SK Hynix has drawn orders for more than seven times the amount of American Depositary Receipts it plans to sell in its U.S. market debut, signaling strong demand for one of the most closely watched semiconductor listings in years despite a sharp correction across chip stocks.
The South Korean memory-chip maker is preparing to list on Nasdaq under the ticker SKHY, with its U.S. debut expected on July 10 after filing an F-1 registration statement with U.S. regulators in late June. The planned offering involves 177.9 million ADRs and is expected to raise about $24.5 billion based on the company’s Korean share price at Wednesday’s close. Some market estimates have placed the potential proceeds closer to $28 billion depending on final pricing and the ADR structure.
If completed at the upper end of expectations, the sale could become the largest U.S. listing by a foreign company, underscoring the scale of global demand for companies tied to artificial intelligence infrastructure. SK Hynix is a leading supplier of high-bandwidth memory, or HBM, a critical component used in AI accelerators and data centers.
The strong demand comes at a difficult moment for the semiconductor sector. Chip shares have been under pressure after two years of gains fueled by heavy AI-related spending from major technology companies. Concerns have grown that the pace of data-center expansion may be slowing, especially after reports that Meta was considering selling unused computing resources and Blackstone canceled a large data-center project.
Those developments have forced traders to re-check assumptions about how long the AI infrastructure boom can continue at its current speed. SK Hynix’s Korean-listed shares have fallen from a June 25 peak of 2,917,000 won to 2,076,000 won at the previous close, a drop of nearly 30%. The broader semiconductor market has also weakened, with the Philadelphia Semiconductor Index recently suffering a steep decline that erased more than $1 trillion in market value.
Still, demand for the U.S. offering suggests many large global funds remain willing to back the company’s long-term role in AI hardware.
Demand remains strong despite the selloff
The contrast between SK Hynix’s falling share price in South Korea and heavy demand for its U.S. ADRs has become one of the clearest signals of tension in the chip market. Short-term traders have been reducing exposure to semiconductor shares after a sharp rally, while large long-term funds appear focused on the company’s position in advanced memory chips.
Orders for the ADRs came mainly from global long-only funds, technology-focused portfolios and sovereign wealth entities. Baillie Gifford and Coatue Management were among the major names linked to the offering. A fund managed by Leopold Aschenbrenner, the former OpenAI researcher known for his writings on artificial intelligence, reportedly showed interest in buying about $7 billion worth of the ADRs.
Roughly 1,000 global institutions are said to have taken part in the order book. That level of participation is notable because it comes during a market pullback, not during a broad rally. For SK Hynix, the strong subscription gives the company flexibility as it seeks to raise capital for an aggressive production buildout.
The deal also gives U.S.-based traders direct access to one of the most important suppliers in the AI chip supply chain. While Nvidia often attracts the most attention because of its graphics processing units, HBM suppliers such as SK Hynix are essential to the performance of AI accelerators. Advanced AI chips require large amounts of fast memory placed close to the processor, and HBM has become one of the most important bottlenecks in the sector.
Proceeds to support Korean production expansion
SK Hynix plans to use the proceeds from the ADR sale to expand production capacity in South Korea. The company has pointed to three main areas of spending: its Yongin wafer facility, advanced packaging lines in Cheongju and equipment related to extreme ultraviolet lithography, commonly known as EUV.
The Yongin project is central to South Korea’s plan to strengthen its domestic semiconductor supply chain. The site is expected to support next-generation memory production and help SK Hynix meet rising demand from global technology companies building AI data centers.
The Cheongju expansion is also important because advanced packaging has become a key part of the chip industry’s growth strategy. As chips become more complex, performance gains increasingly depend not only on the chips themselves but also on how they are connected, stacked and packaged. HBM production requires advanced packaging techniques because multiple memory layers must be stacked and linked with high precision.
EUV tools are another major cost area. These machines are among the most expensive pieces of equipment used in semiconductor manufacturing. They are required for the most advanced chipmaking processes and are essential for companies trying to maintain leadership in high-performance memory.
By raising capital in the United States, SK Hynix is seeking to fund this expansion while also broadening its global shareholder base. A Nasdaq listing could improve visibility among U.S. technology-focused traders and funds at a time when AI supply chains are becoming a central theme across global markets.
AI spending concerns weigh on chip shares
The offering arrives after a period of growing caution around AI-related capital expenditure. Major technology companies have spent heavily on data centers, networking equipment and AI processors over the past two years. That spending has supported revenue growth across the semiconductor industry, especially for companies linked to AI computing.
But markets have started to question whether the pace can continue. Reports that Meta may sell unused computing capacity raised concerns that some AI infrastructure may have been built faster than demand required. Blackstone’s decision to cancel a large data-center project added to worries that real estate, power and financing constraints could slow further expansion.
