Thank you for joining us for the SK hynix 2022 third quarter earnings release conference call. We're ready to begin. This conference call will consist of the presentation by the executives of SK hynix, followed by a Q&A session. To ask a question, press star one, that is star and one on your phone. The business results portion will be simultaneously interpreted. We're now ready to begin.
Good morning, and good afternoon and evening to those calling in from abroad. This is Park Seong-Hwan, the Head of IR at SK hynix. Welcome to the SK hynix 2022 third quarter earnings release conference call. Before starting the conference call, allow me to introduce the executives present here with me today.
First, Kevin Noh, Chief Marketing Officer, Kim Woo-hyun, Chief Financial Officer, Park Myeong-soo, Head of DRAM Marketing, and Park Chan-dong, Head of NAND Marketing. 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'll now begin SK hynix earnings release conference call. Kevin will first present the earnings for the third quarter of 2022, followed by the company's plan and market outlook. There will then be a Q&A session with the executives.
Good morning. Allow me to first report to you the company's performance in the third quarter of 2022. The third quarter saw deteriorating macroeconomic conditions due to high inflation and interest rate hikes. This led to a rapid decline in customers' memory demand, and the company's bit shipments of both DRAM and NAND fell quarter on quarter.
Prices also fell at a faster pace than expected, resulting in third quarter revenue of KRW 10.98 trillion, down 20% quarter-on-quarter despite the strong dollar. Third quarters generally see a pickup in market conditions with favorable seasonality. This year, the market environment remains weak with almost unprecedentedly soft demand. While demand for consumer products such as PCs and smartphones is weakening further, procurement demand from servers also softened as customers try to work down their inventories. The company tried to maximize its sales by expanding sales to mobile customers who introduced new models and to hyperscale customers who had relatively deeper pockets. DRAM bit shipment fell by a mid-single digit percent lower than previous guidance, and ASP fell by around 20% quarter-on-quarter.
For NAND flash, the company aimed to offset the weakness in consumer product demand by expanding sales of eSSDs, but the bit shipment fell short of the guidance, falling by low teens% quarter-on-quarter, including Solidigm sales volume, and by high single-digit% when excluding Solidigm. ASP also fell by more than 20%, both for SK hynix alone and for the integrated entity. As for profitability by product, the 1a-nm DRAM and 176-layer NAND products showed smooth ramp up during the quarter, reaching stable yield rates and expanding the sales share of these leading-edge products, which in turn led to unit cost reductions for both DRAM and NAND. However, such cost reductions were insufficient to offset the fast decline in prices. Operating profit in Q3 was KRW 1.66 trillion, down 61% from the previous quarter.
Operating profit margin was 15%, falling from 30% of the previous quarter. Depreciation and amortization in Q3 was KRW 3.57 trillion, increased by KRW 0.81 trillion quarter-over-quarter. EBITDA in Q3 was KRW 5.22 trillion, and EBITDA margin 48%. There was non-operating profit of KRW 27.7 billion. This includes net interest expense of KRW 121.3 billion and a net foreign currency related gain of KRW 132.7 billion that was driven by the transaction gain on accounts receivables and translation gain on Kioxia investments, despite a translation loss on foreign currency denominated debts due to the strong dollar.
Pre-tax income was KRW 1.68 trillion, with corporate tax expense of KRW 0.58 trillion. Net profit was KRW 1.1 trillion with a net profit margin of 10%. Consolidated cash balance at the end of Q3 was KRW 7.21 trillion, down by KRW 0.28 trillion from the previous quarter. Interest-bearing debt was KRW 22.02 trillion, increasing by KRW 2.64 trillion compared to a quarter ago. Debt to equity ratio and net debt to equity ratio at the end of Q3 was 32% and 22% respectively, rising from the previous quarter. Next is the company's market outlook and plans. Contrary to expectations earlier in the year, the memory market in the second half continues to struggle with demand contracting sharply.
