SK Hynix has already shipped samples of its sixth-generation HBM4 memory to key customers and plans to start mass production of 12-layer HBM4 products by the end of 2025.

The company is set to invest over $14 billion in a new DRAM fabrication facility in Cheongju, South Korea, with construction beginning by the end of April 2025 and mass production expected to start in November.

However, SK Hynix cautioned about potential demand volatility in the second half of the year due to macroeconomic uncertainties and tariff policies.

SK Hynix has reported a significant increase in quarterly revenue and operating profit, primarily driven by robust demand for high bandwidth memory (HBM) used in AI chipsets.

Despite concerns regarding potential U.S. tariffs on chip sales, the company anticipates limited impact, as approximately 60% of its revenue comes from U.S. customers, with much of the product delivered outside the U.S.

While analysts express concerns about a potential decline in demand in the latter half of 2025 due to inventory build-up, SK Hynix currently sees no signs of a significant inventory glut.

In the first quarter of 2025, SK Hynix began supplying LPCAMM2, a high-performance memory module for AI PCs, and plans to launch SOCAMM, a low-power DRAM module for AI servers, as demand continues to rise.

Despite strong financial results, SK Hynix shares fell by more than 1% in morning trading, reflecting investor caution regarding future uncertainties rather than immediate performance concerns.

The company emphasizes the significant role of AI in boosting memory market performance, positioning itself as a leading player in this rapidly evolving sector.

Looking ahead, SK Hynix forecasts that its HBM shipments will double in 2025, with 12-layer chips expected to account for over half of its HBM3E sales in the second quarter.

The company reassured stakeholders that global customer demand remains stable, with some clients even requesting short-term supply advances amid uncertainties in semiconductor demand.

In addition to data centers, there is increasing demand for AI-enabled smartphones, indicating a dual-track growth strategy in both enterprise and consumer markets.