SK Hynix, $29 Billion Bid to Win Over Wall Street

The transaction and its industrial rationale

SK hynix brings its most important industrial cycle to Wall Street. The company aims to raise up to $29.4 billion, defend its leadership in HBM and secure in the United States a valuation more closely aligned with the role it has assumed in the artificial intelligence supply chain.

Editor's note. HBM stands for High Bandwidth Memory: high-bandwidth memory used to power GPUs and AI accelerators with greater speed and fewer bottlenecks.

Operational read. The ADR programme can lower the cost of capital and finance the facilities needed to stay ahead of Samsung and Micron. Maximum dilution remains contained at 2.50%, but the risk is clear: capacity decided during a period of scarcity will come to market at a time when AI demand must still be strong enough to absorb it.

Key data

1. Executive summary

SK hynix is preparing an American Depositary Receipt offering on the Nasdaq, backed by a maximum of 17.79 million new shares. The potential raise amounts to approximately $29.4 billion, placing the transaction among the most significant of the current AI cycle.

This is not the company's stock market debut. Shares are already listed in Seoul and depositary receipts are traded in Luxembourg. The move to the US market is intended to raise fresh capital and bring SK hynix into the market that currently assigns the highest multiples to artificial intelligence infrastructure.

The timing is favourable. The company holds a 58% share of the HBM market, has surpassed Samsung in market capitalisation in Korea and closed the first quarter of 2026 with approximately $34.0 billion in revenues and approximately $24.4 billion in operating profit. Memory, for years regarded as a cyclical commodity, has become one of the bottlenecks in computing capacity.

The read remains twofold: dilution is contained, but the industrial plan requires years of execution. If AI demand remains strong, the capital raised may defend its leadership position; if the cycle slows, the new capacity could weigh on memory prices.

2. This is not a traditional IPO

Describing the transaction as an IPO is convenient but imprecise. SK hynix already has a market price, a consolidated shareholder base and a listed history. The Nasdaq listing adds a new ADR line and a second price discovery centre.

As of 31 March, the company had 712.7 million ordinary shares outstanding. The new shares expected represent approximately 2.50% of pre-offering share capital. SK Square's stake would fall to close to 20%, without altering the control structure.

Transaction structure

The new ADR line complements the Korean market with a channel more accessible to investors evaluating Nvidia, Micron and the rest of the AI supply chain. Source: SK hynix Investor Relations and ADR transaction documentation.

3. Nasdaq as a second reference market

The US listing opens SK hynix to a deeper pool of capital that is more accustomed to pricing growth, technological leadership and supply-chain positioning within AI. For many American investors, purchasing securities traded directly in Seoul is less straightforward; the ADR reduces this barrier.

The most important variable remains the Korea discount. If the American market awards SK hynix a premium closer to that of its technology peers, the Seoul-listed shares could also benefit. If, on the other hand, the ADR trades at a discount, the market will be signalling that HBM leadership alone is insufficient to offset the cyclicality of the business.

The Nasdaq transaction

SK hynix plans to issue up to 17.79 million new shares and raise an indicative maximum of $29.43 billion. Each ordinary share will be represented by ten ADRs, and the final price will be set following the book-building process. Source: Reuters, 24 June 2026.

In the following chart we observe how the market has already recognised leadership in AI memory

4. HBM has transformed SK hynix's competitive position

For decades, memory has been treated as a technology commodity. HBM has altered this equilibrium by stacking multiple DRAM layers, increasing bandwidth and reducing the bottleneck between memory and AI accelerators.

SK hynix invested before the technology became central. That head start translated into operational relationships with Nvidia, packaging capabilities, production-yield expertise and faster qualification timelines for new generations. In the first quarter of 2026, the company still controlled 58% of the HBM market.

Leadership remains clear-cut, but it is no longer unassailable. A year earlier, the market share stood at 69%; Samsung is accelerating on HBM4, Micron is expanding capacity, and customers are seeking to limit dependence on a single supplier. Maintaining primacy therefore requires continuous investment.

The multi-year partnership with Nvidia provides visibility on the Vera Rubin roadmap and strengthens the integration between GPUs, packaging and memory. At the same time, it increases concentration: a significant portion of SK hynix's value depends on the continuity of capital expenditure by major hyperscalers and on Nvidia's market leadership.

In the following chart we observe SK hynix's HBM market share relative to Samsung and Micron.

In the following chart we observe the evolution of SK hynix's share of the HBM market

HBM growth is occurring within a semiconductor market that is still expanding. The most important factor is not solely the size of the sector, but the rising weight of high-bandwidth memory in AI spending.

In the following charts we observe the expected growth of the global semiconductor market and the projected HBM market in 2026.