Rising interest rates in major countries in response to inflation in combination with the strong dollar are heightening concerns of a recession, weighing down on both consumption and investment. Demand deceleration is being felt more severely in IT products due to the base effect of strong growth during the pandemic. As such, demand growth for DRAM and NAND for the year are expected to drop to unprecedentedly low levels, being low- to mid-single-digit % and around single-digit % respectively. Given the current environment, the company's bit shipments for DRAM and NAND in Q4 are expected to be similar to that of Q3. PC shipment growth, which had recorded high growth in the past 2 years, is now expected to be down by mid-teens %, with correction likely to continue into next year.
Smartphone unit shipment is also expected to fall by a high single-digit % due to lengthening replacement cycles and the impact of COVID-19 lockdowns. Given the economy-sensitive nature of consumer goods, demand for low to mid-end products are falling steeply, while demand for high-end and flagship models launched in the second half remains relatively solid. Thus, increase in memory content per phone is offsetting the slowdown in smartphone unit shipments. For servers, demand remains relatively healthy compared to other applications this year, but demand in the second half is being affected by investment reduction by corporates and inventory correction by the customers due to macroeconomic uncertainties. Nevertheless, growth in cloud services such as AI and big data analytics, as well as investment by big tech companies, are projected to continue. Thus, servers will continue to drive overall memory demand.
Although the ongoing uncertainties are making predictions increasingly difficult, bit demands of DRAM and NAND are estimated to grow by low-teens% and mid- to 20% respectively next year. Meanwhile, new applications such as AR and VR devices and content increases from price elasticity are possible upside factors for memory demand. Against the sharp demand decrease, memory suppliers have begun taking actions such as cutting back on CapEx and adjusting utilization rates. However, as there is a time lag for the impact to materialize, the current supply-demand imbalance is expected to persist for the time being. In light of such market environment, the company is aggressively cutting back on next year's capital spending to align supply with demand. This year's CapEx is likely to close at around high 10 trillion KRW, which is higher than that of last year.
Next year's CapEx will be reduced by more than 50% compared to that of this year. The rate of reduction is as aggressive as what the industry had back in 2008 and 2009 during the global financial crisis. Given the high levels of inventory expected throughout the industry at the end of the year, the company will minimize investment for wafer capacity and even look into delaying certain investments for tech migration. In order to effectively respond to the rapidly changing demand environment, the company is also planning to reduce production of relatively low margin products. Moreover, we are reconsidering our product mix and equipment allocation, which may sacrifice some of the wafer capacity in the near term, but will improve efficiency of our fab operations in the long term.
Thus, the company's DRAM and NAND wafer production next year will be reduced compared to this year, and the pace of migration to the leading-edge technology nodes will also be slower than what was originally planned. Yet, we will maintain investments that are critical to mass production of new products like DDR5, LPDDR5, and HBM3, which will be the drivers of the future demand growth so that we remain fully ready to support our customers' demand. Next, let me explain the company's technology roadmap. For the 1a-nm DRAM and 176-layer NAND, both yield improvement and ramp up are continuing smoothly. By the end of Q3, the 1a-nm accounted for around 20% of total DRAM production, and the 176-layer, 60% of total NAND.
In particular, the share of the 176-layer has gone up considerably in all applications, contributing to improvement in cost competitiveness. The company also completed the development of the 238-layer NAND in August this year. With the significantly improved NAND and SSD competitiveness, the company aims to close the gap with the market leader while seeking to respond to the upcoming market downturn by focusing on creating business synergy in collaboration with Solidigm. The DDR5 market is finally expected to grow in 2023 in terms of the ecosystem and market demand. This has been held back by the delay in a CPU launch, but that gave time for the ecosystem to prepare and for demand to be built up. The recent market situation has lowered the price burden as well, which is likely to stimulate faster transition to DDR5 by server customers.
The company will secure leadership in DDR5 market based on its full lineup of leading-edge 1a-nm products, including high density products like 3DS. Regarding the recent U.S. government's restriction on semiconductor equipment exports to China, the company has received a 1-year waiver for licensing requirements to move equipment into China. The company will work closely with relevant stakeholders, including governments, customers, and partners in different countries to prepare against external uncertainties and to ensure business continuity. Throughout this year, a multitude of factors have magnified uncertainties, and the memory industry is now facing a market downturn we have never seen before. At the same time, the company today is not what it was in the past, and we will live up to our role as a global semiconductor leader by striving to maintain stability in the ecosystem. Last is the company's ESG management activities.
