Global artificial intelligence infrastructure growth is actively reshaping the memory semiconductor market, driving major shifts in high-bandwidth memory and general-purpose DRAM supply chains. According to market data from Counterpoint Research, Samsung Electronics captured a 33% global revenue share in the high-bandwidth memory sector during the second quarter, marking a significant rebound from 21% in the first quarter.
Memory Competitors Race for HBM and DRAM Market Share
The latest market figures show SK Hynix holding a 50% revenue share in the second quarter, down from 58% in the previous period. That shift narrowed the gap between the two South Korean memory makers from 37 percentage points down to 17 percentage points, according to Counterpoint Research data. Meanwhile, American memory producer Micron Technology held an 18% share in the second quarter.
Micron plans to scale its production capacity significantly by the end of the year, targeting a wafer capacity of roughly monthly 100,000 wafers compared to last year’s range of monthly 40,000 to 50,000 wafers. Micron is installing manufacturing equipment across its Taiwan and Singapore production bases, with plans to raise the share of HBM4 12-stack production as high as 50% by year-end.
While South Korean and American firms compete on high-end memory output, Chinese manufacturer ChangXin Memory Technologies has expanded its presence in general-purpose DRAM. According to market data, CXMT secured a 10% global DRAM revenue share in the second quarter, reaching double digits for the first time after recording under 1% in 2023, 4% in the second quarter of last year, and 8% in the first quarter of this year. CXMT aims to scale its monthly DRAM capacity from roughly monthly 300,000 wafers up to 600,000 wafers by 2028, while also preparing for HBM3E mass production.
Executive Meetings and Next-Generation AI Accelerator Plans
Industry leaders are also realigning supply partnerships for upcoming artificial intelligence hardware. Samsung Electronics Chairman Lee Jae-yong and SK Group Chairman Chey Tae-won are scheduled to meet with Nvidia CEO Jensen Huang in New York. The discussions follow Huang’s scheduled recognition by the Korea Society, with talks expected to cover supply arrangements for next-generation AI accelerators.
Nvidia plans to integrate HBM4 memory into its upcoming Vera Rubin AI accelerators. As major memory suppliers pursue design wins for next-generation hardware, manufacturing execution and joint development partnerships with key artificial intelligence infrastructure providers will determine long-term market dominance across the sector.
Worth a look