SK hynix Aktie

Good morning, afternoon and evening. Thank you for joining SK hynix 2026 Q1 Earnings Release Conference Call. Following SK hynix presentation, there will be a Q&A session. [Operator Instructions]

Please note that presentations will be interpreted simultaneously and a Q&A session with consecutive interpretation. With that, we are now ready to begin.

[Interpreted] Good morning, afternoon and evening. This is Park Seong Hwan, Head of IR at SK hynix. Welcome to the SK hynix 2026 first quarter release -- earnings release conference call. Allow me to introduce the executives present here with me today. We are joined by CFO, Kim Woo-Hyun; Head of DRAM Marketing, Park Chan-Dong; Head of NAND Marketing, Kim Jeong Tae; and HBM Sales and Marketing Head, Kim Gi-Tae.

Let me issue a disclaimer that our first 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 SK hynix's earnings release conference call for first quarter of 2026. CFO Kim Woo-Hyun will first present the earnings followed by the company's future plans and market outlook and a Q&A session with the attending executives.

[Interpreted] Good morning, everyone. This is CFO Kim Woo-Hyun. Allow me to first introduce the company's performance for the first quarter of 2026. Although the first quarter is typically a seasonally slow period, robust demand driven by expanding AI infrastructure investments more than offset seasonality and the supply environment remains tight. As a result, prices for both DRAM and NAND rose meaningfully led in particular by server DRAM and enterprise SSDs.

Accordingly, first quarter revenue increased by 60% Q-o-Q, and 198% Y-o-Y to KRW 52.6 trillion, surpassing KRW 50 trillion mark for the first time on a quarterly basis and setting a new record high for quarterly revenue. Starting with DRAM. Shipments in Q1 were similar to Q4 levels, in line with our previous guidance. Within our supply capabilities, we focused sales on products with robust demand, including HBM and high-density server modules of 128 gigabytes and above, while ASP rose by mid-60% as the strength in conventional DRAM pricing accelerated.

For NAND, shipments declined by approximately 10% Q-o-Q, reflecting a high base from the previous quarter sales as well as reduced discrete product sales and longer production lead times associated with the shift in mix towards high value-added products, where demand is rising rapidly. On the other hand, ASP rose sharply by mid 70%, supported by strong pricing in all product lines. As rising prices across both DRAM and NAND coincided with a greater contribution from high value-added products, first quarter operating profit reached KRW 37.6 trillion, increasing nearly twofold from that of the previous quarter. Operating margin also improved by 13 percentage points Q-o-Q to 72%, marking another all-time high on a quarterly basis. In the first quarter, depreciation and amortization totaled KRW 3.7 trillion, EBITDA came in at KRW 41.3 trillion, and EBITDA margin was 79%.

Net nonoperating profit reached KRW 14 trillion, reflecting, among other factors, foreign exchange-related net gain of KRW 1.6 trillion due to the rise in exchange rates and valuation gains on investment assets of KRW 9.9 trillion. Accordingly, pretax profit was KRW 51.6 trillion. Net profit was KRW 40.3 trillion and net profit margin was 77%.

As of the end of the first quarter, cash and cash equivalents including short-term investments stood at KRW 54.3 trillion, an increase of KRW 19.4 trillion from the end of the previous quarter, while interest-bearing debt declined by KRW 2.9 trillion to KRW 19.3 trillion. As a result, company recorded net cash of KRW 35 trillion and the debt-to-equity ratio at the end of the first quarter improved by 6 percentage points from the end of the previous quarter to 12%.

Now let me share our market outlook. AI technology is now evolving rapidly beyond the training phase of large-scale models into the inference and agentic AI stage where user requests are processed in real time across a wide range of service environments. Accordingly, AI workloads are shifting from simple question-and-answer tests to more complex processes that involve planning, execution and verification repeated until the optimal result is achieved.

As the volume of data generated by AI agents at each stage continues to increase, demand is rising for a broad range of memory products to process and store such data efficiently. Beyond demand for high-performance memory like HBM, the total volume of memory required across overall systems is also increasing, including server DRAM modules and enterprise SSDs. As a result, the demand base is broadening for both DRAM and NAND.

In addition, ongoing software and hardware optimization efforts across the AI industry are serving as another driver of memory demand growth. While memory efficiency technology may appear to reduce memory usage per device, in practice, it is evolving in a way that maximizes the amount of context that is handled per unit of memory and a number of users.

This is expected to improve the economics of AI services, expand the overall market and create a virtuous cycle that further drives memory demand. Meanwhile, in the PC and mobile markets, some signs of demand softening have emerged, including adjustments in such shipments and changes in product portfolios due to cost pressures stemming from higher memory prices. However, robust server memory demand continues to offset this weakness and is driving overall market growth.

As memory becomes increasingly critical in AI computing, demand for high-performance memory is surging, while supply remains constrained. Amid the supply-demand imbalance, customers are prioritizing securing volume over pricing, which is sustaining the current pricing strength. Accordingly, we expect a favorable pricing environment to continue for the time being.

Next, I will discuss the company's plan. For the second quarter, we plan to actively respond to demand for high-density server modules and mobile products with DRAM shipments expected to increase by high single-digit percent Q-o-Q. For NAND, we expect shipments to increase by mid-teen % Q-o-Q through expanded sales of 321-layer-based products and enterprise SSDs.

In the HBM business, where we are leading the market, competitiveness is determined not only by DRAM process technology but also by a combination of diverse technological capabilities, including TSV and packaging.

In addition, comprehensive execution capabilities encompassing performance, yield, quality and supply stability are critical. Therefore, we are concentrating our efforts on further strengthening these capabilities. For HBM4, we have been working closely with customers from the early stages of development, and we plan to ramp up volume in line with the agreed schedule of products that meet the customers' required performance level.

In response to increasingly diverse memory requirements driven by the evolution of AI technology, we continue to develop and supply new products across both DRAM and NAND. In DRAM, during the first quarter, we completed development of the industry's first 1C-nanometer-based LPDDR6. Compared with existing LPDDR5X, the product delivers a 33% improvement in data processing speed and more than 20% better power efficiency. We plan to begin full-scale market supply in the second half starting with adoption in a major smartphone customer's next-generation flagship model. In addition, we began mass production this month of our 192 gigabyte SOCAMM2 product based on our most advanced 1c-nanometer process. Optimized for NVIDIA's Vera Rubin platform, this product delivers more than twice the bandwidth of conventional RDIMM and over 75% improved energy efficiency.

For NAND, as demand for high-performance, high-density storage continues to expand in earnest, we will proactively capture growth opportunities. First, we have begun supply of PQC21 client SSD, the first product from our company to adopt CTF-based 321-layer QLC technology. Starting with the client segment, we plan to build a well-balanced product lineup across the entire enterprise market, spanning high-performing TLC and high-capacity QLC in order to respond flexibly to broad-based AI demand.

In particular, based on synergies with Solidigm, which has strength in high-capacity QLC enterprise SSDs, we will further strengthen our competitiveness in addressing a wide range of customer needs in the AI data center and AI PC storage market. As customer demand continues to exceed our supply capabilities, securing sufficient supply capacity to respond to structural demand growth in the AI era is emerging as a core competitive advantage.

Accordingly, our CapEx this year is expected to increase significantly compared with last year with the majority allocated to infrastructure preparation centered on the Yong-in cluster and ramp-up of M15X and to the procurement of key equipment such as EUV tools. Given that semiconductor manufacturing infrastructure takes several years from initial construction to actual operation and thus securing key equipment also involves substantial lead times, we will strategically secure the production base needed to respond proactively to mid- to long-term demand growth. While adhering to CapEx discipline, we will execute investments based on demand visibility, thereby ensuring both supply stability and financial soundness.

Lastly, let me address our financial soundness targets and shareholder return policy. The memory market is facing an unprecedented growth opportunity at the center of the AI era. Given the high return on investment demonstrated by our first quarter results, we believe that reinvesting the cash we generate back into the business is currently the best use of capital. Therefore, we will strengthen our financial soundness to secure the foundation for long-term growth so that we can continue to execute the investments that are strategically necessary over the long term under any market conditions. Considering our all-time high profit generation, we believe that achieving financial soundness with net cash of more than KRW 100 trillion and expanding shareholder returns are goals that can be pursued in parallel.

In addition to dividend, we will also actively review additional shareholder return measures, such as share buybacks and cancellations and establish an implementation plan within the year. Meanwhile, on March 24, we confidentially submitted a registration statement to the U.S. Securities and Exchange Commission related to the proposed ADR offering, and we are proceeding with the goal of listing on the U.S. securities markets within the year. However, specific details of the proposed offering, including its size, structure and timing has not yet been determined and the final decision on whether to proceed with the listing will be made after comprehensively considering the SEC's review of the registration statement, market conditions, investor demand and other relevant factors.

We will provide further updates once specific details are finalized. Going forward, we will continue striving to enhance shareholder value by maintaining the optimal balance among investment for future growth, financial soundness and expanded shareholder returns.

With that, we are now ready to take your questions.

[Interpreted]

[Operator Instructions] The first question will be provided by [indiscernible] Kim from Hana Securities.

[Interpreted] Now my question is the memory spot prices have been on a steep upward trend, but are now showing some signs of weakness. Some are concerned that this may be a signal of a peak out with the demand beginning to soften due to the price burden. But does the company see this as just a temporary adjustment caused by short-term supply and demand factors? So what is the company's view?

[Interpreted] Now we understand that there are various interpretations in the market regarding the recent spot price movements. So let me share with you the company's perspective.

Now first of all, the stock -- the spot market itself takes up a very small part of the overall DRAM market and the types and volume of products that are traded there differ considerably from our business. So we can't really view these changes in the spot market as reflective of the overall market.

Now on one hand, we see that the memory demand from our major customers is increasing across the board, including HBM, server DRAM and eSSD. On the other hand, for the suppliers, the reality is that the suppliers find it hard-pressed to increase supply in the short term. And as such supply-demand imbalance persists, the rising memory price cycle is likely to last longer than in the past.

So the moderate trend in spot prices rather than being a sign of market peak out appears to be a temporary phenomenon resulting from some inventory entering the market from some distribution channels due to the recent price increase.

[Interpreted] The following question will be presented by Sun Woo Kim from Meritz Securities.

[Interpreted] So the prior question was about the spot price, and I -- my question is more about the overall price cycle because at this time, it seems as if the unprecedented super cycle is driving up memory prices. Then what is the company's outlook on the memory price trends down the road?

[Interpreted] Now we see that the current price strength is driven by structural changes in the market, not by the temporary supply-demand imbalance. And that is why we also expect that this trend to be different.

First, the importance of memory has grown more than ever due to AI, and IT companies are competitively increasing their purchase to secure more memory supply.

On the other hand, industry supply has been constrained by investment slowdown following the last downturn and also by the shortage of available space, which limits potential for production expansion for the short term.

Suppliers are expanding their investment by resuming fab construction and infrastructure investment to secure ramp-up capacity, but it will take some time to complete meaningful new clean rooms and production capacity.

As this supply-demand imbalance persists, customers are prioritizing procurement over price and the growing importance of memory in AI computing is also being priced in. Therefore, the favorable pricing environment is expected to continue for the time being.

[Interpreted] The following question will be presented by Nicolas Gaudois from UBS.

Yes. Could you provide us an update on the current progress on your enhanced new long-term agreements and the time line for the expansion? Could you specify how do they differ from past LTAs, whether they apply to DDR, NAND flash and HBM like or more so DDR? And could you share the specific terms and conditions to the extent you can disclose them?

[Foreign Language]

[Interpreted] Thank you very much for the question. And yes, I will try to explain to the extent possible. Now as the memory shortage persists, customer requests to secure medium- to long-term supply volumes have significantly increased. Memory today has become so critical that customers now see memory price and supply uncertainties as key business risks. And for the company as well, reducing investment burden by securing demand visibility is one of our key priorities.