In July, the company published its sustainability report 2022 and announced the ESG strategy framework, PRISM. PRISM is a new and more developed initiative that includes all social value 2030 goals the company announced last year, while also incorporating new requests by relevant stakeholders. PRISM consists of five pillars: pursue, restore, innovate, synchronize, and motivate. It explains the reason and purpose behind the company's pursuit of ESG management, as well as the detailed goal. The first pillar, pursue, relays the ultimate purpose of ESG management. The company practices ESG management to create economic value as well as social value, and thus create a happier world and a brighter future for all. To that end, the company will try to protect the Earth by restoring the environment, create a better environment by developing innovative and future-oriented technologies, which are captured in the second and third pillars, restore and innovate.
Achieving all of this requires cooperation from not only the company's employees, but all like-minded partners. That is why the 4th and 5th pillars of PRISM, motivate and synchronize, set goals to muster collective efforts of internal members and external stakeholders. Under these five areas, the company disclose around 20 quantifiable goals to be achieved by 2030.
In addition, the company also published a separate TCFD report for the first time this quarter. It contains an analysis of the risks and opportunities related to climate change, as well as an estimation of potential financial impact on the company. The company will provide regular updates to our progress towards these goals and will transparently disclose the process of our unique ESG management efforts based on PRISM. By doing so, the company will try to exert a positive influence and help create a brighter future. Thank you. With that, we're now ready to take your questions.
[Foreign language]
Now Q&A session will begin. Please press star one. That is star and one if you have any questions. Questions will be taken according to the order you have pressed the number star one. For cancellation, please press star two. That is star and two on your phone.
In order to allow as many Q&A chances as possible within the restricted time, we would appreciate only two questions per each participant. [Foreign language]The first question will be provided by Do Hyun-woo from NH Investment & Securities. Please go ahead with your question.
Yeah. [Foreign language]
Thank you for taking my question. I have two. They are about the inventory investment reduction as well as production cut. Now, first about the investment cutback. The presentation did explain a bit about the plan, but can the company elaborate a bit further about the details? For example, what is going to be the extent of the investment cutback? And also, what is going to be the portion between the, let's say, infrastructure versus equipment and also between new investment versus investment into tech migration. Second question is pertaining to production cut. Again, the question is what is going to be the share between DRAM and NAND? And also is the production cut going to be in the form of reduction in the wafer input, or is it going to be suspension of some equipment utilization?
을 낮추는 방향으로 검토하고 있고, 그것이 결과적으로 감산의 효과를 불러올 것으로 판단하고 있습니다. 일부 그간 이, 업턴 마켓에서 생각하기 어려웠던 팹 내의 효율성 증가를 위한 장비 재배치 혹은 팹 간의 제품 재배치, 뭐 이런 것들을 통해서 지금 현재 향후 미래 효율성을 올리지만, 단기적으로는 감산에 준하는 효과를 갖고 오게 되는 다양한 시나리오들을 검토하고 있고, 실제로 일부는 이미 적용되어서 실행이 되고 있습니다.
Please understand first off that it is not going to be easy for me to give you the details about the proportion of production cuts or investment cutback by different categories and for all of our investment. Having said that, in terms of the total CapEx, so compared to this year, including the investment year to date as well as the remaining investment until the end of the year. Compared to the investment for the year, we expect the CapEx to go down by around 50% next year. That is the business plan that we are currently preparing. In other words, a CapEx reduction by slightly over 50%. We are also considering some scenarios where the CapEx reduction could be higher than that. For now, the plan is to be slightly over by 50%.
In terms of the investment reduction between infrastructure versus equipment, I would say that they are going to be roughly proportional. It's not going to be like investment is going to be maintained in infrastructure while it is cut back in equipment. That is not the case. It is going to be proportional. Likewise for DRAM and NAND, although the investment cutback on NAND is going to be slightly higher than on DRAM, again, they are going to be roughly proportional. Then for the production cut, now, given that the demand is slowing down in the industry overall, now what we are going to do is, try to reduce the production on low margin products.
