South Korea’s SK Hynix is investing $720 billion to build what it says will be the world’s largest network of memory factories, betting that soaring demand for artificial intelligence will outlast the boom-and-bust cycles that have historically defined the memory chip industry.
The world’s leading maker of high-bandwidth memory, or HBM, has seen its market value surge more than fivefold over the past year to more than $1 trillion as AI companies race to secure supplies of the chips needed to power advanced computing systems.
SK Hynix controlled 58% of the global HBM market in the first quarter, followed by South Korean rival Samsung and U.S.-based Micron, each with 21%, according to Counterpoint Research.
The company raised $26.5 billion in July through a Nasdaq listing, the largest amount ever raised by a foreign company through a U.S. stock market listing, to help fund its expansion.
Its U.S.-listed shares have since fallen about 21% from a July 14 peak of nearly $195, closing at $154.41 on Wednesday, amid broader volatility in AI-related stocks and a decline in South Korea’s stock market.
SK Hynix’s expansion forms part of a wider South Korean government plan announced by President Lee Jae Myung in June to double the country’s memory production over five years. The plan includes construction of several new Samsung megafabs.
At SK Hynix’s Yongin Cluster, south of Seoul, the first of four planned fabs is under construction. The company says the facility will reach the height of a 50-story apartment building and contain six cleanrooms spread across multiple floors.
That contrasts with newer U.S. fabs being built by Taiwan Semiconductor Manufacturing Co and Intel, which generally spread cleanroom space across a single floor.
HBM is made by stacking dynamic random access memory, or DRAM, chips to provide the high bandwidth required by AI processors. Rising demand for HBM has tightened supplies of memory for consumer electronics as major AI companies sign longer-term contracts to secure capacity.
“It’s like a war,” said Chey Tae-won, chairman of SK Group, the controlling owner of SK Hynix. “Everybody wants to buy the memory chips. Without that, they cannot produce their AI computing and AI chips.”
Tae-won says that “prices went up too fast” and that he hopes the industry’s expansion will help. “I’m trying to do my best,” he said.
SK Hynix said in July it had signed 10 long-term agreements with major customers. Nvidia has “secured a stable supply of HBM” and agreed to co-develop next-generation memory as part of a $500 billion deal with SK Group that also includes plans to build new data centres with SK Telecom by 2027.
Tae-won showed CNBC a wafer carrying a message from Nvidia CEO Jensen Huang that read, “Please make more.”
He said he occasionally meets major technology executives including Microsoft’s Satya Nadella, Google’s Sundar Pichai and Meta’s Mark Zuckerberg. Huang and AMD CEO Lisa Su have also visited South Korea in recent months for talks on memory supplies.
MS Hwang, research director at Counterpoint Research, said hotels near SK Hynix and Samsung’s memory-fab expansions were fully booked.
“If you name any company in Big Tech, they are all in Korea to sign a contract,” he said.
SK Hynix also operates three memory fabs in China but is barred by U.S. export controls from selling its most advanced HBM there. China is developing its own HBM industry, including memory maker CXMT, which recently made a strong debut on the Shanghai stock market.
“It’s a race, and now the counterparty of the race is China,” Hwang said. “That’s going to be a lot more dangerous than anything else in the world.”
South Korea is urging chipmakers to accelerate construction of new fabs as the United States also moves to expand memory production.
Micron is spending $50 billion on two new memory fabs in Boise, Idaho, with the first scheduled to begin wafer production in 2027. It is also developing a $100 billion campus in Clay, New York, with capacity for up to four fabs.
SK Hynix is building its first U.S. facility in West Lafayette, Indiana, where construction is expected to be completed in 2028. The $4 billion facility will focus on advanced packaging, where memory chips are stacked and connected, rather than front-end manufacturing.
SK Hynix also has existing U.S. operations through Solidigm, the subsidiary created after its $9 billion acquisition of Intel’s NAND business in 2020. In January, the company restructured the business and launched a $10 billion “AI Company” in the United States, but has no current plans for front-end manufacturing there.
Tae-won said SK Hynix had been studying potential locations for additional U.S. chip facilities for more than a month.
“I’m willing to do it, but the problem is that I really have to find the right place,” he said.
He said the memory industry was at a “turning point” as the development of custom HBM could help shield SK Hynix’s massive investment from the risks associated with the industry’s traditional cycles.
Custom HBM is increasingly being co-designed with AI processors, a trend that is accelerating with the development of next-generation HBM5.
“Nvidia wants their own custom chips and Google wants their own customized HBM, so it’s not just a commodity,” Tae-won said. “It actually changes the memory chip’s status.”
Stay ahead of the Stories shaping our world. Subscribe to Impact Newswire and join our
WhatsApp Channel for updates on global tech, business, and innovation—all in one place.
Dive deeper into the future with the Cause Effect 4.0 Podcast, where we explore the ideas, trends, and technologies driving the global AI conversation.
Got a story to share? Contact Us to reach a global audience with Impact Newswire.
Discover more from Impact AI News
Subscribe to get the latest posts sent to your email.
