SK Hynix Inc's chairman, Chey Tae-won, told reporters on Monday that the current memory chip shortfall is expected to persist until 2030 due to unusually strong demand from the artificial intelligence sector. The remarks came on the sidelines of NVIDIA's GTC conference in San Jose, California.
Chey also disclosed that SK Hynix is reviewing the possibility of an American Depositary Listing. In addition, the company is reportedly preparing to outline measures aimed at stabilising DRAM chip prices as supply constraints tied to AI-driven demand continue to strain the market.
As one of the world's largest memory manufacturers, SK Hynix is a major supplier of high-bandwidth memory to NVIDIA. Chey attributed the current tightness in supply to outsized demand for HBM from AI customers, saying that the company faces a shortage of wafers and will require more time - "at least four to five years" - to ramp wafer production sufficiently to meet that demand.
Over the past three years, SK Hynix has seen significant gains from elevated AI-related demand for memory products. The company has recently benefited from rising memory chip prices, a consequence of the sector's supply shortages, which pushed prices higher.
Independent industry data cited in recent reporting indicated that server memory chip prices climbed between 60% and 76% in the fourth quarter of 2025, and that further increases were expected into the first quarter of 2026. Against that backdrop, market reaction included a rise in SK Hynix shares of over 2% on Tuesday.
Key details:
- Chairman Chey Tae-won said the memory shortage is likely to last until 2030, driven by AI demand for HBM.
- SK Hynix is reviewing a potential American Depositary Listing and may announce actions to help stabilise DRAM prices.
- Company estimates it needs at least four to five years to build up additional wafer capacity to address the shortage.
Market context:
- SK Hynix is a leading memory chip maker and a key supplier of high-bandwidth memory to NVIDIA.
- Server memory chip prices surged substantially in Q4 2025, with further rises expected in Q1 2026 according to the recent data referenced.
Implications for sectors:
- Semiconductors and hardware suppliers face prolonged supply constraints affecting pricing and availability.
- AI infrastructure and data center operators may continue to encounter elevated procurement costs for memory components.