From “the most hopeless company in South Korea” to the king of trillion-scale storage, what exactly has SK hynix done right?

Tonight, SK hynix ADR will officially list on the US stock market Nasdaq.

This company, now at the center of the AI memory-chip boom, saw its market cap reach $1 trillion in May this year, becoming South Korea’s second-largest listed company by market value, behind Samsung Electronics. Since the beginning of 2025, its stock price has risen by about 13 times in total.

But if you rewind 30 years, SK hynix was once widely regarded in Korean corporate circles as a “problem company”: high debt, consecutive losses, reliance on creditor rescues, and even at one point preparations to sell to the US storage giant Micron.

From facing possible exit to becoming an important supplier of Nvidia’s high-bandwidth memory chips, SK hynix’s comeback—on the surface—hits the AI wave; in reality, it is the accumulation of several key decisions.

“Downsizing” during the crisis, defending the core memory business

The predecessor of SK hynix was Hyundai Electronics, founded in 1983.

At the time, South Korea was heavily developing strategic industries such as semiconductors. Hyundai Electronics started with DRAM—dynamic random-access memory chips—and in the 1990s became one of the world’s leading DRAM suppliers.

Problems soon followed.

Aggressive capacity expansion required huge funding. After the Asian financial crisis broke out in 1997 and 1998, memory chip prices fell, and debt pressure was quickly exposed.

After that, the company underwent restructuring and changed its name to Hynix Semiconductor in 2001.

At its toughest point, although the company’s annual sales reached 4 trillion won, losses exceeded 5 trillion won, and liabilities totaled as much as 7 trillion won.

Its stock price once dropped to 125 won per share, which is less than 60% of one yuan RMB, and it was dubbed “South Korea’s penny stock.”

In 2002, Hynix had planned to sell for about $4 billion to Micron, but the deal ultimately did not go through. After that, the company entered a reorganization phase, cutting costs on a large scale, selling assets, spinning off businesses, and laying off staff to survive.

The price was high, but the outcome was clear: Hynix would concentrate resources on its most core memory business.

Businesses such as display, packaging, and non-memory chips were spun off or sold. What remained was a more focused, leaner, and more cycle-resilient memory chip business.

In the semiconductor industry, doing everything doesn’t necessarily mean you can build strength in everything. SK hynix’s later rise first came from the fact that, during the most difficult period, it did not lose its technical foundation.

Bringing in SK Group to solve the problem of “no money to expand”

Memory chips are an extremely brutal industry.

For a long time, memory has not only been highly cyclical; building plants, buying equipment, and developing advanced process nodes all require continuous massive capital investment. When the industry is booming, companies compete to expand capacity. When the industry is sluggish, price declines can quickly wipe out profits. Many competitors exited the market amid this kind of repeated volatility.

Today, the global memory market is mainly dominated by three companies: Samsung Electronics, SK hynix, and Micron.

Although Hynix regained its momentum after reorganization, it still faced a reality: relying solely on creditors and bank shareholders makes it hard to sustain a long-term capital race against Samsung.

The real turning point came in 2012. SK Group of South Korea gained control from creditors, and the company officially changed its name to SK hynix.

SK Group’s entry brought capital and credit support. The company received fresh equity injections and began increasing capital expenditures: in just 2012 and 2013, it invested 4 trillion won and 2 trillion won, respectively.

This money was not used for blind expansion. Instead, it was continuously invested in memory manufacturing capacity and advanced technology.

Looking back later, SK Group’s acquisition not only filled SK hynix with capital, but also provided the long-term investment capability needed to ride out memory cycles.

When HBM was still not widely favored, it didn’t give up

What truly put SK hynix at the core position of the AI era was HBM.

In 2013, SK hynix partnered with AMD to launch the world’s first high-bandwidth memory chip, HBM. Its core idea was to vertically stack multiple layers of DRAM to achieve higher data transfer speeds at lower power consumption.

Looking at it today, it almost seems tailor-made for AI computing workloads.

But back then, HBM was not a widely favored track. The market performance of products using HBM in the early days was not ideal, and Samsung later also took the lead in HBM. Even in 2018, when Samsung cut its HBM team, many people in the industry judged that this technology might not have a big future.

SK hynix did not follow suit and give up.

HBM consumes more wafers than standard DRAM and is also more difficult to manufacture, involving multiple issues such as stacking precision, heat dissipation, and yield. But SK hynix kept refining its processes and improving its stacking and packaging capabilities.

This is a classic kind of investment where there is no visible payoff in the early stage, but later determines one’s ranking.

Until late 2022, when ChatGPT was released, demand for storage bandwidth for AI model training and inference suddenly surged. AI accelerators needed large quantities of HBM, and the market quickly saw a supply bottleneck.

What had seemed like “extra” investment suddenly became a first-mover advantage that is hard to replicate.

Today, SK hynix holds about 51% of the global HBM market, ahead of Samsung Electronics at about 26% and Micron at about 23%. It is also an important supplier of Nvidia, which is a dominant player in the AI accelerator space.

AI changed the valuation logic of the memory industry, and it also validated SK hynix’s坚持 on HBM over the years.

When demand is at its strongest, continue betting on capacity and global capital

Becoming a leader in HBM doesn’t mean the competition is over.

Samsung, SK hynix, and Micron are all accelerating capacity expansion and competing for the market of the next-generation HBM4. HBM4 will be used in Nvidia’s next-generation Vera Rubin accelerator. The related products from all three manufacturers have entered production, and more deliveries are expected in the second half of 2026.

Facing competition, SK hynix has chosen to keep investing.

In South Korea, the company plans a large-scale capacity expansion. The wafer-fab cluster in Yongin alone will invest $10k, and the four factories are planned to be completed by 2033. At the same time, the company is also speeding up procurement of key equipment such as extreme ultraviolet (EUV) lithography machines.

In the US, SK hynix is building its first production facility in Indiana, expected to be completed in 2028, mainly for advanced packaging—an important link in HBM production for connecting and stacking chips.

Listing on Nasdaq tonight is also part of its globalized capital deployment.

In this ADR issuance, each ADR corresponds to one-tenth of a share of common stock. SK hynix sold 177.9 million ADRs at a price of $149 per ADR, raising $26.5 billion, with a scale of about 2.5% of the company’s market value. The raised funds will be used to build new factories and purchase advanced manufacturing equipment, including EUV lithography machines needed to produce leading-edge chips.

From relying on creditor rescues to raising funds in the US to expand capacity, SK hynix’s changing identity is already enough to show the point.

Conclusion: not just accidentally catching AI, but being prepared before AI arrived

SK hynix’s story is not simply a “AI concept stock boom” story.

It struggled at the bottom of the memory cycle, and it was also seen as a company that should be sold or liquidated. It survived by focusing on its core business, regained the capability to invest through the capital support from SK Group, and—most importantly—by persisting with HBM technology for years, it occupied the most critical position when AI demand surged.

Of course, the memory industry cycle has not disappeared. Historically, every wave of demand mania can lead to subsequent supply overhang and price volatility. Samsung’s and Micron’s catch-up efforts, along with the competition for HBM4, also mean SK hynix cannot afford to be complacent.

But at least so far, SK hynix has proven something:

What truly determines the fate of a technology company is often not whether it can tell a story after the windfall arrives, but whether it can last through the pre-surge period, invest, and build the technology.

Risk warning and disclaimer

        The market is risky; investment is to be done cautiously. This article does not constitute personal investment advice, nor does it consider the special investment objectives, financial conditions, or needs of individual users. Users should consider whether any opinions, viewpoints, or conclusions in this article align with their specific circumstances. Invest at your own risk based on this.