A multiyear LTA must provide business stability for both parties by ensuring supply stability to the customer and demand visibility and stable revenue structure to the seller. Accordingly, unlike past LTAs, we are comprehensively reviewing various approaches and structural options. But due to current supply constraints, we are limited from accommodating all customer requests.

If multiyear LTAs are successfully established, we expect investment efficiency to naturally improve, thanks to demand visibility and stable profitability. Not only that, it can also reduce the volatility that has repeatedly plagued the memory industry, which can improve the market's overall view on the memory business as well.

[Interpreted] The following question will be presented by S.K. Kim from Daiwa Securities.

[Interpreted] Congratulations. Now my question is about the memory or the AI technologies. And I believe that it was also mentioned during the presentation, but then we see that efforts today continue to develop technology and products to improve memory efficiency to address the recent memory shortage. So there are also concerns that such trends could dampen memory demand. What is the company's outlook on this?

[Interpreted] Now as the AI inference market grows rapidly, it is that new technologies are emerging to process the exponentially increasing volume of data. The company believes that these technological advancements will broaden the AI ecosystem and ultimately serve as a catalyst for driving overall memory demand.

Some suggest that LPUs using SRAM could serve as a memory alternative. But unlike GPUs, LPUs operate primarily by using the internal SRAM, which makes them very fast, but also limits the physical capacity. So it is highly likely that we move toward a hybrid architecture where LPUs handle tasks requiring fast response and HBM-based GPUs, handle complex and massive computations. In other words, as AI services diversify, memory will become increasingly tiered and demand for high-performance memory is expected to continue.

At the same time, with the recent introduction of technologies that optimize KV cache by compressing data, there has been speculation that memory demand might decrease, but the intent of this technology is not to use less memory, but to use the same memory more efficiently to provide a wider variety of AI technologies and services. To process longer context and perform more from current inferences, more memory is required. So as AI services become more widespread, memory demand is expected to increase.

While memory optimization technologies have been steadily emerging, the overall trend has been towards diversifying the AI market lowering barriers to entry and expanding the market's overall size. In response to the diversifying needs, we will continue to solidify our market leadership by supplying world-class AI memory products in a timely manner.

[Interpreted] The following question will be presented by Ricky Seo from HSBC.

[Interpreted] Congratulations on the outstanding performance. My question is about HBM4. Now regarding the HBM4 qualification and mass production, it seems as if the market views currently remain mixed. So what is the update on HBM4 qualification and the expected timing for full-scale shipments? And also, if possible, by end of this year, what is the company's expectation about the HBM4's share out of the total HBM?

[Interpreted] Thank you very much for the question. Now from the customers' perspective, HBM is a business where overall competitiveness seems to be more important than any single factors like performance metrics such as speed and power efficiency or quality, yield or supply stability. For our HBM4, we have established a proactive development and supply system in close collaboration with major customers from the early stage, and we are getting ready to ramp up production and supply products that meet customers' requirements in a timely manner in line with each customer's mass production schedule.

Since the launch of HBM2E, we have maintained the highest level of overall product competitiveness in time-to-market, cost, yield and performance as well as customers trust. We have built on that basis to keep strengthening our technological leadership and lead the market with next-generation products such as HBM4 and HBM4E. It is evidenced by the fact that customers' demand for the next 3 years far exceeds our current supply capacity.

Within the limited supply capacity, we are doing our best supply as much HBM as possible to our customers. Given the severe supply shortage is affecting not only HBM, but also the general-purpose DRAM, we are trying to achieve optimal allocation between HBM and general DRAM for the sake of balanced growth in the AI ecosystem rather than trying to maximize revenue.

Moving forward, we will sustain our market leadership based on our product capabilities, including HBM3E and HBM4.

[Interpreted] The following question will be presented by Youngmin Koh from DAOL Investment & Securities.

[Interpreted] Thank you for taking my question which is about investment. Now recently, we see other companies are announcing aggressive investment expansion. Does the company also plan to expand investment to maintain market share and competitiveness? And also on the other hand, it is understandable if there are some concerns about past oversupply issues recurring. So I wonder what the company's view would be regarding investment and also about such concerns.

[Interpreted] Thank you very much for the question. The company's policy is to execute investment with CapEx discipline, taking demand visibility into account. Now given the current sustained robust demand for memory, we expect investment in 2026 to increase significantly Y-o-Y to prepare future infrastructure and secure key equipment to meet demand. Mainly, we are rapidly proceeding with the construction of the Yong-in cluster fab to secure mid- to long-term production capacity. After completion of Phase 1 early next year, we plan to execute investment in stages to complete Phases 2 through 6.

In addition, we plan to move ahead as scheduled with investment in the equipment necessary for the migration to advance processes through next year. So as you can see, we are actively expanding supply to meet customer demand, for example, increasing our investment, but also at the same time, it appears that supply will remain short of demand for the time being. So it is not like -- so we -- so for the time being, the supply will fall short of the structural increase in demand.

And that is also why we are closely monitoring demand changes by strengthening our customer intelligence system and continuously reviewing supply-demand environment against the medium- to long-term demand visibility secured through our long-term customer relations. Now at this time, both customers and suppliers agree on the importance of securing long-term visibility into supply and demand. So there should be no major concerns about oversupply as in the past.

[Interpreted] The following question will be presented by Jay Hyun Kwon from JPMorgan.

[Interpreted] So in the opening presentation, it was also mentioned that the enterprise SSD demand is also growing. So with demand driven by AI expected to fuel rapid growth in NAND demand as well, then what is SK hynix's plan to meet future market demand?

[Interpreted] Yes. As it was rightly observed, NAND is no longer a simple storage device, but a core component that determines computational speed and efficiency, and it is also projected for long-term growth.

Now as AI models develop further, the volume of intermediate data processing, known as KV cache is increasing exponentially. And that is why customers are now adopting high-performance high-capacity eSSDs on a large scale. So the company plans to actively address this demand by strengthening our technological capabilities and expanding production capacity even under the constraints of limited supply and investment conditions.

First, in April this year, we developed the world's first 321-layer QLC and completed customer qualification, securing an overwhelming technological lead. And looking ahead, we aim to flexibly address the increasingly tiered AI storage demand by establishing an optimal product lineup that covers both high performance and high capacity, and we will ultimately further strengthen our eSSD-centric product mix to proactively address market changes driven by the growth in KV cache.

We are also accelerating our tech migration. To maximize speed production, we plan to migrate more than 50% of our domestic production to 321-layer technology by the end of this year. And this 2-generation jump from 176 layer to 321 layer is expected to yield significant productivity gain for the company. As the growth potential continues to strengthen in the NAND market, along with the AI market, we plan to expand our influence within the NAND market through flexible and proactive investment in line with market conditions.

[Interpreted] The following question will be presented by Young Ho Ryu from NH Investment & Securities.

[Interpreted] Congratulations on the performance. My question is more about the long-term view. So as the AI industry continues to develop with growing need for computation, it seems as if there are growing demand for other applicability [ and purposes ] for memory. So the question is then for the company, then how is the company preparing for such next-generation memory market post HBM?

[Interpreted] Thank you for the question about the next-generation memory. So I will respond to the question about DRAM and the other will follow the -- follow up with the question on NAND. Now as you see, the AI market continues to grow. New platforms are emerging and technologies are creating a multi-tiered memory architecture and broadening application areas leading to increasingly diversified demand for AI memory.

The company is closely following these technological developments and customer needs, trying to identify them early on and are systematically preparing to respond to the next-generation AI memory market in a timely manner.

First, we plan to begin mass production and supply of our 192-gigabyte SOCAMM2 product based on 1c-nanometer LPDDR5X starting this month. This product delivers more than twice the bandwidth and over 75% improved energy efficiency over existing RDIMMs making it optimal for high-performance AI computing.

As the demand for inference grows, there is the CXL pooling solution, which can serve as one option for offloading the rapidly increasing KV cache. And building on our first-generation CXL memory modules, based on CXL 2.0, which completed customer qualification last year, we aim to maintain our leadership in this emerging market by delivering enhanced capacity and performance in our second-generation product that supports CXL 3.0.

In fact, we already signed an MOU with a cloud provider last year to validate and optimize next-generation AI solutions such as CXL and PIM in new AI service environment to be used as the starting point to actively expand our technical partnerships with global customers.

[Interpreted] Meanwhile, in NAND, we are preparing next-generation storage solutions to build high-performance, high-bandwidth and high-capacity storage infrastructure tailored for AI workloads. In particular, HBF is a technology that can deliver ultra-high bandwidth through 3D stacking, and we launched a consortium last February to standardize its specifications. And looking ahead, we aim to continue to lead new markets, not only in DRAM but also in NAND through the push for global standardization and commercialization of HBF.

We will continue to develop optimized memory solutions that are right for the AI environment like edge AI and physical and we will continue to strive to further solidify the AI memory leadership that we have established through HBM.

[Interpreted] The following question will be presented by Peter Lee from Citigroup.

[Interpreted] My question is also about the next-generation product. So we just got an update on the HBM4. And there is also growing interest in the market about the next generation, the HBM4E. So the question is what are the key factors that will contribute to the company's technological competitiveness in the HBM4E and also how can you differentiate from other companies? And please also explain the packaging technology and logic die deployment plan.

[Interpreted] Thank you for the question. Now for the HBM4E, we are preparing it in close consultation with our customers regarding the shipment schedule and product specifications. Our internal plan is to start supplying samples in the second half of the year, and we are moving ahead smoothly with development targeting mass production in 2027.

First, the base die to be used in HBM4E is under development based on the optimal technology to meet customer performance requirements. And the work is moving on smoothly in collaboration with our customers. And the core die will adopt 1c-nanometer technology to meet growing customer performance requirements.

Our 1c-nanometer technology is proven to deliver industry-leading performance. Mass production began in late 2025 and both yield and mass production capabilities have already reached a mature stage. And this will help us supply HBM4E to customers with stable performance and volume. And we plan to develop HBM4E in a timely manner through further technology internalization and customer validation and to keep maintaining our leadership in HBM technology through unparalleled mass production capabilities and product quality.

[Interpreted] The following question will be presented by Jong Wook Lee from Samsung Securities.

[Interpreted] Now there was a response about investment earlier. And my question is more specifically about infrastructure investment. So the -- what is the operational direction for the Y1 fab, which is scheduled to open early next year? And also, are there any plans by the company to build or acquire additional fabs outside of Yong-in to secure medium- to long-term production capacity?

[Interpreted] To meet the medium- to long-term demand, we have decided to speed up the opening of the Phase 1 clean room in Fab 1 by 3 months from May 2027 to February 2027 and construction of the Yong-in fab is progressing on schedule.

The Yong-in cluster will become the largest state-of-the-art production complex in history and will serve as the bedrock of our mid- to long-term operations. Phase 1 is to produce DRAM and we intend to continuously review the products and technologies to be deployed in Phases 2 through 6 to ensure efficient operations that are aligned with market demand.

Now at this point, we have no plans to build or acquire additional fabs outside Yong-in, but we also recognize that the ability to reliably supply the products that the customers want when they want them is emerging as a key competitive advantage in the AI era, which means securing large-scale production capacity and a stable supply system is more important than ever before. So we will keep making every preparation necessary to respond flexibly to the global memory demand growth over the medium to long term.

[Interpreted] The following question will be presented by Simon Woo from Bank of America.

[Interpreted] My question is about the commodities that SK hynix requires. So for example, the helium and bromine that are dependent on the Middle East, tungsten that is imported from China and also LNG that is used for self-power generation by SK hynix. So with regards to the shortage of these materials, what is the impact on the company? And what is the company's strategy to secure the supply over the longer term? So if you could provide us with an overall update about the commodities front?

[Interpreted] Now based on past experience in international conflicts, we are fully aware of the risks associated with commodities and energy supply resulting from geopolitical changes, and we have already secured our responses. As for the key industrial gases like helium and bromine, we have already diversified our suppliers, and we also have built up sufficient inventory as well. So any short-term or long-term impact on our production capacity will be quite limited.

