of the DRAM cell itself, and some in the industry fear the move could leapfrog South Korean technology. Many analysts are more cautious, however, noting that announcing a product and actually mass-producing it are two different things. Experts were asked to examine CXMT's 4F² technology across four key questions.
① Is this really the world's first commercial 4F²?
No mention of samples leaves the timeline unclear; it should be read as a launch announcement, not mass production
CXMT said on the 10th that it plans to unveil a server-grade DDR5 memory module (RDIMM) combining a 4F² structure with hybrid bonding technology before the end of the year. It did not disclose the product's capacity or performance, nor a specific release date or mass-production schedule.
According to a recent social media post by Kwon Seok-joon, a professor of chemical engineering at Sungkyunkwan University, CXMT's announcement contains no reference to either "mass production" or "samples." Kwon said the statement should be read as a launch announcement rather than the start of full-scale mass production. He expects meaningful production volumes no earlier than the end of 2027, and realistically in the first half of 2028. Unlike building a prototype, mass-producing server DRAM still requires clearing the hurdles of securing yields and winning customer qualification.
② Is it two years ahead of Korean technology?
Samsung and SK hynix have already shown prototypes and research; density is comparable to 6F²
The term 4F² refers to the area occupied by a single memory cell. F stands for the minimum line width used as the benchmark in circuit processing. Unlike the prevailing 6F² structure (3F wide by 2F deep), a 4F² cell measures 2F in each direction. Assuming the same line width, the cell area shrinks by one-third and theoretical density rises 50%. The key is converting the conventional horizontal-channel transistor into a vertical structure to save space.
Even so, it is hard to conclude that actual density has pulled ahead. Kwon explained that CXMT's 4F² cells use a 15-nanometer line width while Korean memory makers use 12.2-nanometer 6F² cells, leaving the two cell areas similar in size. By that analysis, China's first 4F² product may be comparable in density to Korea's current 6F² products.
Samsung Electronics (005930.KS) is also developing 4F² technology. Song Jai-hyuk, president and chief technology officer at Samsung Electronics, attended the Semiconductor Exhibition on the 14th and answered "yes" when reporters asked whether the company was developing 4F² technology. On progress, the CTO said only that the company was "working hard at it." Samsung presented a working prototype of a 16-gigabit DRAM chip with a 4F² structure at this year's International Solid-State Circuits Conference (ISSCC).
SK hynix (000660.KS) has also disclosed related research results. Comparing CXMT's plan to launch this year with projections that Samsung will reach mass production around 2028, Kwon said it was a stretch to interpret China as being two years ahead.
③ Has it overcome the EUV restrictions?
Switching cell structures makes a workaround real, cutting into the time U.S. export controls bought
Behind CXMT's bet on 4F² lie U.S. export controls on extreme ultraviolet (EUV) lithography equipment. Samsung Electronics and SK hynix use EUV to shrink circuit patterns, while CXMT relies on multi-patterning, etching circuits repeatedly with deep ultraviolet (DUV) equipment. Applying the smaller-area 4F² structure instead of 6F² opens room to achieve comparable density even with a relatively wide line width.
Last month CXMT announced mass production of fifth-generation (G5) DRAM using self-aligned quadruple patterning (SAQP) without EUV. That, however, is a separate achievement from mass-producing 4F² products. Hwang Cheol-seong, a chair professor of materials science and engineering at Seoul National University and a globally recognized semiconductor scholar, said that "for China, which has difficulty securing EUV, switching the cell structure may be a realistic alternative." Kwon likewise said: "The threat in this announcement is not that China has overtaken Korean technology right now, but that the time bought by U.S. equipment export controls is running out."
④ Are yields and stability secured?
Initial yields may fall short of 10%; success would help in the 3D DRAM race
With 4F² technology, the difficulty of forming vertical-channel transistors uniformly, current leakage, the "floating body" phenomenon in which the electric potential inside a transistor becomes unstable, and wafer bonding are all cited as challenges. Even if CXMT launches a product this year, some experts estimate that initial yields could fall short of 10%. The assessment is that Korean companies, which have access to EUV, may hold an advantage in miniaturization and process efficiency.
Hwang noted, however, that "experiencing and solving these challenges first could prove advantageous in the future competition in three-dimensional (3D) DRAM." Kim Yong-seok, a chair professor at Gachon University's College of Semiconductors, said: "A gap clearly still exists, but it is narrowing."
Korean companies are also accelerating development of future technologies. Beyond HBM4E, its seventh-generation high-bandwidth memory, Samsung Electronics is pushing a concept model for HBM5 (the eighth generation), "zHBM," which stacks memory on top of an artificial intelligence accelerator, and "LPDDR5X-PIM," which incorporates computing functions, as its next-generation memory chips. SK hynix is developing HBM4E, high-bandwidth flash (HBF) and next-generation stacking technology.
With Korea focused on advancing AI memory and China on technology to work around EUV, competition in future technologies is expected to intensify further. Ahn Ki-hyun, executive director at the Korea Semiconductor Industry Association, said: "China has now climbed to the point where you can see it coming up behind you," adding that "we need to accelerate the pace of technology development."