The issue is not whether AI demand exists. Few market participants doubt that artificial intelligence will require far more computing capacity over time. The question is whether the current level of spending is sustainable quarter after quarter.
The result has been a sharp revaluation of chip shares. Traders who bought into the AI theme early have locked in gains, while others have stepped aside to wait for clearer signals from technology companies. SK Hynix has not been immune, even though demand for its HBM products remains strong.
Its nearly 30% slide from late June highs reflects that broader concern. When one of the strongest AI-linked hardware names can fall that quickly, it shows how sensitive the market has become to any sign of slower data-center spending.
Big technology companies remain central to the outlook
The next major test for the sector will come from earnings reports and spending guidance from Microsoft, Alphabet, Meta and Amazon. These companies are among the largest buyers of AI infrastructure, and their capital expenditure plans will shape expectations for chip demand through the rest of the year and into 2026.
Current projections still point to very large spending. Combined 2026 capital expenditure for Amazon, Microsoft, Alphabet and Meta has been estimated near $725 billion. Microsoft is expected to spend about $190 billion in 2026, with finance chief Amy Hood previously pointing to rising component and memory prices as a major factor behind higher costs. About $25 billion of that increase has been linked specifically to components and memory.
Amazon has projected about $200 billion in capital expenditure for 2026, much of it tied to data centers. Alphabet has guided for spending between $175 billion and $185 billion. Meta has raised its own forecast to a range of $125 billion to $145 billion.
Those figures remain enormous by historical standards. If they hold, suppliers such as SK Hynix would likely continue to benefit from demand for advanced memory, storage and networking components. If they are reduced, the current chip-sector correction could deepen.
This is why the upcoming earnings cycle matters so much. Traders will be watching not only revenue and profit but also comments on data-center utilization, AI model training needs, power availability and supply constraints. Any sign that major technology firms are spreading out their purchases over a longer period could affect semiconductor valuations.
The long-term case for HBM remains intact
Even with recent volatility, the strategic importance of HBM has not changed. AI accelerators depend on memory bandwidth to handle large models and heavy workloads. Without enough HBM, expensive processors cannot operate at full efficiency.
SK Hynix has been one of the early leaders in supplying advanced HBM products to the AI market. That position has made it one of the main corporate beneficiaries of the boom in generative AI computing. Demand from chip designers, cloud platforms and server makers has pushed HBM from a niche product into one of the most important segments of the memory industry.
The broader semiconductor industry is still expected to grow sharply if AI spending continues. Some industry forecasts suggest global semiconductor revenue could move above $1.5 trillion in 2026, helped by demand for AI servers, advanced memory and cloud infrastructure. AI-related technology spending has also been projected to rise strongly, as companies continue building systems for training and running large models.
For SK Hynix, the question is whether it can expand capacity fast enough without oversupplying the market later. Memory has historically been a cyclical business. Periods of strong demand often lead to aggressive capacity expansion, which can later pressure prices if supply catches up too quickly.
However, HBM is more complex than standard memory. Production is harder, qualification cycles with major customers are longer, and advanced packaging capacity is limited. These factors may reduce the risk of a sudden oversupply compared with older memory categories, though they do not eliminate it.
A listing shaped by two different market views
The ADR offering highlights two competing views of the AI trade. One view focuses on the recent correction, rising spending doubts and the risk that technology companies may slow their data-center expansion. The other focuses on the long-term need for more computing power and the limited number of companies able to supply the most advanced components.
The heavy order book suggests the second view remains powerful among large funds. For many traders, the decline in SK Hynix’s Korean shares may have created a more attractive entry point ahead of the Nasdaq debut. The U.S. listing also offers a way to gain exposure to AI memory without buying shares in South Korea.
That does not remove the near-term risks. The offering will price into a nervous semiconductor market, and the first days of trading could be volatile. Demand for the ADRs has been strong, but public trading after the debut will depend on broader market sentiment, final pricing and the tone of upcoming technology earnings.
Still, the scale of interest shows that SK Hynix remains central to the AI hardware story. The company is raising capital at a time when demand for advanced memory is strong but market confidence is being tested. That makes the listing more than a single corporate event. It is also a gauge of how much faith global traders still have in the next stage of AI infrastructure growth.
For now, the message from the order book is clear: despite the selloff, demand for exposure to high-end AI memory remains deep. The next confirmation will need to come from the companies spending hundreds of billions of dollars to build the data centers that use it.
Want to explore how AI and trading intersect? Start with AI copy trading to see algorithm-powered investment strategies in action.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