And for tungsten, it is true that the prices have risen due to geopolitical issues recently, but then we already have secured sufficient inventory, and there appear to be no disruptions in supply. So there is no impact on our production. And regarding electricity, energy prices have risen due to delays in oil and LNG exports, but we source LNG through long-term agreements, which keeps any price fluctuations to the minimum. And any impact on our business will be limited. We will continue to closely monitor market conditions and try to minimize new risks in our business operations.

[Interpreted] The last question will be presented by Dong Hee Han from SK Securities.

[Interpreted] My questions are about the company's effort to enhance corporate value, for example, shareholder return and ADR. Now regarding the goal of KRW 100 trillion in cash announced at the shareholders' meeting, there are also concerns that shareholder return may fall short of market expectations along the way. Then what is the company's direction or plan for shareholder return policies? And also, could you give us an update about ADRs?

[Interpreted] As the scale of capital expenditure structurally increases due to growing AI demand, the company decided that securing global leading financial strength is essential to ensure stable investment that is not affected by market conditions and also to respond promptly to customer demand.

As explained in the presentation, given the company's significantly enhanced profit generating capability, we believe that we can sufficiently balance between financial soundness and expanding shareholder returns. We plan to develop additional shareholder return measures within this year, including not only dividends but also share buyback and cancellation and communicate with the market accordingly.

The company already demonstrated our commitment to shareholder returns through a total annual dividend payout of KRW 2.1 trillion and cancellation of KRW 12.2 trillion in shares in 2025. Going forward, we will actively explore ways to steadily increase returns to shareholders in line with our earnings growth.

And about ADRs, the U.S. SEC review is currently underway. And I ask for understanding that we cannot provide any information beyond what has already been disclosed in accordance with domestic and international laws and regulations. Having said that, we will communicate more details with the market once they are realized.

[Interpreted] With that, we conclude the SK hynix 2026 First Quarter Earnings Release Conference Call. Thank you, everyone, for your participation.

[Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]

Good morning, afternoon and evening. Thank you for joining SK hynix' 2025 Q4 Earnings Release Conference Call. Following SK hynix presentation, there will be a Q&A session. [Operator Instructions] With that, we are now ready to begin.

Good morning, afternoon and evening. This is Park Seong Hwan, Head of IR at SK hynix. Welcome to the SK hynix 2025 Q4 Earnings Release Conference Call. Today, we are joined by President of Corporate Center, Song Hyun Jong; Chief Financial Officer, Kim Woo-Hyun; Head of DRAM Marketing, Park Chan-Dong; Head of NAND Marketing, Chong; and Head of HBM Sales and Marketing, Kim Jeong Tae.

Let me issue a disclaimer that all outlooks presented by the company are subject to change depending on the macroeconomic and market circumstances. With that, we will now begin SK hynix's earnings release conference call for fourth quarter of 2025. Song Hyun Jong, President of Corporate Center, 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. This is Song Hyun Jong, President of Corporate Center. Allow me to first introduce the company's performance for the fourth quarter of 2025. During the fourth quarter, strong demand for memory products continued, driven by intensifying global investment in AI infrastructure. Demand increased significantly, not only for HBM, but also for conventional memory. The pace of industry supply growth could not keep up with demand, leading to broad-based price increases and the formation of a highly favorable market environment. As prices for both DRAM and NAND rose sharply and NAND shipments increased, fourth quarter revenue reached KRW 32.8 trillion, up 34% Q-o-Q and 66% Y-o-Y, thereby achieving the highest quarterly revenue.

DRAM shipments grew by low single-digit percent, driven by increased sales of HBM3E products and DDR5 for servers. Shipments of high-density DDR5 modules increased by approximately 50% Q-o-Q, leading demand growth centered on AI and HPC. ASP rose 20% Q-o-Q, reflecting a significant increase in conventional DRAM prices. For NAND, with the base effect from lower shipments in the previous quarter, demand growth in mobile products and eSSD led to approximately 10% increase in shipments Q-o-Q exceeding guidance. ASP also increased low 30% Q-o-Q as price increases accelerated. Driven by substantial price increases, profitability improved for both DRAM and NAND. As a result, fourth quarter operating profit reached KRW 19.2 trillion, up 68% Q-o-Q and 137% Y-o-Y, marking a record high-quality operating profit with the operating margin of 58%.

Depreciation and amortization for Q4 was KRW 3.6 trillion. EBITDA amounted to KRW 22.7 trillion with an EBITDA margin of 69%. Net nonoperating loss reached KRW 1.5 trillion, including KRW 6.6 trillion in valuation gains on investment assets and KRW 8.4 trillion in loss on derivatives related to exchange rights on exchangeable bonds. Pretax income amounted to KRW 17.7 trillion. Net profit totaled KRW 15.2 trillion, and net profit margin was 46%.

Now moving on to the annual performance for 2025. The year 2025 marked a fundamental shift in the structure of memory demand driven by the broad adoption of AI. The memory market has moved beyond increasing content with growing demand for multidimensional performance that simultaneously meets requirements for speed, efficiency and reliability. The bar for product competitiveness has been raised significantly. The company proactively strengthened its technological competitiveness with a focus on AI memory while expanding the share of high value-added products in the portfolio successfully securing both profitability and growth. As a result, full year 2025 revenue reached KRW 97.1 trillion and operating profit totaled KRW 47.2 trillion, representing Y-o-Y growth of 47% and 101%, respectively. These achievements are not really the result of short-term favorable market conditions, but rather the outcome of our strategic execution aligned with an increasingly AI-focused demand environment.

The year 2025 is also a milestone year in which the company once again demonstrated its world-class technological leadership. In DRAM, following World First assembled shipments of HBM4 to major customers in March last year, we were also the first in the world to secure mass production revenues in September further solidify our technology leadership in the AI memory market. HBM revenue more than doubled Y-o-Y driven by significant increase in HBM 312 sales and contributed to record high annual DRAM revenue and operating profit. In conventional DRAM, we commenced full-scale mass production of 10-nanometer DDR5, which delivers industry-leading performance and cost competitiveness. Through the development of the industry's highest density 256 gigabit DDR5 RD based on 10-nanometer 32 gigabit die, we have further demonstrated our leadership in fiber model.

In NAND, despite a challenging demand environment, we continue to strengthen our technology leadership, including the successful development of 321 layer QLC products in the first half of last year, in particular, by responding proactively to the recovery in enterprise asset demand in the second half, we achieved record high annual NAND revenue. Cash and cash equivalents amounted to KRW 34.9 trillion at the end of 2025, an increase of KRW 20.8 trillion from the end of 2024, while borrowings decreased by KRW 0.4 trillion to reach KRW 22.2 trillion. The debt ratio declined significantly to 18% and the company transition to a net cash position, resulting in a substantial improvement in our financial structure.

Now let me share our market outlook. AI models are rapidly transitioning to an inference centric base in which large volumes of user requests are processed in real world service environment. Computing workloads are shifting from architectures with a focus on high-performance servers towards more distributed architectures. The core of system design is no longer limited to compute performance online but increasingly focuses on end-to-end system efficiency, including data movement and storage. Accordingly, demand is expected to continue expanding not only for high-performance memory, but also across server DRAM and NAND. In particular, servers that shipments are expected to grow by high teens percentage range in 2026 with solid growth anticipated over the mid- to long term. A key driver of this growth is not only AI servers but also the specification upgrades of general-purpose servers. To support AI workload requirements for memory density and bandwidth and general purpose servers are rapidly increasing. As a result, demand for server DRAM and enterprise SSDs is expected to grow structurally at a pace above the overall market growth. PCs and mobile devices are expected to see short-term shipment adjustments due to rising component costs and weakened consumer sentiment. Memory content per device is expected to grow at a slower pace due to price increases and supply constraints. Memory demand for PCs and mobile applications is expected to grow at a smaller pace than the overall market, reflecting such market conditions, despite the explosive increase in memory demand driven by the server market, demand growth for DRAM and NAND in 2026 is expected to remain at over 20% and high-teen percent, respectively.

Next, I will discuss the company's plans. Despite first quarter generally being a weak season, customer demand remains strong. However, given constrained supply conditions, we plan to maintain DRAM shipments at a similar level to that of last quarter, while NAND shipments are expected to decline somewhat due to the base effect of the previous quarter. As demand continues to surge, both AI and conventional memory, the industry has accelerated capacity additions and tech migrations to advanced process nodes. However, considering physical limitations in production space and the ongoing tech migration, we expect high supply demand condition to proceed for the time being. We are capable of supplying both HBM3E and HBM4 in a reliable manner. And through our technological leadership, proven quality and production capabilities, we have earned strong customer trust. As a result, we have consistently collaborated with our customers to introduce new products in a timely manner, and HBN4 is progressing under the same collaborator framework in line with the schedule agreed with our customers.

After securing mass production revenue in September last year, we are currently in mass production of the volume requested by customers. Looking beyond HBM4, competition is expected to evolve beyond simple backing toward custom HBM We are actively engaging in custom HBM technology discussions with key customers and through on in collaboration with our partner companies, we are progressing seamlessly to ensure optimal product supply.

For conventional DRAM, we plan to increase the production of high value-added products, while accelerating tech migration to the 10-nanometer process, we plan to expand our product portfolio, including For NAND, we will maximize product competitive transition to 321 nano technology.

In particular, through the development of next-generation 245 terabyte products, we aim to secure a leadership position in the ultra high-density storage market, driven by the expansion of AI and workloads. We plan to maximize production within feasible limits to meet customer demand. This year, we will ramp up capacity earlier than planned while accelerating tech migration to 1c-nanometer DRAM and 321 layer NAND. Over the mid- to long term, we aim to rapidly expand our production base in Phase I back while also proceeding without delaying the preparation of -- and the advanced packaging facility in Indiana. We will strengthen our global integrated manufacturing capabilities, enabling flexible responses to changes in customer demand. In 2026, CapEx is expected to increase considerably from that of last year, reflecting the expansion of production capacity and infrastructure. We will continue to adhere to our CapEx discipline based on a comprehensive assessment of demand visibility and investment efficiency. Ultimately, we aim to go beyond being a simple product supplier and further strengthen our role as a full stack creator that enables customers AI performance requirements from a system level perspective. By integrating our capabilities in the high-performance memory with process technology, packaging and solutions capabilities, we will maximize customers' computing efficiency and generate sustainable earnings growth. .

Finally, let me explain shareholder returns for 2025. The company introduced a new shareholder return policy applicable from 2025, under which securing financial soundness was identified as our top priority. Our objective was to transition to a net cash position and to maintain an appropriate level of cash reserves. We also stated that should meaningful free cash flow will be generated, we would consider shareholder returns even before the expiration of the policy period. Based on the financial flexibility secured in 2025, the BOD has resolved an additional shareholder return package in response to our shareholders' continued trust and support.

First, in addition to the fixed dividend, we will pay an additional cash dividend of KRW 1,500 per share. Accordingly, the year-end dividend per share will be INR 1,875. As a result, the total cash dividend per share for fiscal year 2025 will amount to INR 3,000, and the total dividend payout will be approximately KRW 2.1 trillion. Second, excluding the portion of treasury shares reserved for employee compensation, the company plans to retire all remaining 50 million treasury shares, equivalent to 2.1% of total shares outstanding. Based on the closing share price of the day prior to the Board meeting, this treasury share retirement represents a value of approximately KRW 12.2 trillion. This action is intended to enhance per share value and demonstrate the company's long-term commitment to shareholder value creation. The company remains committed to a disciplined capital allocation framework that seeks to maintain an optimal balance among future growth investments, financial stability and shareholder returns. We will continue to make efforts to enhance shareholder value. Going forward, we will close...

[Foreign Language]

Now Q&A session will begin.

[Operator Instructions]

The first question will be provided by Peter Lee from Citigroup. .

[Interpreted] First of all, congratulations on the record high performance for the company. Now my questions are pertaining to the HBM4. So recently, especially last month, there has been some noise about SK hynix's progress on HBM4. So can the company share your current status of the HBM4 development and the expected timing for mass production? And also what are the plans to maintain HBM performance and mass production competitiveness?