These are the products where the demand to begin with is not very strong, but we would produce them first and then look for demand later. As a result, they tend to be low margin. We would be reducing the production of these products first. We are currently reconsidering the wafer input. We would be reducing the input into the fabs as a way of as the approach toward production cut. Another approach that we are considering and already implementing is something that we were not able to do during the upturn. In other words, in order to enhance the efficiency in a fab, we could reallocate the equipment inside one fab or reallocate the product mix between the fabs.
in the long run, it would serve to enhance the efficiency of our fab operations, while for the short term it would also have the effect of reducing production.
네, 다음 질문해 주십시오.
다음으로 질문해 주실 분은 Citigroup의 이세철 님입니다. The following question will be presented by 이세철 from Citigroup. Please go ahead with your question.
네, 안녕하세요. 씨티그룹 이세철입니다. 저는 미국의 중국 반도체 장비 규제 관련해서 질문 좀 드리겠는데요. 아까 발표 때도 말씀하셨지만, 저희도 중국의 팹을 좀 운영하고 있는데, 관련해서 최근 중국향 반도체 장비 규제로 인해서 향후 당사의 팹 운영 전략하고 그리고 미국 포함해서 다른 지역에 투자를 확대할 가능성에 대해서 답변 부탁드리겠습니다.
Now my question is pertaining to the U.S. restrictions against the semiconductor equipment in China. This was also mentioned briefly during the presentation, but the company also has some fabs in China. I wonder whether the U.S.'s restrictions against the equipment going into China would have an impact on the company's fab operation strategy. Does the company now plan to expand its investment in other geographies, including the U.S.?
네, 질문 감사드립니다. 최근에 다들 보시다시피 여러 가지 geopolitical issue들이 비즈니스의 의사결정에 많은 영향을 주고 있습니다. 특히 현재 중국의 operation을 갖고 있는 당사 입장에서 그런 issue들로부터 오는 여러 가지 제약 조건들은 여러 가지로 고통스러울 수밖에 없는데요. 다만 이미 실적 발표 때 말씀을 드렸던 것처럼 미국 정부가 이 당사의 중국 내 fab 포함 다른 경쟁사의 중국 fab 등등에 1년 간은 향후에 license를 받아야 하는 그 상황을 1년 정도 유예를 시켜준 상황입니다. 그리고 그 license 유예된 상황에서 현재 그 fab 내에서 개발되거나 투자되고 있는 것들 수준까지는 일정 정도 투자에 대한 유
Now, as we all know, there are geopolitical issues that are also affecting businesses' decision making process these days. Yes, there is also the restrictions against the equipment being brought into China. For a company that has operations in China, I would say that these restrictions can also be painful. As also was explained earlier, the U.S. government granted a one-year waiver on licensing requirements for the company's China fab operations. Not just for the company, but also for other players too that have operations in China. This one-year waiver, we believe that it also applies not only to. It also applies to the development that is currently underway in the fab.
We believe that this would also be applicable to at least to some of the investment for the development that is currently underway at the fab. Also for the one year waiver, we are expecting and hoping that there would be a one year extension thereafter. Again, this is not certain given the high level of geopolitical uncertainties these days. This means that what we used to take as common sense in business, for example, to produce out of the most efficient and cheapest region and then to supply globally. This used to be the business common sense. Now on top of this, we have other layers of factors that we have to consider, which is magnifying the uncertainties from the business perspective as well.
Now, having said that, I do believe that it would be essential for us to diversify our production bases down the road for the mid to long term, but for the short term, for us to change production bases at this time will not be easy as you can imagine.
The following question will be presented by SK Kim from Daiwa Capital Markets. Please go ahead with your question.
[Foreign language]
I have questions about the demand. Currently, there are ongoing concerns about the demand deceleration and in terms of the server demand, for example, the demand for DDR5 and HBM3. What is the company's outlook for the demand on the server side? Also for next year, new applications are expected to be launched, for example, the XR device. How does the company foresee these new applications to affect the memory demand? Does the company expect these new applications to be the driver of new demand growth? Well, in response to your questions, I would like to provide my answer in three ways. One is with regards to DDR5 demand, so server demand plus the overall demand, and then the HBM3 demand, followed by the new applications which could potentially be growth drivers.
