SK hynix (SKHY) — Q2 2026 Earnings Call Transcript

Date: July 29, 2026 (KST) | Source: Investing.com earnings call transcript (English/Korean, interpreted)

---

Park Seong-hwan (Head of IR): Good morning. Thank you for your participation in the SK hynix earnings release conference call. Today, we will begin with SK hynix's presentation and move on to a Q&A session. If you have any questions, please press star one on your phone. Note that presentations will be interpreted simultaneously, and the Q&A session will be consecutively interpreted. With that, we will now begin SK hynix earnings release conference call for the second quarter of 2026. We are joined by President of Corporate Center, Song Hyeon-jong; CFO, Kim Woo-hyun; Head of DRAM Marketing, Park Joon-deok; Head of NAND Marketing, Song Chang-seop; and Head of HBM Sales and Marketing, Kim Ki-tae. Let me issue a disclaimer that our second quarter results included in this conference call are consolidated figures and provisional in nature, as the external auditor's review has not yet been completed. Accordingly, they remain subject to change. In addition, forward-looking statements, including market outlook and the company's plans, may vary depending on changes in macroeconomic and market circumstances. With that, we will now begin. President Song Hyeon-jong will first present the earnings, followed by the company's future plans and market outlook, and a Q&A session with the attending executives. Good morning, everyone.

Song Hyeon-jong (President of Corporate Center): This is Song Hyeon-jong, President of Corporate Center. Allow me to first introduce SK hynix's performance for the second quarter of 2026. In the second quarter, strong demand driven by the expansion of AI infrastructure investment and a tight supply environment continued, leading to a sustained upward trend in prices. Both DRAM and NAND recorded meaningful price increases following the previous quarter, with AI-related products such as server DRAM and enterprise SSDs driving this growth. As a result, second quarter revenue increased by 51% quarter-on-quarter and 257% year-on-year, reaching KRW 79.3 trillion, marking a record-high revenue following the previous quarter.

In DRAM, amidst limited supply capabilities, we expanded sales centered on HBM3E and AI server DRAM products, achieving high single-digit % increase in bit shipments in line with our guidance. In particular, sales of LPDDR products for servers, including SO-DIMM 2, grew significantly. [ASP rose by approximately 30%, driven by the continued price strength in conventional DRAM.] In NAND, from a low base of reduced shipments in the first quarter and with the expansion of enterprise SSD sales, we recorded a mid-teen % increase in bit shipments in line with our guidance. Our enterprise SSD revenue increased twofold from that of the previous quarter, and Solidigm's revenue from high-capacity enterprise SSDs of 30 terabyte and above also expanded more than threefold from that of last quarter. ASP increased by mid-50% due to strong pricing across all products.

With price increases across both DRAM and NAND segments and improvement in stock cost structure, second quarter operating income reached KRW 60.5 trillion, up 61% quarter-on-quarter and 557% year-on-year. Operating margin also improved by 5 percentage points from that of the previous quarter to a record 76%, achieving all-time highs for both operating income and operating margin. Second quarter depreciation and amortization amounted to KRW 4 trillion. EBITDA stood at KRW 64.6 trillion with an EBITDA margin of 81%. Net non-operating profit reached KRW 62.2 trillion, including foreign exchange-related net gains of KRW 1.1 trillion due to the rising exchange rate and gains from the sale and valuation of investment assets totaling KRW 63.3 trillion. Consequently, pre-tax profit was KRW 122.7 trillion, net profit was KRW 93.9 trillion, and the net profit margin was 118%.

As of the end of the second quarter, our cash and cash equivalents, including short-term investments, stood at KRW 88 trillion, an increase of KRW 33.6 trillion from the end of the previous quarter. Interest-bearing debt decreased by KRW 0.7 trillion to KRW 18.6 trillion. Accordingly, net cash expanded to KRW 69.4 trillion, and the debt-to-equity ratio improved by 5 percentage points from the previous quarter end to 7%.

Now let me share our market outlook. AI technology is evolving into an agentic form that performs complex tasks on behalf of users for extended periods. As AI spreads across various services such as search, coding, and productivity tools, the scope of demand is widening from a memory perspective. Alongside high-performance memory like HBM, which is necessary for improving AI server performance and expanding system scale, demand for server DRAM to support agent services is also increasing. Additionally, the role of high-performance enterprise SSDs is widening for more efficient processing of continuously generated AI outputs. We are witnessing a structural shift in demand where both AI memory and conventional memory are growing together.