In the following chart we observe the growth of the global semiconductor market

In the following chart we observe the projected growth of the HBM market in 2026

The transition from HBM3E to HBM4 will be the first true test of competitive advantage. SK hynix holds the lead, but will need to defend it generation after generation.

5. Twenty-nine billion is industrial capital

The $29.4 billion finances an industrial cycle comprising fabs, clean rooms, electrical installations, water treatment facilities, EUV scanners and advanced packaging.

The first fab in the Yongin cluster and the initial infrastructure require approximately $6.1 billion. The project encompasses four fabs and a supplier ecosystem. In Indiana, SK hynix will invest a further $3.87 billion in advanced packaging and research, with production scheduled from the second half of 2028.

The equity route takes advantage of a favourable window: issuing a limited tranche of capital at a high valuation reduces the need to increase debt before the facilities begin production. For shareholders, the key question will be the return on the new capital.

The following chart shows the investment breakdown for the first Yongin fab

6. The US capital raise also reshapes the competitive landscape

The first effect concerns Micron. For many US investors, it was the most direct way to gain exposure to memory and HBM growth. SK hynix's arrival on Nasdaq introduces an alternative with a higher HBM share and a deeper relationship with Nvidia.

Capital allocated to SK hynix will not necessarily flow out of Micron, as the entire sector can continue to attract new resources. However, the uniqueness premium of the American manufacturer is reduced, while Samsung will need to respond on both the product and capital fronts.

7. The less visible side: governance and dilution

The offering drew criticism in South Korea, as some investors would have preferred less dilutive structures. The assessment of the transaction will ultimately depend on the deployment of proceeds: the capital will create value if it funds facilities generating returns above the cost of capital.

Maximum dilution remains around 2.5% of pre-offering capital. This is a manageable level if the new capacity defends margins and market share; it becomes more burdensome if the facilities come online during a period of weaker demand. Source: SK hynix Investor Relations; calculations based on 712,702,365 existing shares and 17.79 million new shares.

8. Memory remains a cyclical sector

The 2026 results may suggest that the memory cycle has been overcome. In reality, it has become more complex. Producers are few, the technology is more demanding and HBM absorbs more capacity than conventional DRAM, but cyclicality has not disappeared.

The risk is that capacity built for HBM is repurposed towards conventional DRAM should AI demand slow. As long as HBM absorbs wafers, it supports prices; if demand normalises, a supply glut may re-emerge.

SK hynix enters from a stronger position than in previous cycles, thanks to its technology leadership and customer relationships. This protects market share and profitability, but does not eliminate the risk of lower prices and underutilised facilities.

9. Scenario map

The base case remains constructive, but hinges on the duration of AI demand and the ability to qualify HBM4 within the expected timeframe.

10. Investment view

The most compelling thesis rests on industrial advantage, not merely on index inclusion. Passive flows may amplify the initial move, but it will be HBM4, manufacturing yields and free cash flow that determine whether the US premium can be sustained.

11. What determines the validity of the thesis

The first test will be the offering price. A raise close to the top of the indicated range would confirm the depth of US demand; a heavily discounted placement would instead signal greater caution on valuation.

The second test will be industrial. SK hynix will need to demonstrate how the proceeds are to be deployed across Yongin, existing capacity, equipment, packaging and the United States. Capex growth will need to be measured against capacity that is actually qualified, not merely against floor space built.

The third test will be a market one: the relationship between the ADR and the Korean share, US trading volumes, fungibility and demand composition. A liquid ADR can become a global benchmark; a thinly traded float would remain a secondary instrument.

12. Operational conclusion

SK hynix's US listing is one of the most significant transactions to emerge from the artificial intelligence boom. The company raising capital is already dominant in the component that currently constrains AI accelerator capacity.

The Nasdaq provides the resources to translate technological leadership into productive capacity and brings SK hynix closer to the investors who follow Nvidia and the broader supply chain. Dilution is limited, but the plan requires years of execution and tens of billions of dollars in investment.

The industrial quality is evident: SK Hynix developed HBM ahead of its competitors, tied its roadmap to Nvidia, and brought a niche technology to the group's economic core. The risk remains that inherent to memory: investments are decided during periods of scarcity and completed when the market may have already shifted.

The operational reading is therefore constructive, but disciplined. The stock can benefit from the American premium and HBM leadership, provided AI demand remains strong enough to absorb the incoming capacity.

Main sources consulted

• SK Hynix Investor Relations and Q1 2026 results.
• SK Hynix Newsroom, updates on Yongin, Indiana and HBM outlook.
• Reuters, announcement of the ADR transaction dated 24 June 2026.
• Counterpoint Research, DRAM and HBM market share data.
• Micron Technology, SEC filings on capacity risks and the memory cycle.