내년도에 연간 전체적으로 한 20% 이상의 비중을 차지하면서 내년 말로 가면 30% 이상까지 비중이 확대될 걸로 이렇게 보고 있고요. PC 같은 경우에는 연간 전체적으로 약 한 30% 비중, 근데 연말로 가면 그 이상으로 PC도 늘어나는 그림으로 보고 있습니다. 관련해서 저희는 DDR5의 그 성능 그리고 완제품 포트폴리오, 그다음에 고객과의 인증 프로그레스 이러한 그 강점들을 바탕으로 해서 시장 성장세보다는 좀 더 적극적으로 DDR5의 비중을 늘려가려고 하고 있습니다.
First about the DDR5. Out of the server demand, we expect this to take up over 20% share by next year. By the end of next year it is expected to reach over 30%. Likewise for PC, its share would be around 30% and it is bound to go higher as the year gets closer to the end. In addition to that, for the DDR5, based on our strengths in performance as well as our finished product portfolio and customer qualification process, the company plans to increase the sales of DDR5 that would outperform the market trend.
HBM의 경우에는 올해 전 세그먼트 내에서 HBM 마켓의 성장률이 가장 컸습니다. 올해가 전년 대비해서 HBM 세그먼트가 약 50% 이상 성장한 걸로 보고 있고요. 내년도에는 그것보다 더 성장률이 올라갈 걸로 이렇게 전망하고 있습니다. 다만 HBM 세그먼트도 전체적으로 글로벌 이코노미 영향을 지금 받고 있기 때문에 당초 예상했던 규모보다는 절대 규모 자체가 조금은 줄어든 모습입니다. 하지만 자사의 경우에는 저희가 갖고 있는 제품의 아주 우수한 성능, 그다음에 검증된 품질, 그다음에 고객 및 SoC 파트너와의 협업 만족도 이런 것들을 바탕으로 해서 압도적이고 선도적인 지위를 계속 해 나갈 것으로 믿고 있습니다.
For the HBM, out of all segments, HBM market grew the fastest this year at a pace of 50% YOY, and we expect the pace of growth next year to be even faster. The HBM segment was also affected by the macro condition in the global economy. The sheer size is not as big as we had expected, but we believe that the HBM segment will continue to grow and especially for the company based on our performance, quality and the collaboration satisfaction from the customers and the SoC partners. The company will maintain overwhelming leadership in this segment.
DDR5와 HBM 외에도 자사 같은 경우에는 LPDDR5 제품의 차별적 성능 우위, 그리고 주요 하이엔드 세그먼트, 폰 메이커들과 그다음에 SoC 파트너들과의 협업이 지금 굉장히 잘 확대되고 있어서 내년도에 저희가 DDR5, HBM과 더불어서 저희 사업의 안정성 및 확대를 도모하고 수요 안정화에 기여하는 데 크게 활용을 할 수 있을 걸로 보고 있습니다.
In addition to DDR5 and HBM, the company also has advantage in LPDDR5. We have performance advantage and we also have our collaboration advantage with the high-end segment phone makers as well as SoC partners. Based on these advantages that we enjoy, we will continue to expand the HBM segment for the company, and we believe that we will be able to help the product, will be able to help stabilize demand.
마지막으로 신규 성장 분야에 대한 의견을 드리겠습니다. 지적하신 것처럼 XR, VR 디바이스나 한편으로 또 Automotive 관련된 영역이 향후에 디바이스 관점에서 기존의 클래식한 PC나 스마트폰에 이어서 미래 디바이스 shipment 성장을 충분히 이끌 수 있는 그러한 영역이라고 보고 있습니다. 현재 XR, VR 디바이스 같은 경우에는 set shipment 자체는 그렇게 크지 않지만, 당장 내년도에도 올해 대비해서 30% 이상 set shipment 성장이 있을 걸로 보고, 중장기적으로도 비슷한 수준의 성장률을 가지고 갈 걸로 보고 있습니다. 자동차 산업 같은 경우에는 저희가 ADAS 탑재 기준으로 봤을 때 올해 대비해서 약 5년 후 정도에 올해 대비 약 두 배 정도의 탑재율 증가를 전망하고 있고요. 그다음에 향후 한 10년까지 좀 넓혀서 보면 관련해서 들어가는 메모리 수요량이 현 시점 대비해서 멀티플, 꽤 높은 멀티플로 성장할 걸로 보고 있습니다.