As AI models improve and software optimization advances, the computational volume and cost per individual task are continuously decreasing. We anticipate that these improvements in efficiency will not dampen overall infrastructure demand. Rather, they will lower price and adoption barriers for AI services, thereby expanding the user base and scope of applications. Major big tech customers are expanding infrastructure investments due to increased AI service usage and shortage of computing capacity. Based on the growth in revenue and profits generated from AI services, they appear to be continuing to expand their memory procurement. In fact, our major customers are still requesting more memory supply. In PC and mobile applications, temporary sales adjustments are occurring due to difficulties in securing memory. We expect these segments to gradually regain growth momentum as supply shortage eases and AI services become more widely adopted.

Amidst constrained supply conditions, DRAM and NAND demand are projected to grow by mid-20% and high-teen % respectively. Should supply constraints ease going forward, the market's growth trajectory could expand further as latent demand is met. On the supply side, however, it appears difficult for the supply-demand balance to improve meaningfully in the near term. This is due to the increasing complexity of advanced processes applied to HBM and AI server memory, as well as lead times required for constructing new production facilities.

With tight supply-demand conditions expected to persist for a considerable period, discussions regarding multi-year contracts to secure mid- to long-term supply stability with customers are ongoing. To date, we have concluded LTA negotiations with around 10 customers, including our key customers, and are continuing further discussions with our major industry players. These LTAs represent strategic partnerships that go beyond simple volume supply. They are designed to secure mid- to long-term supply stability and facilitate the development of next-generation memory aligned with our customers' technology roadmaps. While specific pricing structures vary depending on the customer and product characteristics, they are designed to address price volatility. Furthermore, financial mechanisms such as deposits are incorporated to support contract fulfillment and enhance the visibility and reliability of customers' mid- to long-term demand plans. Building on this, we will enhance the efficiency of our investments and production operations, thereby strengthening the foundation for mid- to long-term business stability and sustainable growth.

Next, I will discuss the company's plans. For the third quarter, DRAM shipments are expected to increase by approximately 10% from Q2 as we actively respond to demand with a focus on server products. For NAND, we plan a bit shipment increase of low single-digit percent quarter-on-quarter. As AI models become more sophisticated, the performance levels required for memory are rising even further, and the scope of competitiveness is expanding beyond the design of individual memory products to include system architecture and packaging technologies. Leveraging our competitive portfolio of DRAM and NAND, including HBM, and our co-development capabilities with customers, we will lead memory innovation from a system level.

First, regarding HBM4, through continuous product optimization, we have demonstrated differentiated technological competitiveness by achieving the data processing speeds required by customers, while attaining industry-leading power efficiency and cost competitiveness. We began mass production shipments in Q2 and plan full-fledged ramp-up of production in the second half of the year. For HBM4E, we supplied samples to major customers in the first half of the year. HBM4E is produced with the optimal tech node that is mature and has proven mass production stability, and we expect the subsequent development schedule to proceed smoothly. Based on our comprehensive competitiveness, which includes stable supply capabilities and cost competitiveness backed by superior quality and high yields, as well as industry-leading performance, we will continue to maintain our HBM leadership.

Additionally, for DRAM, we have fully commenced the supply of SOCAMM2 products based on the 1z-nanometer process in the second quarter. Moving forward, we will optimize our product lineup in alignment with customer development schedules and prepare for sample shipments to expand our customer base. For NAND, we will accelerate the transition to advanced nodes and strengthen our portfolio with a focus on high-capacity, high-performance products to meet market demand. In the previous quarter, our 321-layer product accounted for the largest proportion of NAND production, and we plan to expand the share within our domestic capacity to the 50% level by the end of the year as planned.

In a market situation where supply-demand imbalances persist, stable supply capability — the ability to deliver the volumes customers want in a timely manner — is emerging as a core business competitiveness alongside technological proficiency. Therefore, to respond to robust customer demand and mid- to long-term growth opportunities, we are continuing investment plans to expand our production capacity. In the short term, to enhance our supply responsiveness, we are pulling forward the mass production schedule for M15X and proceeding with investments to rapidly expand production capacity following the clean room opening of the Yongin Phase 1 in early 2027. Due to the schedule acceleration and investment expansion, our 2026 CapEx is expected to reach the high KRW 40 trillion range.

In the mid- to long-term, based on discussions with customers and market demand forecasts, we plan to proactively secure infrastructure for future supply capacity. Recently, we announced new investment plans for P&T7 to strengthen advanced packaging capabilities and M17, a new NAND production base. Furthermore, we have announced a mid- to long-term plan to create a new domestic semiconductor cluster to prepare for long-term demand beyond Yongin. Going forward, actual fab construction, equipment installation, and production capacity expansion will be pursued in stages comprehensively, considering customer demand visibility, investment efficiency, et cetera. We are planning to prepare for mid- to long-term growth opportunities without delay while maintaining CapEx discipline, thereby strengthening both our supply responsiveness and financial soundness.

