SK Hynix: No Signs of Slowdown in AI Infrastructure Investment, Expecting Steady Growth Beyond 2027

South Korean memory giant SK Hynix stated on its earnings call on the 29th: 'We have noticed that some large tech companies are re-evaluating data center leasing businesses and the rise of high-performance AI models, leading to concerns that AI infrastructure investment may slow down. We believe these actions are not a process of reducing AI investment but a process of improving utilization rates and accelerating the monetization of large-scale AI infrastructure.'

SK Hynix believes that the proliferation of high-performance AI models is unlikely to lead to a decrease in demand for infrastructure and memory. As model and system efficiency improves, more users can access various services on the same infrastructure, expanding the accessibility and coverage of AI services. Given that even the recently launched high-efficiency AI models have explosive user demand, efficiency improvements are promoting service proliferation and increasing overall utilization rates.

SK Hynix pointed out that in mid-to-long-term demand negotiations with major customers, the sustainability of AI investment has been confirmed, and it is believed that AI infrastructure investment will remain robust after 2027.

In response to market demand growth, SK Hynix is accelerating the mass production of the DRAM production base M15X in Cheongju, Chungcheongbuk-do, South Korea. Mid-to-long-term investment plans include the development of the advanced packaging facility P&T7, the NAND production base M17, and the construction of new semiconductor cluster development projects, which will be carried out in stages based on customer demand and investment benefits.

Although SK Hynix's financial data grew rapidly in the second quarter of this year, its operating profit was below the previous market forecast of approximately 64 trillion won, and its revenue was below the market expectation of approximately 84 trillion won.

Analysts pointed out that SK Hynix failed to meet analyst expectations because the company's business share in the field of high-end memory chips used in AI data centers was higher than that of its competitors. This means that it benefited less from the strong rise in traditional memory chip prices, which further fueled concerns that the AI-driven semiconductor industry boom may be slowing down.