현재로서는 약 한 다섯 배 이상, 그다음에 그 주변 그 유관 산업들, 예를 들면 In-car Memory 외에 Network이나 Edge, 그다음에 그
For new applications like XR and VR devices is, and also in the automotive segment, we believe that the devices are going to have the potential to increase a bit shipment in addition to the PC and the smartphones. For the XR and VR devices, although the set shipment currently is not too big, they will be increasing by about 30% next year. The set shipment of these devices is expected to grow by 30% next year. For the medium to longer term, this trend is expected to continue. For the automotive segment, for example, the attachment on ADAS. In this case, we believe that the attachment rate is going to double in five years time compared to what it is today.
In terms of the memory content, so when we look at the 10-year time horizon, then we believe that there is going to be an increase in memory content by high multiples compared to what it is this year, so by around 5x multiple. Also when we look at the other related industries, in addition to the in-car memory, there's also the network and edge computing and also hosting. When we look at the other sectors as well, it is highly likely that the growth is going to continue.
The following question will be presented by Nicolas Gaudois from UBS. Please go ahead with your question.
Yes, good morning, and thanks for taking my questions. Just coming back to how you would manage your operations in China going forward. Beyond the U.S. restrictions, there is also currently the lack of availability of EUV lithography tools into China due to how the Dutch government is applying the Wassenaar Arrangement. You know, it's arguably unlikely to change. That places a timeline in terms of what you'll be able to do in terms of a tech migration in the Wuxi fab as you eventually would move otherwise to 100 nanometer. With that in mind, can we be confident that you have the flexibility and strategic options, including the possibility of moving back tools to Korea if needed over time?
Secondly, you have seen obviously downside in DRAM and NAND flash shipments to customers, as did your peers in the course of Q3. Do you believe that we're now at a stage where destocking at the customers level is sufficient for bit growth to stabilize in the near term? You just guided for bit growth to be flat quarter-over-quarter in Q4. If so, would you assess when looking at combined upstream and downstream inventories that we may have passed a peak level already of combined inventories? Thank you.
Yes. [Foreign language]
Tech을 가장 extension 한다고 보면 일부 EUV layer를 한국에서 백업하는 것까지 고려해서요. 그렇게 되면 timeframe상 2020년대 후반까지도 조금 어려움은 있겠습니다만, 우시에 있는 DRAM 팹을 운영하는데 아주 critical한 문제는 없을 것으로 보입니다. 물론 일부 cost의 상승과 어려움이 있을 것으로 예상됩니다. 다만 현재 1년씩 이렇게 license를 받는 것을 유예받아 가면서 이제 새로운 규제가 들어왔고, 이것을 고려하면 1년 이후에 이런 license를 받는 것이 유예되지 않는 상황이라면, 이게 tool by tool license를 받아야 하는 상황이라면 메모리 산업의 특성상 장비를 도입하는데 어려움이 있을 수 있고, 그렇게 되면 EUV issue가 2020년대 후반에 오는 것보다 훨씬 더 빠른 시점에 팹을 운영하는데 어려움이 있을 수 있을 것 같습니다.
Now regarding the first question. The Wassenaar Arrangement, as far as I understand now, this is based on unanimous decisions or consensus, meaning that it has the kind of structure where imposing regulations would not be easy. I do believe that the current situation comes down to issues between certain countries. Now for the EUV and in the Wuxi fab. It is a very cautious forecast, but it appears at this time that it is unlikely that the EUV will be able to be entered into our DRAM fab in Wuxi. Again, I'm being very cautious here.
Given that situation, lacking the EUV equipment, in terms of the DRAM tech expansion now we could go ahead with it with some of the EUV layers being backed up in Korea. If this is a scenario, I do believe that at least until into the late 2020s, there will not be any critical issues in operating our DRAM fab in Wuxi, although there would be some increase in cost and perhaps a bit of complexity. Now, let's say, currently there are new restrictions in place with one-year waivers. Right now we are hoping for yearly extension.