[Interpreted] Thank you very much for your questions. Now SK hynix has been a leading pioneer in the HBM market since HBM2E, working together as one team with customers and infrastructure partners. And it is not only about superior technology, our mass production experience and customer trust in our quality built up over time are not something that can be overtaken in a short period of time. Likewise, for HBM4, customers and infrastructure partners show strong preference and expectations for our products, prioritizing our products over others. And we aim to extend this to HBM4 to take up overwhelming market share, just as we did with HBM3 and HBM3E.

As mentioned in the presentation, preparations for HBM4 are underway as planned, according to the time line agreed with customers, and we are currently mass-producing volumes requested by customers. represents a major technical achievement that will enable customers' requirements based on existing 1B nanometer process. Using our proprietary packaging technology, advanced we plan to secure yields comparable to the 12 high HDM3E products.

Even as we maximize production, we cannot meet HBM demand 100%. So some competition is expected to enter the market. Such developments notwithstanding, our market leadership and leading supplier position will continue based on performance, producibility and quality. Next question, please. .

[Interpreted] The following question will be presented by Hyung-keun Ryu from Daishin Securities.

[Interpreted] Congratulations on the performance. The question is about the LTA. So now with the growth -- continued growth of the market, there are talks of LTAs being reached, so can the company share some content -- can the company share an update about the LTAs that are underway for the year 2026. And how are the LTAs different from past LTAs?

Thank you. Please understand that I cannot divulge details about LTA is currently under discussion with our customers. But to your question on how the long-term supply agreements or LTAs are different today, now as you know, LTAs already existed, but they were generally lose contracts on volumes and tended to be quite fluid, depending on market conditions. The LTAs being discussed today are expected to reflect strong mutual commitments between customers and suppliers, not simply indicating the intent to buy. Because memory production today requires cutting-edge technology and far bigger investment, compelling suppliers to seek high visibility into demand. And this is also why customers now prefer multiyear contracts, but capacity constraints make it difficult to accommodate all customer requests. And bearing this in mind, SK hynix will keep exploring ways to enhance both our customers and our own long-term operational stability.

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

Now regarding the unprecedented surge in memory demand, some suggest that it could be driven by pull in demand to secure inventory, but most see this as growth in real demand due to tight customer inventory. So what is the customers' inventory level in key applications as seen by the company? And what is the company's own inventory status?

Thank you. Now as you would know, memory market today is experiencing explosive demand growth along with continued expansion in AI infrastructure investment, but supply cannot keep pace with demand, resulting in a severe supply-demand imbalance. Most customers are struggling to secure memory volumes and are persistently demanding increased supply. So it appears that customer inventory levels are -- have decreased overall. Server customers, in particular, as soon as they secure volume, move on to building sets, which apparently keeps driving down inventory. Customers can hardly secure enough volume to build up inventory and set builds use up memory fast.

With memory being seen as debottleneck in data center infrastructure expansion, server customers are expected to keep trying to increase purchase to secure their volume. At the same time, PC and mobile customers are also experiencing supply constraints as well as the direct and indirect impact from the strong server side demand and their inventories also continued to decline. For the company, even as we strive to increase production, DRAM inventory decreased in Q4 Q-o-Q. The tight inventory trend, particularly for server DRAM, is expected to continue throughout the year. With memory selling out as soon as it is produced, our inventory is projected to decline even further in the second half of the year.

and now about NAND. NAND inventory is also observed to be falling rapidly among server customers. This trend of falling inventory is expected to continue, particularly for enterprise SSD products. And the company's inventory level is also declining rapidly with NAND inventory weeks at the end of last year, nearly matching that of DRAM.

The following question will be presented by Simon Woo from Bank of America.

[Interpreted] And also, congratulations on the record high performance. Now my question is pertaining to the company's plan on managing the capacity by customer or product. And this is because while the sharp rise in memory prices coming from tightening supply/demand positively affects the company's performance, SK hynix' customers may be struggling to secure enough volume to sustain their operations. So the question is how the company plans to manage its both existing and newly acquired capacity.

[Interpreted] It is true that the AI industry's explosive growth has brought unprecedented changes to the memory market as well. In particular, it has deepened the supply-demand imbalance because while demand for AI memory has surged, ramping up supply takes time.

And at a time like this, we believe it is important to prioritize meeting customer needs and building trust in the market, not focusing solely on short-term results. And that is why although the space is limited, we are working to maximize production to accommodate the rapidly increasing demand. To meet the HBM demand, we are adding new capacity for 1d nanometer at M15x, which was completed last year, while enhancing productivity through yield improvement, not only that, to address demand for conventional DRAM and NAND, we are accelerating tech migration to 1c-nanometer and 321 layer. As a memory leader, Hynix remains committed to creating a sustainable semiconductor ecosystem where we grow together with our customers.

The following question will be presented by Hyun Kim from Meritz Securities.

[Interpreted] The company's performance is also expected to improve significantly this year as well. So then my questions are twofold. Are there any plans to continue with the extra dividends and share cancellations going forward? And the second question is the company also recently canceled the treasury shares. Then in order to further enhance the shareholder value, are there plans to issue ADRs? And if yes, then through what approach.

[Interpreted] Thank you for the question. Now as explained last quarter, the company's goal in trying to achieve financial soundness is maintaining an adequate level of cash reserve that enables stable business operations even during industry fluctuations and that also allows necessary CapEx to sustain competitiveness.

Market conditions today point to continuing increase in the CapEx required to meet demand. Given the memory market's growth potential and high investment returns, our belief remains unchanged, that reinvesting available funds into our business to enhance corporate value will be the best use of cash. Having said that, our financial soundness improved faster than expected at the time when we announced the current shareholder return policy following last year's record performance, and as the question rightly mentioned, performance improvement is likely to continue this year. So leveraging the financial room secured last year, we are implementing additional shareholder return to show our appreciation for their support and to enhance shareholder value. And we plan to continue reviewing additional shareholder return measures and timing based on performance and cash flow at the time. Each method of shareholder return, be it dividend payout or share buyback has different characteristics. While we maintain the current shareholder return policy, we will ensure flexibility in its administration to seek the best option at the given time that can meet market expectations.

Meanwhile, as mentioned in yesterday's inquired disclosure, we are looking into various options to enhance corporate value. Nothing has been finalized to this date, and we plan to make careful decisions comprehensively considering internal and external conditions.

The following question will be presented by S. K. Kim from Daiwa Securities.

[Interpreted] And my question is pertaining to the NAND flash. So we see that the AI storage market continues to grow. So what is the company's outlook on the AI storage market down the road? And also in the same context, what is the company's plan for response to this demand in the NAND storage?

[Interpreted] Thank you for your question. NAND today is changing completely not only a data store, but becoming a storage solution that directly supports AI computation workflows.. As AI inference keeps advancing, GPU and CPU memory alone cannot satisfy all requirements. So the key value cash offloading has become essential to ensure smooth inference services. As AI's data utilization becomes more precise and fast, it is driving a structural surge in demand for high-performance, high-capacity enterprise SSDs that can support high-speed data and IO. What we are particularly interested in are the fundamental changes at the AI server architecture itself as a result of this. So as I briefly explained earlier, whereas the SSD was a peripheral under the CPU-centric architecture, in the recent context memory environment and GPU-centric I/O server architecture, it is increasingly becoming a central part in the compute pipeline.

To address these changes, we are developing next-generation storage products alongside our conventional storage lineup based on which we intend to strengthen our competitiveness in the enterprise business. So what we are preparing first is to develop the ultra performance enterprise SSD. Demand is rising for storage with very fast I/O and ultra-low latency for real-time inference and GPU-based servers. So we plan to secure technological leadership in the growing market by preparing for future technology with high IOPs, SST.

And we will keep responding to the ever evolving AI server market. We will further develop the HBF technology, which is an extension of HBM and expand our lineup of ultrahigh capacity enterprise SSD that enables key value cash and diverse data offloading to meet customer requirements and overcome data center power and space limits.

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

[Interpreted] Now with the recent strong -- very strong demand for DRAM, especially coming from the server side, there is literally a skyrocketing prices which are -- which is likely to heighten the burden on the customers. So if the prices continue to rise, then the cost burden on the PC and mobile customers could also grow. So I -- my question is then, are there any request or demand to adjust the set shipment or the -- or to downgrade the content?

[Interpreted] Thank you for the questions. Following the recent sharp rise in memory prices, some volume adjustments have appeared mainly among PC and mobile customers. This appears to be due to the set manufacturers raising the price on finished products to defend their margin, which has temporarily dampened consumers' purchasing power. Some customers are becoming more conservative revising their shipment plans or reviewing spec adjustments for price-sensitive, lower-tier products. But despite this, expectations for on-device AI are driving replacement demand, especially towards high-end products. And as a result of this, it is not likely that the impact from the adjustments in PC or mobile shipment will spill over into a broader demand contraction. Not only that, over the longer term, as AI features keep improving, they will become part of the default spec, not just an option, which means there will be structural increase in memory content per device. And this should partly offset the price-driven content adjustment. In this market environment, SK hynix will strive to ensure stable supply as much as possible within the limited resources by reflecting customers' product strategies and changes in demand.

The following question will be presented by Min Chae from Korea Investment & Securities.

[Interpreted] My question is with regards to the AI company that has been already disclosed. Now the head offices CapEx size is also significantly increasing. And at this time, the company is also planning on a large-scale external investment. So what are the benefits or the synergies that the company is looking for in establishing the AI company.

[Interpreted] Thank you for your questions. as aI technology continues to advance, memory emerged as key in AI competition. It's no longer about improving individual chips performance but there is a full-fledged race to optimize the system efficiency. As the market continues to change, Hynix plans to become not only a component provider, but a partner in the AI data center ecosystem. And that is part of the reason why we decided to set up an AI co as a way to proactively address the AI business environment and secure future growth engines. The AI co will be established in the U.S., the center of AI technology and market changes, and it will explore companies with key AI capabilities and discover and follow up on opportunities for commercializing AI solutions.

And for your information, the investment commitment in AI co is not large relative to our financial performance or cash generation capacity and the investment will be dispersed sequentially once the investment decision is finalized. So based on our AI memory competitiveness and leveraging the AI company, SK hynix will actively respond to global AI market changes, strengthen technological and business capabilities and grow into a full stack AI memory creator.

The following question will be presented by Young Ho Ryu from NH Investment & Securities.

I would also like to congratulate the company on good performance, which is on CapEx. So it was mentioned that the 2026 CapEx will increase significantly year-on-year. So by how much exactly? And also, can the company maintain CapEx to revenue around the mid-30% level this year?

[Interpreted] Now as explained, the CapEx in 2026 is expected to increase significantly over -- year-over-year, due to capacity expansion, accelerated tech migration and investments in future infrastructure. But at the same time, we will maintain CapEx discipline by monitoring market conditions and balancing demand visibility with investment efficiency. So there will be increase in CapEx, but we also expect revenue to grow substantially as well. So we do not anticipate any difficulties in maintaining CapEx discipline at the mid-30% range. And incidentally, the investment in the AI company discussed just now is not included in CapEx and therefore, does not affect FCF calculation.

The last question will be presented by Lee from DS Investment & Securities.

[Interpreted] My question is on tariffs. The U.S. government recently talked of a 100% tariff on semiconductors if the fabs are not built in the U.S. So what is the company's position and plan? And will there be a need for a plan to build additional fabs in the U.S.?

[Interpreted] Thank you for your question. Now building fabs overseas involves numerous factors to be considered both factors, both inside and outside of the company. So for now, we will monitor the discussions between the government and communicate the company's direction at a later date.

Thank you very much. With that, we conclude the SK hynix 2025 Fourth Quarter Earnings Release Conference Call.

[Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]

[Interpreted] Good morning, afternoon and evening. Thank you for joining SK hynix' 2025 Q3 Earnings Release Conference Call. [Operator Instructions] Please note that presentations will be interpreted simultaneously and Q&A session consecutively.

With that, we are now ready to begin.

