
The huge profits of SK Hynix and Samsung are opportunities for Chinese memory companies

If the imbalance between supply and demand for storage chips continues until the first half of 2027, the period of high profits for South Korean companies may become an opportunity window for Chinese manufacturers to narrow the technological gap. Samsung, SK Hynix, and Micron's production capacity cannot meet global demand, providing market entry opportunities for Chinese storage manufacturers. If South Korean companies fail to convert their high profits into technological barriers, Chinese competitors may erode their advantages after the market cools down
The surge in storage chip prices has brought substantial profits to Samsung Electronics and SK Hynix, but this ongoing supply shortage is opening a market window for Chinese memory manufacturers. Industry experts warn that if the supply-demand imbalance persists until the first half of 2027 or even longer, the era of exorbitant profits for Korean companies may instead become a golden opportunity for Chinese competitors to narrow the technological gap.
According to a report by DIGITIMES on Tuesday, at a recent seminar on next-generation semiconductor technology trends held in South Korea, Professor Seok-jun Kwon from Sungkyunkwan University pointed out that the memory industry is experiencing a super cycle, with DRAM prices soaring by 300-400% in just a few months. However, Samsung, SK Hynix, and Micron are unable to meet global demand, creating opportunities for Chinese memory manufacturers to enter the market.
Kwon emphasized that the expansion potential of Chinese companies is not limited to the consumer retail sector but also includes the enterprise market. Even capturing just 5-10% of the market share would be sufficient to accumulate momentum for future growth.
If Chinese companies can gain more technical experience during this window period, it will have a profound impact on the global memory chip landscape.
The Speed of Technological Catch-Up Exceeds Expectations
Chinese memory manufacturers are catching up to their Korean counterparts at an astonishing pace. By early 2025, when SK Hynix supplies the third-generation high-bandwidth memory (HBM3E) chips, ChangXin Memory Technologies (CXMT) is still developing HBM2 technology, but by mid-2025, reports indicate that it has already advanced to the HBM3 development stage.
Kwon pointed out that CXMT's accelerated progress is not only due to its own funding but also benefits from investments by local governments and Huawei. New technology testing is not only conducted in CXMT's own factories but also frequently takes place in factories in Guangdong and Shanghai invested in by Huawei, allowing the company to identify feasible solutions before scaling up production internally. This model has enabled CXMT to shorten its R&D cycle and quickly commercialize technological innovations.
Strategic Significance of the Market Window
For Chinese companies, the current market opportunity has dual value. On one hand, the global supply tightness has reduced customers' sensitivity to technological gaps, creating market space for relatively lagging products. On the other hand, the sustained high-price environment provides Chinese companies with ample profit margins to support investment in technological R&D.
Kwon warns that if the supply shortage continues until the first half of 2027 or later, it effectively gives Chinese manufacturers more time to acquire technical expertise. The real risk facing Korean companies is that if they fail to convert their current exorbitant profits into higher technological barriers, once the market begins to cool, Chinese competitors may erode their advantageous position.
Technological Watershed in the AI Era
Leading memory companies must adapt to the AI era to maintain their positions. The performance bottleneck in AI has shifted from core computing power to storage, with power consumption particularly determining whether memory manufacturers can enter the edge AI market. Although HBM remains the dominant technology, its structural limitations mean that alternatives will inevitably emerge.
**According to Kwon, a key trend is the increased use of SRAM: despite being more expensive and taking up more space, its ultra-low latency makes it more attractive than traditional DRAM in certain AI models. Future technological advancements will focus on silicon through-silicon vias (TSV) density, shortening the distance between memory and core, and developing new interconnect technologies that go beyond silicon interlayers **