Next, I'd like to speak about the ADR issuance. On July 10th, we successfully listed our ADRs on the Nasdaq market in the United States. This ADR offering was the largest ever for a foreign company's IPO in the U.S. This listing holds significance not only from the funding perspective, but also more so as a confirmation of the global market trust in our technological competitiveness and growth potential while broadening our connection points with the next-generation computing ecosystem. Building on this, we will strengthen strategic cooperation with major customers and partners and explore new business opportunities.

Finally, I will address our goals for financial soundness and shareholder returns. Driven by profits and cash generation capabilities expanded to record levels, our financial capacity has been further strengthened. Meanwhile, as structural growth opportunities in the AI era expand, the scale of investment required to realize these opportunities is also increasing significantly compared to the past. In this environment, we prioritize investments in growth opportunities that can generate high profitability and strategic value. At the same time, we aim to secure a financial structure capable of ensuring stable business operations even amidst market fluctuations, and through this, we intend to continuously share the resulting outcomes with our shareholders. While investment requirements are expected to increase going forward, we believe that our significantly strengthened cash generation capabilities will allow us to meaningfully expand shareholder returns, all while achieving our investment goals for future growth and maintaining our financial soundness targets. We're currently reviewing various additional execution measures for shareholder returns from multiple angles.

---

Q&A Session (key excerpts, as interpreted)

Park Seong-hwan (Head of IR): The first question will be provided by Jay Kwon from JPMorgan.

Jay Kwon (JPMorgan): Recently we see that some big tech companies are considering leasing data centers and more efficient AI models emerging. As a result, there are some concerns that AI infrastructure investment could slow or even decline. Based on the company's talks with customers, how do you see some of the major CSPs' AI infrastructure investment evolving? Also, could you also explain the implications for demand for HBM, DRAM, and NAND?

Park Seong-hwan (Head of IR): Thank you for the question. We are well aware of the concerns that investment in AI infrastructure may slow down following news reports of some big tech companies exploring data center leasing and introduction of more efficient AI models. We view these developments not as a sign of AI investment slowdown, but more as a transition towards higher utilization of the AI infrastructure that has already been built at scale, as well as accelerated monetization efforts. Given that for major CSPs, AI competitiveness is closely tied to their core competitiveness, including search, advertising, cloud services, and software, we believe investments aimed at strengthening AI capabilities are likely to remain solid.

Likewise, we do not see these more efficient AI models reducing infrastructure demand. Rather, we believe that they will further broaden the AI base, like accessibility and adoption, because as models and systems become more efficient, the same infrastructure can support more users and services. It is evidenced by the explosive demand for the recent high-efficiency AI models, which suggest that higher efficiency is driving broader AI adoption and usage rather than reducing demand for infrastructure. This view is also supported by the medium- to long-term demand outlook we have discussed with our key customers.

Song Hyeon-jong (President of Corporate Center): We expect the CSPs' AI-related investment to continue over the medium to long term, and the kind of memory demand being discussed with our customers reflects this trend. Of course, the timing of individual projects may differ due to physical constraints, such as power availability and data center construction. We do believe AI infrastructure investment will remain solid beyond next year, supported by ongoing AI competition among CSPs and continued expansion of AI services. We expect memory demand in general to keep expanding, not only for HBM for AI compute, but also for server DRAM to support agentic AI and high-performance, high-capacity NAND to accommodate the expansion of AI services as well as growth in data.

Moderator: The following question will be presented by Rokho Kim from Hana Securities.

Rokho Kim (Hana Securities): The company recently presented plans to significantly expand its capacity over the medium to long term. What is the basis for your long-term memory demand outlook that supports this strategy? Does it include demand secured through the long-term agreements? Given the increase in the capacity, there are understandably some concerns in the market about potential oversupply. What is the company's view regarding such concerns?

Song Hyeon-jong (President of Corporate Center): Thank you very much for the question. We plan for our medium- to long-term capacity strategy based on the structural growth in memory demand driven by AI expansion, as well as our ongoing discussions with key customers on their longer-term demand. Recently, our collaboration with customers is evolving beyond transactional relations into more strategic long-term partnerships. The stronger intent by the customers to reach long-term agreements with suppliers, as well as build partnerships, also serves as evidence of the sustained demand coming from the AI ecosystem. The capacity expansion that SK hynix is currently planning for is based on the visibility into market demand that has been secured in our partnerships with the customers. Of course, the actual capital investment and production ramp-up will be implemented in phases while considering demand visibility, investment efficiency, as well as other factors. Given that our capacity expansion will be executed flexibly in alignment with confirmed customer demand, we do not believe our medium- to long-term investment plans will lead to oversupply right away.