[Interpreted] Good morning, afternoon and evening. This is Park Seong Hwan, Head of IR at SK hynix. Welcome to the SK hynix 2025 Third Quarter Earnings Release Conference Call. Allow me to introduce the executives present here with me today. We're joined by Chief Financial Officer, Kim Woo-Hyun; Head of DRAM Marketing, Woo-Hyun Kim; Head of NAND Marketing, Kim Seok; and Head of HBM Sales and Marketing, Kim Kitae.

Let me issue a disclaimer that all outlooks presented by the company are subject to change, depending on the macroeconomic and market circumstances.

With that, we will now begin SK hynix' earnings release conference call for third quarter of 2025. CFO, Mr. Kim Woo-Hyun 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. Allow me to first introduce the company's performance for the third quarter of 2025. Earlier in the year, we expected more moderate demand conditions in Q3 due to external uncertainties and impacts of some preemptive purchases and deferred tax. However, we ended up witnessing a highly favorable market environment with a spike in demand for memory products for servers, including HBM, driven by surging AI infrastructure investments by big tech companies.

Third quarter revenue again recorded record quarterly revenue of KRW 24.4 trillion, up 10% Q-o-Q and 39% Y-o-Y. This was driven by stronger DRAM and NAND pricing as well as increase in DRAM shipments from rising demand.

DRAM bit shipments exceeded guidance by increasing high single digits sequentially, driven by growing sales of HBM3E 12 high products and server DDR5 to support AI demand as well as seasonal demand recovery for LPDDR5 products. In particular, shipments of high-density DDR5 models of over 128 gigabytes, doubled Q-o-Q for 2 quarters in a row, clearly demonstrating robust growth in HPC-related DRAM demand.

ASP rose by mid-single-digit Q-o-Q with strong ASP growth for conventional DRAM products.

For NAND, bit shipments decreased by mid-single-digit Q-o-Q given high base from the previous quarter, but enterprise SSD shipments grew by double digits amid rising demand from AI servers.

ASP increased by low teen percent compared to previous quarter, supported by NAND price recovery and higher mix of enterprise SSDs with pricing premium.

Operating profit reached KRW 11.4 trillion, up 24% Q-on-Q and 62% Y-o-Y, also marking an all-time high. Operating margin improved by 5 percentage points Q-o-Q and 7 percentage points Y-o-Y to 47%, driven by strong sales of leading-edge products such as HBM, high-performance DRAM and enterprise SSD. This marks the first time in the company's history that quarterly operating profit has exceeded KRW 10 trillion.

Depreciation and amortization expenses in Q3 were KRW 3.6 trillion, resulting in EBITDA of KRW 14.9 trillion and an EBITDA margin of 61%.

Nonoperating income net of expenses was KRW 3.4 trillion, including KRW 0.21 trillion of foreign currency-related net gain due to stronger U.S. dollar at the end of the quarter and KRW 3.3 trillion of valuation gains on investment assets.

Pretax income was KRW 14.8 trillion. Net income was KRW 12.6 trillion and net profit margin stood at 52%, again reaching record high level.

At the end of Q3, cash and cash equivalents stood at KRW 27.9 trillion up KRW 10.9 trillion from that of last quarter.

Interest-bearing debt increased by KRW 2.2 trillion to KRW 24.1 trillion, resulting in a net cash position of KRW 3.8 trillion. Accordingly, [indiscernible] to equity ratio improved by 1 percentage point Q-o-Q to 24%.

Now let me share our market outlook. In 2025, despite ongoing geopolitical and macroeconomic uncertainties, such as tariffs, the memory markets saw mixed expectations, optimism regarding exclusive AI growth alongside concerns about monetization. Recently, however, global investments in AI infrastructure has become the top priority for AI market expansion, driving significant demand growth, not only for HBM, but for broader memory demand such as DRAM for general purpose servers and enterprise SSD.

The AI market is now shifting rapidly from the training phase of large models to the inference phase where users actively utilize AI services. As AI models Evolve into multimodal forms and inference-based AI services spread across industry, the number of concurrent users and the need for faster, more accurate responses are rising dramatically, causing an exponential increase in the number of tokens processed.

While the training stage primarily involved computation on large-scale AI servers, the evolution toward inference requires handling vast numbers of tokens with no latency. This is driving efforts to distribute computing workloads across not only AI servers but a variety of infrastructures such as general purpose servers and edge devices.

KV cache or key value cache, which is intermediate computation results generated during inference grows proportionally with context plans. When HBM alone cannot store all of this data, it is offloaded sequentially to conventional DRAM and SSDs. With AI systems processing multiple user requests in parallel, the combination of growing output tokens and longer context windows is exponentially increasing memory usage during inference. As a result, expansion of the AI inference market is driving demand not only for HBM and high-performance DDR5 but also for enterprise SSD, signaling a structural shift in both DRAM and NAND demand.

Leading AI companies are now accelerating investments backed by monetization and forming strategic partnerships to support this growth. This trend will lead to further expansion of AI data centers, creating robust demand across a wide range of memory products from HBM to conventional DRAM and NAND.

Meanwhile, the smartphone and PC markets are expected to show moderate growth, reflecting ongoing inflationary and macroeconomic uncertainty. However, as users increasingly experience on-device AI, AI functionality is spreading even to low- and mid-end smartphones. While AI PCs are expected to account for over half of total PC market, Therefore, we expect content growth will continue to drive memory demand for consumer applications. Reflecting such demand environment, DRAM demand growth is expected to rise from high teen percent this year to over 20% next year, while NAND device growth is projected to improve from midteen percent this year to high teen percent in 2026.

Next, I will discuss the company's plan. In the fourth quarter, we plan to continue to expand sales of HBM server DRAM and enterprise SSD. However, considering our normalized levels of inventory, we expect DRAM and NAND bit shipments to increase by low single-digit Q-o-Q for both products.

For HBM, we have completed discussions with key customers for next year's HBM supply. Our HBM4, which we have completed development and mass production preparation in September not only fully missed customer performance requirements but also support highest speed in the industry. We will start HBM4 shipments in Q4 this year with further expansion planned for 2026, reinforcing our leadership position in the HBM market.

For conventional DRAM, we plan to meet increasing customer demand by securing a full line up of the most advanced 10-nanometer-based product across server, mobile and graphics segments.

Mass production of 1c nanometer is already ongoing smoothly, and we plan to accelerate migration in 2026 to maintain our technology and cost leadership.

For NAND, where demand recovery has been slower, we are deploying world's highest 321 layer technology on various solution products to be ready when market conditions improve. We will also focus on supporting the growing enterprise SSD demand, all the while continuing to operate with a profitability-focused approach.

Furthermore, in line with growing demand opportunities in AI servers, we are investing in tech migration to expand supply of both TLC and QLC products based on the 321 layer platform next year. Meanwhile, we have secured customer demand across all DRAM and NAND products, including HBM through next year.

While we are currently doing our utmost to meet customer demand, AI memory demand is significantly exceeding expectations, and this trend is expected to continue for the foreseeable future.

To respond with me, we have recently opened a clean room ahead of schedule at M15X and begun equipment installation to rapidly secure our new capacity. For conventional DRAM and NAND, we will accelerate the transition of existing capacity to advanced nodes to ensure robust responsiveness to rising demand. As a result, our CapEx in 2026 is expected to increase from this year's level. While Continuing to maintain CapEx discipline, we will plan our investments in an optimal manner to support market demand.

AI technological innovation is heralding fundamental changes of course industries and society as a whole. The memory market is transitioning into a new paradigm with the emergence of HBM, and AI-driven demand is now beginning to expand across all product lines. We have led the market from the inflection point of this AI-driven industrial transformation, leveraging our HBM competitiveness to deliver differentiated performance.

Moving forward, we will further strengthen our technology and product competitiveness and solidify our leadership in the AI market by delivering products with the highest quality and performance.

With that, we are now ready to take your questions.

[Interpreted] [Operator Instructions] The first question will be provided by Hyun Kim from Meritz Securities.

[Interpreted] This is [indiscernible]. It was mentioned that the HBM supply negotiations for 2026 have been completed. Could you share more details about the contract?

[Interpreted] Thank you very much for the question. We understand that there has been broad and deep interest in the HBM contract for next year. This year, in particular, has been challenging in fixing not only the supply volume but also the product mix. due to various external factors. Not only that, there have been changes in the performance requirements for HBM products which necessitated longer time in, including the, in longer time in discussing the supply contract expected.

That said, our discussions over major issues with our clients have been completed and the HBM supply plan for next year for major customers has been finalized.

Given the explosive growth in demand for HBM to keep building infrastructure and the company's product competitiveness, the company's HBM has been selling out since 2023. And the pricing, which I'm sure is the point of interest for many has also been formed at a level that can sustain the current profitability.

And As the HBM demand continues to accelerate, driven by longer-term growth trend in the AI market. The company believes that it will be unlikely for supply to catch up with demand in a short period of time, which means that the pace of HBM's growth will be determined by supply capacity, and the company's HBM is positioned for a much higher growth than conventional DRAM products.

The company's HBM supply will remain tight compared to demand into 2027, but we will continue to do our best to supply products that meet customers' needs in a timely and secure manner.

[Interpreted] The following question will be presented by Jong Wook Lee from Samsung Securities.

[Interpreted] Congratulations on the performance. This is [indiscernible] from Samsung Securities. My question is on HBM. So it was mentioned that there have been higher performance requirements for HBM4. And my understanding that is that it was to be higher than the specifications. So have there been any difficulties for the company in meeting such higher performance requirements?

And also for the HBM4E, does the company expect the performance requirements to be similarly high to be higher than the specifications?

[Interpreted] Thank you very much for the question. Now with the AI inference market growing, the memory bandwidth is increasingly seen as the key factor that can upgrade AI performance.

As for HBM4, the number of IOs is already fixed at 2048, double the number of HBM3E, so the customers are now looking at higher speed as a way to increase the HBM bandwidth.

Based on our #1 technological competitiveness in HBM, the company is already fulfilling top-level specifications required by our customers. Moreover, we have already sampled products that meet customers' upgrade requirements faster than anyone in the industry and already started production for mass supply.

And as explained earlier, with intensifying competition for AI chip performance, the memory wall phenomenon becomes more pronounced where the memory performance becomes debottleneck for technological development. As such, performance requirements for next-generation memory products, including HBM, will continue to be upgraded.

[Interpreted] With the industry-leading design capability and know-how as the primary supplier, the company will respond to customers' requirements in a timely manner for our next-generation product line as well and maintain our #1 supplier position. Thank you.

[Interpreted] The following question will be presented by Simon Woo from Bank of America.

[Interpreted] This is Woo Dong-je from Bank of America. And congratulations on the recent performance, KRW 10 trillion in operating profit in the quarter.

Now my question is about the memory cycle. So what we have seen in the past is that usually, if there is a boom in the memory cycle, then it will be followed by a downturn. And so I wonder whether the company sees any similarities with the recent memory boom with the historic cycles in the past. It appears as if recently, the memory boom is also driving up demand for conventional memory.

And also, in terms of the inventory level, so there was a also cloud-driven boom some time ago. And at that time, so how does the company see the inventory level from then and today.

[Interpreted] Thank you very much for the question. Now it is true that the memory market this year has entered into what can be called a super boom cycle with surge in demand across all products unlike earlier expectations. Such changes have only recently appeared, but the company sees this cycle to be a bit different from the super cycles that we witnessed in 2017 and '18.

The biggest difference is that the current demand is driven by a much broader range of applications coming from the shift to the AI paradigm.

AI creates upside to the overall demand as it is being added on top of existing applications. At the same time, for the longer term, it is also creating new applications like autonomous driving and robotics AI. So what we are seeing is a fundamental shift in the demand for memory driven by AI.

In particular, computing recently has expanded to inference, promoting demand for not just AI servers but general purpose servers as well. The company believes that our total service set shipment -- total servers that shipment next year will grow at a high 10% level and to serve a DRAM to lead the overall demand for conventional DRAM.

At the same time, looking at the supply side, production can only grow so much even if we use more clean room space and capacity because of the growing share of HBM.

