"$1.755 quadrillion in AI investment next year"... Samsung Electronics and SK hynix's time is coming

[Financial News] Major U.S. technology companies are expected to spend more than 1,700 trillion won on artificial intelligence (AI) next year. Analysts say the figure is higher than the market forecast, creating an opportunity for Samsung Electronics and SK hynix to rebound.
Kim Dong-won, a researcher at KB Securities, said on the 16th, "As of September, there are absolutely no signs of a decline in memory orders from major customers, particularly for high-bandwidth memory (HBM) and high-performance DRAM," adding, "By contrast, the rate at which customer memory demand is being met stands at only around 60% for Samsung Electronics, SK hynix, and Micron Technology."
According to Kim, as of September, United States hyperscalers' AI infrastructure investment next year is expected to reach $1.3 trillion, up 63% from the previous year, or approximately 1,755 trillion won. The market forecast is being revised upward from the previous estimate of $1.1 trillion, or approximately 1,500 trillion won.
He noted, "Semiconductor stock prices are currently pricing in a slowdown in artificial intelligence (AI) investment, but actual orders have continued to show strong momentum. Recently, the semiconductor market has focused less on earnings and more on concerns that the pace of AI technology development could be slower than initially expected."
However, he judged that "it is premature to directly link the possibility of delays in the launch schedules of individual AI models to a slowdown in the AI infrastructure investment cycle."
Kim said, "With memory manufacturers' inventories at less than 10 days' worth, an all-time low, supply capacity next year is expected to become even tighter," adding, "Major customers are therefore moving more aggressively to secure supplies in advance, taking into account the possibility of worsening shortages in 2027."
He assessed, "In the past, semiconductor stock prices underwent preemptive corrections whenever concerns about demand emerged. However, this cycle is different because 70% of total production capacity is covered by binding long-term supply contracts."
Kim's analysis is that concerns about the pace of AI development have grown, but AI infrastructure investment and memory orders are instead continuing to increase. Accordingly, he emphasized, "Now, as the gap between stock prices and fundamentals widens, is the time to prepare for a future rebound."

[email protected] Han Young-jun Reporter