Moderator: The following question will be presented by Sunwoo Kim from Meritz Securities.

Sunwoo Kim (Meritz Securities): My questions are on LTA. The company's memory peers have recently concluded and announced LTAs. While it was briefly addressed in the company's briefing, could you also provide more details on SK hynix's LTA framework, such as contract term and pricing structure?

Song Hyeon-jong (President of Corporate Center): Thank you for the question. The LTAs we are discussing with our customers are designed in various forms to be more specific to each customer and their products. While the contract term normally is around five years, specific conditions may vary depending on the customer and product. Our pricing structure will also not be uniform. We are discussing with customers to adopt a range of pricing mechanisms that can better respond to price volatility. The objective is to reduce uncertainty arising from short-term market volatility while enhancing long-term business stability for both our customers and SK hynix. At the same time, it is equally important to secure effective purchase commitment, given the impact that demand volatility can have on the memory cycle. In addition to long-term volume commitments, the agreements include mechanisms such as deposits that can strengthen contract implementation and demand visibility. While we cannot say how much of our total sales will be covered by LTAs, we intend to maintain it at an appropriate level based on market conditions and customer demand. This approach should enhance the downside resilience of our earnings while maintaining the flexibility to capture incremental demand and growth opportunities when market conditions become more favorable. We have already built a solid base of profitability centered on HBM, supported by long-term collaborations with major AI customers like NVIDIA. Looking ahead, we will continue to strengthen our HBM leadership as we try to balance stability and profitability based on the demand visibility and operational flexibility secured through our LTAs.

Moderator: The following question will be presented by S. K. Kim from Daiwa Capital Markets.

S. K. Kim (Daiwa): Thank you for taking my questions; congratulations on the good performance. My questions are on DRAM. It appears as if DRAM ASP growth in the second quarter fell below market expectations. What are the reasons, and what is the outlook for the second half of the year?

Interpreter (management): We manage the sales mix between HBM and conventional DRAM based on customer demand and some medium- to long-term product strategy. In the second quarter, shipments of some high-value-add products were pushed back into the second half, and changes in the product portfolio appeared to have affected our blended ASP. These factors are likely to gradually ease in the second half. As HBM4 shipments ramp up in earnest and the 1c-nanometer conventional DRAM shipment increases, we expect bit growth in the second half to be higher than the level in the first half. In addition, considering customer demand and the changing product mix, the growing HBM4 sales and the higher contribution from value-add products will have a positive impact on our blended ASP as well. This will result in higher shipment volumes and continued improvement in the product mix, which will push up our ASP as well as earnings in the second half. When running our sales strategy, instead of focusing on short-term price movements or profitability, we consider a comprehensive set of factors including demand visibility, long-term customer relationships, and supply-demand dynamics across individual product segments.

Moderator: The following question will be presented by Dong Han from SK Securities.

Dong Han (SK Securities): Thank you. My questions are on HBM. Now, some believe that competitors have made rapid progress in HBM recently. What is the competitiveness of our HBM4 and the key differentiators that will enable the company to maintain its leadership in the HBM market?

Interpreter (management): The HBM4 competitiveness is completed not only by delivering the required performance, but also by the capability to supply at scale with stable yield and consistent quality. SK hynix has consistently demonstrated these capabilities since the HBM2E generation. Our accumulated competitiveness in time to market, product performance, mass production yield, quality, and customer trust are the differentiators that cannot be replicated in a short period of time. Building on this foundation, we began mass production of HBM4 for key customers in the second quarter, with the yield and quality today nearing the levels of HBM3E, which is already in the maturity stage. Our current focus is on steadily ramping up production capacity. As mentioned earlier, we have also completed HBM4E sample delivery to customers. We have applied an optimized manufacturing process with proven technology maturity and production stability. Development is underway smoothly in line with our road map, aiming at volume production beginning in 2027. Not stopping there, we are also proactively preparing next-generation technologies. In addition to hybrid bonding, we are developing IHBM technology to effectively provide thermal dissipation in future products such as HBM5. The IHBM integrates cooling elements within the package and is expected to reduce thermal resistance by more than 30%, improving system stability and operational efficiency in high-performance, high-density AI environments.

---

*Source: Investing.com, SK hynix (SKHY) Q2 2026 Earnings Call Transcript, Jul 28–29, 2026 (English/Korean, interpreted). Financial results are provisional/preliminary (external auditor review pending).*