These circumstances create a structural constraint against a supply increase in the DRAM industry and are likely to serve as the driver for a long drawn out memory super cycle.

[Interpreted] The following question will be presented by Young Ho Ryu from NH Investment & Securities.

[Interpreted] First of all, congratulations on the performance. Now my question is on NAND. The recently strong demand for eSSD was set to be a structural change following the advent of the AI era. So could you elaborate more along the rationale behind such assessment?

[Interpreted] Allow me to explain more in detail the background to the higher NAND demand that we are seeing recently. First, there is a stronger build demand for both AI servers and general purpose servers with our server customers expanding their investment in AI, which, in turn, is driving demand for TLC products.

At the same time, demand for storage is also accelerating as a result of growth in AI-generated data like images and videos, leading to HDD supply shortage. So for hyperscaler customers with high dependence on HDD, the recent developments have prompted them to turn instead to eSSD based on high-capacity PLC.

Having said that, the company sees the recent change in demand as something that goes beyond the current short-term supply-and-demand issues, and we see this actually as one that can potentially change the demand. In other words, it is one that can potentially structurally increase the eSSD demand.

First, with the ever-advancing AI inference, the importance of RAG or the retrievable augmented generation structure is becoming even greater as a way to overcome the limitations of the existing LLM.

RAG moves beyond the current LLM approach, which basically generates responses based solely on the data that it was trained on. It will surge related documents in external databases and generates the final response based on such surge, which allows it to refer to the latest data as well as user-specific data, resulting in responses with much greater accuracy.

To apply RAG on LLM, we need to additionally build outside databases that express and store data as vectors. In other words, we need better databases. And this is where eSSD becomes a must to enable speedy search of data.

To support the scaling up of vector database and performance upgrade in RAN, demand for storage based on high-performance TLC and high-density TLC eSSD is expected to rise.

In addition, there has been a spike in data in the processing that is needed for the inference process, which led to the need to offload part of the key value cache that was generated at the GPU level to the lower layer memory. This is because data processing and power consumption surged during the insurance process, which calls for more efficient operation of the AI system.

By offloading key value cache that was processed at the GPU all the way to the SSD, depending on the frequency of the data usage, they can increase the throughput per unit of power when providing inference to many users and reduce the response time per user. And this is one of the reasons why I use of high-performance TLC eSSD is expected to grow.

As AI utilization keeps spreading, so will the role of the eSSD, resulting in its higher content growth. So essentially, what we are seeing now is the benefits of AI infrastructure spreading from DRAM to NAND as well.

[Interpreted] The following question will be presented by Minsook Chae from Korea Investment & Securities.

[Interpreted] The memory market appears to be changing into a specialty market with order first, sell later approach, thanks to AI. On the field then, does the company see any differences from the past in your discussions or interactions with customers?

[Interpreted] Yes, it is true that in the memory market, some businesses have shifted to order first, produce later approach with the emergence of HBM marked by massive investment and long lead time.

In addition, with strong HBM demand coming from our customers, the company was able to secure visibility into the customers' demand from the contracting stage with long-term agreements and respond with consistency.

For both the memory industry and the company, this has led to greater market predictability and business stability than in the past when it was much more volatile.

And the custom HBM will gradually increase from HBM4E. So the products will be developed in close collaboration with customers from the early stage of design of the customers' GPU or [ AZ ] products unlike in the existing standardized HBM. This will lead to much longer term and strategic transactions between customers and a small number of suppliers, contributing even more to business stability and profitability improvement on the part of the memory suppliers.

[Interpreted] Now if I may add, memory companies are allocating capacity to ramp up HBM supply. And this has led to supply constraints in conventional memory, resulting in supply shortage of conventional memory for which demand is actually growing.

And as a result, we are seeing an increase in customers who want to sign long-term agreements for conventional memory products as well. Some customers are very much actively responding to the current supply shortage by issuing prepurchase POs for 2026.

Now given the customers' demand and the company's capacity for next year, not only HBM, but DRAM and NAND capacity has essentially been sold out.

The company will try to respond to customers' demand with optimum production and sales strategy. And as mentioned earlier, we will keep discussing the implications of the HBM-driven changes with our customers.

[Interpreted] The following question will be presented by Ricky Seo from HSBC.

[Interpreted] Congratulations on the performance. And my question is on CapEx. Now recently, the investments by global AI companies point to a very high investment level needed for the next few years to fulfill the demand for memory. It was mentioned that the company's CapEx in 2026 will increase over this year. Then what will be the extent of the increase is the first question? And then also for the longer term, it appears as if the investment or the CapEx into the new campuses like Join and others will also have to be far higher than 1 year ago?

[Interpreted] Now as global AI companies competitively expand investment with conviction in the growth and monetization of the AI market, there has been accelerated growth in demand for a wide range of memory products, including HBM, DDR5 and enterprise SSD.

So to respond to such surge in demand, CapEx growth across the memory industry appears to be inevitable. And for the company, CapEx next year will far outpace the level of this year.

For M15X, equipment installation has begun in earnest to be used to ramp up supply of HBM. For conventional DRAM and NAND, we will accelerate tech migration in the existing capacity as a way of responding to the demand.

At the same time, considering the Fab 1 construction in Yong-in and preparation for construction of an advanced package plant in Indiana, U.S., investment in infrastructure is set to keep growing next year. But even with growing, the company will stick to its CapEx discipline and maintain a stable financial structure.

[Interpreted] The following question will be presented by Bo Young Choi from Kyobo Securities, question.

[Interpreted] My questions are on the product and technology. Now first, it was mentioned that there will be conversion to 1c nanometer next year, for which there is a line up for all products. and then also increase in the portion of the 321 layer NAND products. And what is going to be the time line for the ramp up for each product? And then also another question is what is the expected portion of the respective products by the end of next year?

[Interpreted] Thank you for the questions. Now under the principle that we respond with priority to demand with high visibility and profitability, our new capacity next year will center on HBM for which supply contract has already been completed. For DRAM and NAND, we plan to respond to demand through tech migration in the existing capacity.

Or DRAM 1c nanometer, development was completed next year with mass production beginning this year. Ramp-up will begin in full swing next year, and 1c nanometer is planned to take up over half of the conventional DRAM capacity inside Korea by the end of next year.

Based on the 1c nanometer process with the best performance and cost competitiveness, we will build up the lineups for all products, including DDR5, LPDDR5 and graphic DRAM products to respond to customers' demand in time and ensure profitability.

And in the case of NAND, our focus remains on improving profitability. And the plan is to keep improving profitability through tech migration rather than ramping up capacity.

That has been the case this year with tech migration from 176 layer to 238 then to 321. Next year, we will grow our supply not only in TLC but also in QLC, which will require ramp-up of 321 layer products. This means that we are making the preparation for 321 layer products to take up more than half of our NAND bit production by the end of next year.

[Interpreted] The following question will be presented by Dong Hee Han from SK Securities.

[Interpreted] And my question is on the inventory level. there has been considerable inventory sell-down in the second quarter. And again, inventory appears to be much lower in the third quarter as well due to the very strong demand. So what is the company's inventory level now and also among the customers?

[Interpreted] With customers demand outpacing expectations in the previous quarter, there have been concerns over excessive inventory buildup in the memory supply chain as well as demand slowdown consequently. But customers' inventory level has become lower overall with accelerated [indiscernible] build, and added to that, investment in AI infrastructure has continued to grow, resulting in noticeably lower memory inventory among server customers.

And for the company, the inventory level has also fallen Q-o-Q in both DRAM and NAND as a result of the recent strength in memory demand. This is particularly true for DRAM inventory, which remains extremely low, so much so that in the case of DDR5, the products must be shipped to customers straight out of production to ensure timely response.

The company will continue to try to maintain a healthy inventory level for both DRAM and NAND to seamlessly respond to customers' demand.

[Interpreted] The following question will be presented by Nicolas Gaudois from UBS.

Regarding that you mentioned you opening earlier, are you able to address the faster ramp up pulling in your equipment delivery schedule for fab, first of all? And in that regard, could you more or less complete full equipment installation for a total vertical capacity for M15x by the end of 2026? And then is it possible for you to pull in the schedule as well for Yong-in fab on clean readiness, which I think initially had been down for May 2027?

[Interpreted] Thank you very much for the questions, and allow me to respond. Now let me respond to the question about the company's plan for the fabs.

The company had decided at the end of 2023 to make new investments in M15X. To preempt the fast rising demand for HBM, which requires relatively bigger wafer capacity.

After around 2 years of construction, the fab finally opened early in M15X a while ago with equipment installation starting. We are now making the preparation for M15X to contribute to HBM production ramp-up starting next year.

And as the memory demand growth continues to accelerate much faster than expectation, we are also speedily moving ahead with the capacity ramp-up at M15X.

As for the Fab 1 in Yong-in that just started construction this year, we are working to pull up the schedule in light of the pace of demand growth and the earlier ramp-up at M15X.

The company will keep trying to preempt capacity and fab space by building state of the art production infrastructure from M15X to Yong-in fab to enable flexible response to the ever-growing AI memory demand.

[Interpreted] The following question will be presented by S.K. Kim from Daiwa Securities.

[Interpreted] Thank you very much for taking my question. It is on demand. Now there have been a series of announcements of GPU and ASIC supply cooperation between Big tech and AI companies, building expectations of further AI market growth. Then against this backdrop, what is the company's outlook on HBM demand growth as well as a broadening of the customer base?

[Interpreted] Thank you for the question. Now with upward adjustment in big tech's CapEx and increased investment by AI companies, the HBM market, even by a conservative estimate, will keep growing at an average of over 30% for the next 5 years.

I will point to our recent LOI with open AI for a large-scale DRAM supply as an example of the very strong market demand for AI as well as the need to secure AI memory based on HBM more than anything else when developing AI technology.

As the primary supplier of not only GPU, but ASIC for many customers, we are working with customers in the development of next-generation products and contributing to the development of the AI industry based on our differentiated product competitiveness and mutual trust.

Thus, we are positioned to maintain a high share in the newly arising HBM demand among a broad range of customers and keep increasing our supply.

[Interpreted] The following question will be presented by Jay Kwon from JPMorgan.

[Interpreted] And my question is on DRAM. The DRAM spot price is rising almost daily. And with the DDR4 and DDR5, almost in supply shortage as seen through the price premium. Of course, I'm sure that the company's contract price with the customers would be different from the spot price. But if the DRAM price maintains the current trend, it appears likely to reach the kind of profit margin similar to HBM. So what is the company's view on the DRAM profitability? And even if temporarily, it surpasses HBM's profitability, is it also foreseeable to shift the capacity mix, perhaps a bit away from the HBM to the conventional DRAM?

[Interpreted] Now it is true that the recent hike in DRAM price has narrowed the profitability gap between HBM and DRAM. But for the company, HBM's profitability remains high.

If supply remains tight next year, DRAM's margin could rise closer to HBM, but the company does not plan to immediately adjust the capacity mix based on what can be a short-lived change in profitability.

Given the nature of HBM products, it is important to agree on the long-term volume with customers to make sure that there is a seamless supply.

And when we discuss long-term volume with customers, we also consider various factors like customer relations, long-term growth potential as well as profitability.

Now having said that, there have also been discussions over stronger binding contracts for conventional memory products as well with customers issuing prepurchase POs or asking for multiyear LTAs. And our decisions on capacity mix will be made in a way that can ensure optimum productivity.

The company also sees the current trend to potentially prompt changes in the nature of memory business for the future.

As the leading supplier of AI memory, the company was able to improve our fundamentals in memory business based on the high and stable profitability from HBM and have achieved differentiated performance. Looking ahead, we will keep responding to customers' demand with a long-term view and achieve sustained growth along with the AI market.

[Interpreted] The last question will be presented by Peter Lee from Citigroup.

[Interpreted] Now as was explained several times because of the good performance in the market as well as the growth in the AI market, good performance for the company has also been achieved and is expected to continue for some time. So as a result, the company has turned around to net cash position this quarter, and also, its FCF is expected to continue to improve on the back of much stronger performance. So I realize perhaps this is a bit early, but then can we expect any changes to the shareholder return policy that was announced in the early part of this year?