Assuming that the waiver is not granted anymore, meaning that if it comes to a situation where we have to receive tool by tool licenses, then as you would know by the nature of the semiconductor equipment, this would create a very difficult situation for us to bring in the equipment. If that does transpire, then, the difficulty in not being able to bring in EUV into the Wuxi DRAM fab would probably occur earlier than the late 2020s.
답변을 마무리 드리면, 지금 Wuxi 포함 중국에 있는 팹에서 문제가 생겼을 경우는 사실 그것은 전략적 flexibility라기보다는 일종의 contingency plan에 해당되는 것일 텐데요. 사실 팹을 운영하기 어려운 상황이 온다고 가정할 경우, 팹을 매각하거나 장비를 매각하거나 혹은 장비를 한국으로 가지고 오는 등 다양한 시나리오들을 당연히 검토를 하고 있습니다. 다만 아시다시피 그러한 것들은 매우 extreme한 상황에서 벌어져야 하는 일종의 contingency 상황이라고 보는 것이 맞을 것 같고요. 그러한 상황이 오지 않고 팹을 운영을 할 수 있기를 바라고 있습니다.
To wrap up my response to this question. If the issue is that we are no longer able to operate the fab in Wuxi, then this actually is not in the realm of strategic flexibility. I believe that would fall under contingency plan. If the time comes when it appears it is difficult to maintain operations of the fab in Wuxi, then we might have to sell off the fab or equipment or bring in the equipment to Korea. We are looking into various scenarios, but again, this would amount to contingency. This would be, you know, tantamount to an extreme situation. We are currently hoping that we will be able to maintain our fab operations without such contingency.
두 번째 질문 주신 업계 재고 수준 전망에 대해서 간단히 말씀드리겠습니다. 현재 공급단과 고객을 합친 업계의 재고 수준은 평균 대비해서도 높고, 아마 이 상황이 내년 1분기에 seasonality까지 감안을 하게 되면 당분간 내년 1분기 정도까지는 peak 상황이 될 수도 있겠다 이렇게 보고 있습니다. 다만 고객들도 현재 재고 소진 우선 정책을 펴고 있고, 그다음에 공급단에서 생산 증가 여력도 좀 줄어들게 된다면 결국은 peak 이후에 업계 재고 수위는 점차 내려갈 걸로 그렇게 지금 기대를 하고 있습니다.
To your second question about the inventory. For the inventory in the industry, both on the supplier side and the customer side, the inventory level is higher than average, and we expect the high inventory to remain into the first quarter of next year, especially given the seasonality in the first quarter. We believe that it will peak around the first quarter of next year. The customers are currently trying to work down their inventory first. Also from the supply side, the suppliers have also reduced their production ramp-up capacity. We believe that after reaching the peak, the inventory level across the industry will start to come down.
The following question will be presented by JJ Park from J.P. Morgan. Please go ahead with your question.
[Foreign language]
Now, the prices for both DRAM and NAND are expected to keep declining in the second half of the year. Now when we look at the past trend, then along with the price decline, there is also price elasticity growth in content. What does the company foresee in terms of the content growth per application next year? Also looking at the bit shipment in the third quarter and the guidance for the fourth quarter, it appears as if there is going to be DRAM and NAND inventory carry forward next year. The company's guidance for next year's bit shipment is much higher. What would be the guidance for production growth next year? I would expect that the production growth will fall below the level of the bit shipment growth, but what is the company's expectation?
Now, first about the content growth projection. It was also mentioned in the presentation, that yes, with declining prices there is some expectation of content growth down the road. When we look at the past 10-year history of memory prices, then the memory price decline has also slowed down. To give you more details, compared to DRAM, the price elasticity is higher for NAND, which has stronger consumer orientedness. Having said that, these days consumers' purchasing power is also down, so it is unlikely to expect any short-term sharp rebound in content growth. By application for PC and smartphones, we expect the DRAM demand to grow by about low 10% level, and for NAND, by mid- to high-teens %.
