The AI boom is turning memory chips, long seen as a cyclical part of the semiconductor industry, into one of the hottest deals in the market.
One of the standout winners in this trend is South Korea’s Samsung Electronics, whose market capitalization reached the $1 trillion mark last week. The stock is up more than 140% this year.
This week, another South Korean semiconductor giant has emerged as a candidate for this milestone.
SK Hynix stock has soared more than 200% this year as demand for high-bandwidth memory, a specialized chip used in AI systems, has exploded.
The company’s phenomenal stock price rise has increased its market capitalization to approximately 1.395 trillion Korean won ($935 billion).
SK Hynix is Nvidia’s top supplier of memory chips, and its AI accelerators are the backbone of the AI race.
SK Hynix’s stock price rise comes as companies rush to build AI data centers and investors increasingly focus on the hardware that powers AI models, not just the companies that design them.
“Memory is becoming one of the most obvious beneficiaries of AI hardware demand,” Charu Chanana, chief investment strategist at Saxo, said on Tuesday.
Demand is increasing as AI models get larger and applications require more computing power and memory to handle increasingly complex workloads, Chanana added.
This shift is reshaping markets far beyond Silicon Valley.
The newly launched Round Hill Memory ETF trades under the ticker DRAM and has quickly become one of the most popular AI trades on Wall Street, nearly doubling since its April 2 launch.
Companies such as Micron, Samsung Electronics, SanDisk, and SK Hynix have invested in the fund.
A surge in memory chip stocks has also fueled a broad-based rally in South Korea, with the country’s benchmark stock index, the Kospi, hitting a record high this month. The index is up more than 80% this year.
The market is so hot that Goldman Sachs recently called South Korea its “most confident view” in Asia, citing the country’s dominance in memory chips and surging demand for AI infrastructure.