[Interpreted] Yes, it is true that the company's financial soundness is fast improving, thanks to the stronger-than-expected performance in 2025. And yes, we have achieved net cash position in Q3 following the higher recovery of receivables with sales growth in Q2.

It was explained as part of the current shareholder return policy that the company's aim in financial soundness is to maintain an appropriate level of cash that would allow us to keep stable business management through the differing industry cycles and to execute CapEx that is necessary to maintain our competitiveness.

Now the recent upturn in the memory market has fueled demand growth, which, in turn, is driving up CapEx necessary to fulfill the demand. So this means that what we see as appropriate level of cash also has to reflect this change.

Furthermore, considering the huge growth potential in the AI memory market and the company's high return on investment, I believe that the shareholders will also agree that for now, the best use of cash is to reinvest it into our business while maintaining CapEx discipline.

[Interpreted] We apologize, but we are experiencing a brief technical issue. Please stand by.

[Technical Difficulty]

[Interpreted] As such, being in the first year of the new shareholder return policy that is announced at a 3-year interval, we are not looking into additional shareholder return at this time, but we will keep looking into how we can maximize shareholder return by taking a comprehensive look into the changes in the environment, both inside and outside, such as market outlook and investment needs.

[Interpreted] thank you very much, and that concludes the SK hynix 2025 Third Quarter Earnings Release Conference Call.

[Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]

Good morning, afternoon and evening. Thank you for joining SK Hynix 2025 Q2 Earnings Release Conference Call. [Operator Instructions] With that, we are now ready to begin.

Good morning, afternoon and evening. This is Park, Seong Hwan, Head of IR at SK Hynix. Welcome to the SK Hynix 2025 Second Quarter Earnings Release Conference Call. Allow me to introduce the executives present here with me today. We're joined by President of Corporate Center, Song Hyun Jong; Chief Financial Officer, Kim Woo-Hyun; Head of DRAM Marketing, Kim Kyu Hyun; Head of NAND Marketing, Kim Tae, and Head of HBM Sales and Marketing, Kim Jeong Tae.

Let me issue a disclaimer that all outlooks presented by the company are subject to change depending on the macroeconomic and market circumstances. With that, we will now begin SK Hynix's earnings release conference call for second quarter of 2025. Song Hyun Jong, President of Corporate Center, will first present the earnings followed by the company's future plans and market outlooks and a Q&A session with the attending executives.

Good morning, everyone. This is Song Hyun-Jong, Head of Corporate Center. Allow me to first introduce the company's performance for the second quarter of 2025. The past quarter began with concerns about a slowdown in demand due to trade tensions and overall economic uncertainty. However, demand for AI memory continued to grow, driven by aggressive AI investments from big tech companies. This was further supported by customers' preemptive purchasing to prepare for external risks, resulting in a more favorable environment than initially expected. Both DRAM and NAND shipments exceeded our original guidance and pricing conditions also improved.

As a result, second quarter revenue increased by 26% Q-o-Q and 35% Y-o-Y, reaching an all-time high of KRW 22.2 trillion. For DRAM, sales of our 12 High HBM3E product was expanded in full scale as planned, while server and PC segments saw strong demand, driving sequential Q2 shipment growth above guidance at mid-20% growth. Meanwhile, ASP rose by low single-digit percent due to a considerable increase in sales of relatively lower-priced conventional DRAM products. Coming from a low base of the previous quarter, NAND shipments significantly exceeded guidance by increasing over 70% sequentially as demand rose across all applications.

ASP declined by high single-digit percent Q-o-Q due to continued pricing weakness in solution products despite price rise in the spot market. Operating profit rose by 24% Q-o-Q and 68% Y-o-Y, reaching KRW 9.2 trillion, also setting a new quarterly record. The operating profit margin stood at 41%. Depreciation and amortization expenses for Q2 were KRW 3.4 trillion, marking a slight increase from the previous quarter.

EBITDA came in at KRW 12.6 trillion with an EBITDA margin of 57%. Nonoperating loss net of gain reached KRW 0.49 trillion. Notable items include foreign currency-related net loss of KRW 0.61 trillion due to depreciation of U.S. dollar against Korean won and KRW 0.12 trillion of other nonoperating income, including valuation gains on investment assets.

Pretax income was KRW 8.7 trillion, while net income was KRW 7 trillion, resulting in a net profit margin of 31%. As of the end of the second quarter, cash and cash equivalents stood at KRW 17 trillion, up KRW 2.7 trillion from the previous quarter. Interest-bearing debt decreased KRW 1.5 trillion to KRW 21.8 trillion, resulting in a reduction of net debt by KRW 4.1 trillion to KRW 4.9 trillion. Accordingly, our debt-to-equity and net debt-to-equity ratios reached 25% and 6%, respectively, which is an improvement by 4 percentage points and 5 percentage points, respectively.

Now, let me share our market outlook. While the memory market was initially expected to recover towards the second half of the year, we saw strong demand growth and favorable pricing conditions in the first half. Despite increased procurement demand during Q2, system build demand also rose, helping customers to maintain inventories at stable levels.

Given demand is expected to grow from new product launches in the second half, we believe the likelihood of sharp demand corrections due to customers' inventory adjustments would be low. Looking at demand by application, we expect healthy demand growth in the server market, driven by increasing investments from big tech companies despite macro uncertainty.

Alongside strong AI demand, general-purpose server demand is also projected to rise due to replacement cycles and adoption of new CPUs. In the AI market, big tech companies have begun full-fledged competition by unveiling AI agents with enhanced reasoning models. These models perform iterative thought processes to solve problems autonomously, requiring systems capable of processing significantly more data than conventional AI models.

This evolution is expected to further drive demand for high-performance, high-density memory. Additionally, ongoing investments by governments and corporations for sovereign AI are likely to become a new long-term driver of AI memory demand. In the PC and smartphone market, the proliferation of AI applications is expected to drive replacement demand this year. Memory content per device are also expected to grow with wider adoption of high-performance memory required to support the AI functions.

Unlike DRAM market, NAND market is yet to see meaningful AI-driven demand growth. However, as power consumption considerations are becoming increasingly important in AI servers, we expect adoption of power-efficient SSD solutions to widen in the near future.

Next, I will discuss the company's plans. In the third quarter, given the significant shipment growth in the second quarter, we expect a moderation in shipment growth with DRAM growing by approximately low to mid-single-digit percent and NAND shipment increase being rather limited. Our plan to double HBM sales Y-o-Y remains unchanged, backed by our strong product competitiveness and mass production capabilities. We expect to continue delivering stable performance, leveraging the HBM business despite external uncertainty.

Our HBM3E products are receiving high marks from all customers for their performance and supply reliability, and we plan to maintain industry-leading competitiveness as we transitioned into HBM4. In March, we supplied industry's first samples of HBM4 to customers. We are currently working closely with partners to optimize system-level performance and are preparing to deliver products on time for our customers' needs.

Beyond HBM, we plan to expand sales of conventional DRAM products where DRAM is increasing. These include high-speed DDR5 products over 8,000 megabits per second that support new CPUs, high-density server modules over 128 gigabytes and LPDDR5X for flagship smartphones as well as LPDDR4X products favored in the Chinese market. We also plan to begin supplying LPDDR-based server modules that enhance data processing speed and power efficiency within this year.

Additionally, for GPUs, we are preparing not only the 16-gigabit GDDR7 products currently in supply, but also 24 gigabit-based products to strengthen our leadership in the AI memory market with a diversified product portfolio. For NAND, we will maintain a prudent investment strategy and a profitability-focused approach in line with downstream demand conditions. However, we are continuing product development efforts to respond quickly once market conditions improved. In the second quarter, we began actively expanding sales of ultra-high-density enterprise SSDs exceeding 120 terabytes based on QLC technology.

In May, we developed a 321-layer NAND-based USS 4.1 product. with random read and bright speed improved by 15% and 40%, respectively, compared to the previous generation. By the end of the year, we also plan to develop both client and enterprise SSDs based on 321-layer technology, thereby building a competitive NAND product portfolio. Looking ahead, our M15X fab is scheduled to open in Q4 of this year to produce DRAM products, including HBM from 2026.

Construction of our Yong-in fab Phase 1 is on track for completion in Q2 2027. And through these facilities, we aim to secure the manufacturing infrastructure necessary to meet future demand. Our total investment for this year is expected to increase compared to the previous plan in accordance with our principle of continuing our investment focus for products with high demand visibility and secure profitability as well as maximizing investment efficiency, although we are yet unable to provide details regarding HBM supply for 2026 as we now have secured visibility on next year's demand figure based on discussions with customers, we assessed that some proactive investment this year is necessary to ensure timely support of our HBM products.

In the AI market, memory is increasingly becoming a critical infrastructure as it plays an essential role in defining overall AI system performance. As a leading AI memory provider, we aim to drive both customer satisfaction and market growth by launching innovative products with top-tier quality and performance that meet the needs of the AI system.

With that, we are now ready to take your questions.

[Interpreted] [Operator Instructions] The first question will be provided by Dong Hee Han from SK Securities.

[Interpreted] I am from Han Dong Hee from SK Securities. Congratulations on your great performance. My question is about forecast for HBM market. Driven by strong demand for AI, HBM is expected for continuous and rapid growth. How do you forecast to meet the long-term demand for HBM beyond 2025? And what will be important momentums to drive HBM demand?

[Interpreted] As you can see from continuous CapEx expansion by big tech companies, increases in token throughput and high growth rate of AI start-ups, the AI market is growing at a rapid pace. In particular, AI demand is expanding from training to inferencing and inferencing has further divided and advanced into areas like reasoning and agents, which has led to a sharp increase in AI workloads and intensifying bandwidth bottlenecks.

In this rapidly growing AI market, HBM is now positioned as a key product that quickly affects performance enhancement. Taking its importance into consideration, we believe there is no doubt about growth potential of HBM demand. As the AI market continue to expand into areas like AI agents and physical AI, computational volume will grow explosively, which will, in turn, drive demand growth in the HBM market.

In the future, the HBM market may not sustain the early stage explosive growth rate, but AI technology will rapidly advance and spread, expanding the customer base and driving the continuous launch of new products and services. Therefore, we expect HBM will continue to enjoy strong growth.

[Interpreted] The following question will be presented by S.K. Kim from Daiwa Securities.

[Interpreted] Congratulations on your great performance. My question is also about HBM because market attention is really centered around this product. So my question is about the next year forecast. So how are negotiations progressing for the 2026 HBM supply with major customers?

[Interpreted] Thank you for the question. We cannot provide customer details, but negotiations are proceeding as planned. As customer projects become increasingly diverse, we are engaged in close negotiation for product mix and pricing.

As a close partner of our major customers in the AI semiconductor field, we are conducting negotiations and offering optimal supply conditions for a long-term partnership. Furthermore, by delivering the highest quality products that customers require in a reliable and timely manner, we plan to prepare for supply expansion to meet customer demands and grow together with our customers.

[Interpreted] The following question will be presented by Young Ho Ryu from NH Investment & Securities.

[Interpreted] So I am Young Ho from NH Securities. My question is about demand. And so to what extent the uncertainties affect pull-in demand in Q2? And how do you assess the resulting increase in customer inventory and its impact on demand for second half?

[Interpreted] Our Q2 shipment volume significantly exceeded the initial guidance, and this was indeed driven by purchasing demand, which was stronger than expected. Initially, customers intended to conservatively maintain inventory levels during the first half of the year, but with growing uncertainty around tariff policies, they shift the strategies towards securing appropriate level of inventory. This led to a great increase in purchasing demand, especially from major customers with previously low stock levels.

In addition, with end-of-life plans released for certain legacy products, competition for securing supply intensified among small to midsize customers, including module makers and therefore, further tightening market supply condition. Although shipment volumes in the first half exceeded our original plan, which raises concerns over potential in slowdown in the second half, we believe the likelihood of sharp fluctuation in supply and demand remains low.

System build activity also increased in Q2 so customer inventories levels did not rise to a significant level. Moreover, memory suppliers have already reduced their inventories greatly, which means going forward, growth will be only driven by product -- supply growth will be only driven by production growth. Depending on how future tariff policies evolve in the future, purchasing demand can be affected. In response, we will do our best to ensure stable business operation, especially focusing on products with secured demand visibility.

[Interpreted] The following question will be presented by John Kangho from Daishin Securities.

[Interpreted] So once again, congratulations for your great performance. With the U.S. government tightening export controls on China, SK Hynix may face growing restrictions on its fab operations in China. So what are your plans for managing your operations in China going forward?

[Interpreted] Thank you very much for the question. As we all worry about geopolitical situations today, however, there has been no change to our validated end user or VEU status that we obtained in October 2023. We are committed to putting our customers' needs first and we'll work hard to provide uninterrupted supply within the boundaries of regulations, and we'll continue operating our China fabs according to our existing plan.

In fact, a recent supply conditions for legacy DRAM can be beneficial for our China fab operations. For example, with increasing DRAM capacity being allocated to expanding HBM production, supply shortages of DDR4 and LPDDR4 legacy products have emerged during the transition to DDR5 and LPDDR5.

So in Q2, we responded to this demand growth with our -- for the legacy product with our existing inventory, but we believe that the demand for legacy DRAM will continue to persist for a while. And therefore, we will also focus on providing legacy products for the long-term customer support, and we are going to utilize our China fab to ensure a stable supply for the legacy product.

Our China fabs are not only critical to our operations, but also play a key role in the global memory semiconductor supply chain. As such, we will continue to closely monitor U.S. regulatory developments and maintain close communication with governments around the world.

[Interpreted] The following question will be presented by Jay Kwon from JPMorgan.

[Interpreted] My question is about NAND products. It seems your NAND shipment growth in this quarter significantly outperformed prior guidance. Can you please elaborate on which customer segment drove this growth?

[Interpreted] In the previous earnings call, we guided for over 20% NAND shipment growth in Q2. At the time, we observed the signs of rising spot prices and declining customer inventories, but demand visibility for certain products remain fairly limited.

However, in Q2, demand for enterprise SSDs increased as hyperscale customers expanded their AI investments while pulling demand for individual components also rose due to tariff-related uncertainty. Additionally, mobile set build demand in the China region grew due to promotional event leading to shipments that significantly exceeded our expectations. With the Q2 sales increase, our inventory levels have now normalized. Moving forward, we'll continue to manage production and inventory flexibly in accordance with market demand.

[Interpreted] The following question will be presented by Hyun Kim from Meritz Securities.

[Interpreted] I would like to ask about HBM4 as HBM4 entails a greater cost increase compared to previous HBM products. There are some concerns about potential profitability decline. What is SK Hynix' perspective on this?

[Interpreted] As you may know, HBM4 exhibits major technological changes, including increased I/O count for higher bandwidth, new designs for improved power efficiency and the adoption of logic processes in the baseline.

Accordingly, we are striving to reflect this cost increase in our pricing strategy for HBM4, and we aim to stimulate the AI market by establishing optimal pricing with customers while maintaining current profitability levels.

[Interpreted] The following question will be presented by Jong Wook Lee from Samsung Securities.

[Interpreted] My question is about DRAM. So DDR4 supply appears to be quite tight recently. What is your outlook on future DDR4 demand? And what is DDR4's current revenue share at SK Hynix? And what are your plans regarding its end-of-life timing and future operations?

[Interpreted] The recent spike in DDR4 prices has been primarily driven by short-term supply concerns following announcements from some suppliers to discontinue DDR4 production. However, as we are currently in a transition phase from DDR4 to DDR5, we view this as a temporary demand concentration rather than a structural shift in underlying demand for DDR4.

Speaking of the DDR4 share in SK Hynix' revenue, since 2023, DDR4 has seen a steady decline in our share -- in our revenue share triggered by growing demand for DDR5 and HBM. In fact, its contribution has dropped from a double-digit percentage last year to a single-digit level this year.

Going forward, we plan to phase out DDR4 for the mass market. However, for selected customers who require long-term support, we'll continue our supply based on mutually upgrade volume.

[Interpreted] The following question will be presented by Peter Lee from Citigroup.

[Interpreted] So my question is about CapEx investment. You mentioned that this year's investment will exceed initial plans. So to what extent will it increase? And will most of the additional investment be allocated to equipment or infrastructure?

[Interpreted] We are pursuing investments for future growth in tandem with efforts to maintain financial soundness. This year, our investment focus is on ensuring seamless support for the supply volumes agreed upon with customers. At the same time, infrastructure investment, such as M16X and the new fab in Yong-in are underway to lay the foundation for long-term growth.

As mentioned earlier, we now have visibility into next year's HBM supply, which calls for proactive investment to ensure a timely response. Accordingly, our total investment for this year will increase from the original plan with the majority of the additional spending allocated to HBM-related equipment. The final investment scale, however, will be confirmed once negotiations with major customers are concluded.

So we'll continue to comply with CapEx discipline while improving investment efficiency going forward.

[Interpreted] The following question will be presented by Ricky Seo from HSBC.

[Interpreted] So a recent article suggests that M16 is already operating at full capacity with no additional capacity for production expansion. In that case, to what extent do you expect the new M15 fab to be utilized by the end of next year given the current investment level? And my second question is that as for the 1C Nano DRAM ramp-up, when will we be able to see the significant level of the expansion at least by 15% to 20%?

[Interpreted] So let me first answer the first part of your question regarding the new fab. As the HBM market expands and demand increased, we have reallocated a significant portion of our capacity from general DRAM to HBM. In this process, due to HBM's nature of requiring more capacity compared to general DRAM, the cleanroom space needed to produce the same output volume has increased substantially.

So our existing fabs are currently being utilized for product mix optimization, technology migration and TSV line operations. At the same time, to secure mid- to long-term production infrastructure, we are simultaneously developing the M15X, Yong-in fab and advanced packaging fab in Indiana, the U.S.

Among these, the M15X is scheduled to open in Q4 of this year and will begin full-scale mass production next year. We will primarily focus on next-generation HBM products to meet rising customer demand. We plan to gradually ramp up capacity in line with the visibility that we gained on customer volume requirements throughout the next year.

And also for your question on that, how much space can actually be utilized, where space increase can be realized through the M15X at the end of next year, I understand your intention for the question, but I believe there will be no cadence where we will not be able to supply product because of the fab-based restrictions.

And also for the conversion into the MC nano, I believe this conversion will begin from the second half of this year, but the full-scale conversion will take place only in next year. Right now, we are in the process of establishing management plan for this. So we will share the details later.

[Interpreted] The following question will be presented by [indiscernible] Kim from Hanwha Securities.

[Interpreted] So the long-term growth potential of the NAND market has slowed down compared to the past. Despite suppliers, maintaining a conservative production stance, profitability has not shown significant improvement. How long do you expect this market condition to persist?

[Interpreted] So the slowdown in the NAND market today is mainly due to stagnant demand for consumer devices as well as the fact that customers transitioning to AI servers have yet to extend their investment into storage devices. However, we believe that the need for storage equipment investment will begin to surface in the not-too-distant future.

As HBM director mentioned at the beginning of Q&A session, this is because AI units and the results in volume of token generation is expected to grow at an astronomical pace far beyond what we can imagine today. And traditional data processing methods will not be able to handle this exclusive surge in demand.

That's why there is a growing attempt to offload AI inference data caching to eSSD. And as that, we expect initial demand to materialize within the next 2 to 3 years. If this trend becomes a reality, eSSD will no longer remain as a mirror storage device or whether it will evolve into part of the compute cash and their roles and positions within the AI system will be changed. And we expect this will mark the beginning of significant demand growth in the NAND market.

[Interpreted] The following question will be presented by Simon Woo from Bank of America.

[Interpreted] So my question is about technologies for the next-generation DRAM production. So when will they be introduced? And what will be the timing for mass production, and how they can contribute to improving efficiency for HBM DRAM production?

[Interpreted] So while memory suppliers have thus far pursued the transition to finer process nodes, their efforts are now approaching the limits in terms of performance and capacity enhancement. And so in response, we are exploring the application of vertical gate platforms and 3D DRAM technologies based on breakthrough changes in architecture, materials and components of sub-10-meter -- 10-nanometer nodes.

3D platform is a next-generation memory technology that minimizes DRAM cell area and vertically structured the gate that function as a transistor switch, thereby dramatically enhancing area efficiency and power characteristics compared to existing architecture.

And 3D DRAM is another next-generation memory policy that enables high density by vertically stacking DRAM cells without miniaturization. And in order for the second local innovation to overcome current limitations and become the new industry standard, we are focusing on securing technological stability.

However, as it will take time for such innovative approaches to become technologically stable and ready for mass production. So we'll continue to advance conventional miniaturization effort such as 1bnm development to maintain our technological leadership.

As you know, our competitive adds in HBM comes from our strong foundational technology for DRAM. And, therefore, if we could -- so as we step-by-step, develop the next-generation memory technology, I believe, this will also be the source of our future HBM production.

[Interpreted] The following question will be presented by Yeong-Min Ko from DAOL Investment & Securities.

[Interpreted] So there have been reports that a major GPU customer has resumed shipments of AI chips with performance just for the export to China. And I believe that this could also be beneficial for the future supply expansion and so on. So can you -- could this lead to an upside in HBM demand?

[Interpreted] So it was only recently reported that a major GPU customer has resumed shipment of AI products bound for export to China. We are currently in the process of assessing concrete information regarding how HBM demand will shift in relation to this product. However, as we were a primary HBM supplier for this product prior to the export control, we believe that if customer demand and our internal supply conditions are aligned, we are well positioned to respond swiftly and capitalize on this opportunity. And leveraging the close partnership that we have built with our customers, we will do our best to generate win-win outcomes for both our customers and the company.

[Interpreted] The last question will be presented by Minsook Chae from Korea Investment & Securities.

[Interpreted] Congratulations on your great performance. My question is back to HBM. SK Hynix has emerged as a leading HBM supplier backed by strong product technology and manufacturing capabilities. However, concerns continue to mount over intensifying competition and uncertainty due to changes in business environment with the AI semiconductor ecosystem. So as an HBM leading supplier, what is your strategy moving forward to navigate these challenges?

[Interpreted] Right. So it is very natural to worry about intensifying competition, especially given the situation that the number of suppliers are growing. But I believe that as long as we are in the memory business, competition is something that we have to take. And I believe this is our mission to deliver good performance despite the fierce competition as well as provide great values to our customers and investors.

So with the emergence of the HBM product, I believe memory market has also shifted from the pure computing stories, and we believe that this can also affect, and as a result, the leading suppliers now has greater leveraging power. HBM influences over computing chips for AI, where HBM is directly employed as well as broader AI computing system performance and cost basis, far greater than conventional general memory products. As a result, the demand for HBM has become very sticky. And leading players now have greater advantage in having all the engagement with customers.

SK Hynix, despite its previous minimal presence in the HBM market, we have been able to grow as a leading player today, and I believe this comes from our corporate culture. We put customer before anything else to address customer pain points in product design, production and supply. We are also backed up by strong teamwork. Such competitive adds in corporate culture cannot easily be innovated or mimicked by others. Going forward, based on such solid corporate culture, we'll make sure SK Hynix will continue to stay top of mind of HBM customers, and we believe this is the way to secure and solidify our leadership in the future in the AI memory market.

Due to various restrictions in AI computing infrastructure today, we are required to develop further granular HBM products as well as design new memory products. Through all engagement with customers, SK Hynix will continue to provide customized HBM products catering to the needs of customers in AI ecosystem and solidify our strong leadership in new memory products for AI, including PIM.

[Interpreted] This will conclude the conference call for the SK Hynix for the second quarter of 2025. Thank you very much.

[Portions of this transcript that are marked [Interpreted] were spoken by an interpreter present on the live call.]