For server, where demand remains relatively solid, DRAM will be about a 10% level, so mid at 10%. For NAND it will be over 30%. Again, looking at history, looking at the past memory downturns, we do see that the content growth has triggered a transition to high density memory. For 2023, there is going to be an increase in PCs or enterprise PCs. There is going to be increase in the sales of enterprise PCs which have higher memory content. Also for the mobile devices, now there is growing demand for larger screen and also, transition to next generation form factor, which are also increasing demand for higher specifications and higher density.
Yes, the company is trying to create demand from the customers by offering higher content products, and this will also serve as a promotion point for the customer's products, and this will also be an element of stronger competitiveness for the customer's products. Now, as someone who is in the memory industry, for me to be negative about production bit growth is quite awkward and even painful because the business model in the memory industry is that we would try to cut costs by increasing bit growth in both production and sales, and then the saved costs would then be transferred to the customers in the form of lower ASP. Now, I mentioned already that the capacity investment in wafer will be minimized, so the production bit growth is going to be minimal.
There is the tech migration bit growth. Usually where we gain the bit growth is from the wafer investment and investment into tech migration. I mentioned already that there is not going to be increase in production, so the wafer capacity will be reduced, meaning that the production would also be reduced. The remaining source of bit growth would be investment into tech migration. As you would know, the bit growth coming from tech migration investment in DRAM is smaller than NAND. For the DRAM next year, we are looking into various investment scenarios, but we are also reviewing the possibility of perhaps reducing bit growth for next year's production.
The last question will be presented by Sung Eun Kim from Meritz Securities. Please go ahead with your question.
[Foreign langauge]
I have some questions about NAND. Now we see that the demand deceleration is much faster in NAND compared to DRAM. When we look at the profitability in the industry, then NAND's profitability is lower than DRAMs. The company recently increased its NAND exposure through Solidigm, and it appears as if it has struggled quite a bit in the third quarter. When we look at the numbers, the bit growth excluding Solidigm, then it is higher. What was the profitability in the third quarter? In the fourth quarter, it is expected that there will be inventory valuation loss and also some provisioning. What is the company's expectation or guidance for the future profitability of NAND? Another question about NAND is about the development.
In the presentation, it was mentioned that the development for NAND is on schedule and that the company has completed the development of 238-layer. As we all know, development has various meanings, and also completion could mean different things. Could you be more specific about your development of 238-layer being complete? And when does the company believe that there you can go through the customer qualification and start selling to customers? And what is the expected cost reduction to come from this 238-layer?
Now the second question will be answered by Mr. Park Chan-dong, who is the Head of NAND Marketing, and this is Kevin Noh responding to the first question regarding Solidigm.
Now, please understand that, given that Solidigm is a non-listed company, it is difficult for me to give you specific numbers about their performance in the third quarter. Excluding the one-off costs, we were expecting Solidigm to record profit in non-GAAP terms and about KRW 200 million deficit or KRW 200 million loss in GAAP terms. That was the expectation in the early part of the year. As you know, the market situation has considerably worsened than what was expected in the early part of the year. As a result, Solidigm's performance was also lesser than what we had expected.
Also at the same time, Solidigm used to be a one division in Intel, but now it has become a standalone company, so it is still undergoing the process of standing up on its own. As it was going through this transition, it was, I would say, hard pressed to deal with many of the challenges that were arising rapidly from the market. When we look at the overall competition landscape in the NAND market, then it is also true that we do need some adjustment or corrections. For SK hynix, we have already gone over the critical mass in terms of the market share.
Yes, although it is struggling at this time, I do believe that once we complete the consolidation in the next 1-2 years, then the strategic benefits out of this will far outweigh the challenges or the issues. Now next is about the NAND development. Now when we were first introducing the 3D NAND and at the time, because of our, let's say lesser competitiveness at the time, we also had some gaps in terms of the cost and the technology. Starting with the 128-layer, we were able to catch up very quickly in terms of the mass production as well as the cost. Coming into the 176-layer, we were able to stabilize the cost as well as the mass production.
[Foreign language]
We smoothly went on to mass production for 176-layer. By the end of the third quarter, it reached 60% of all NAND production. We completed the development as we have notified through the SMS two months ago. The customer sample will start to be provided in the early part of 2023. Given the company's know-how in being the first player to apply the double stack and PUC in the industry, we foresee no complications in the development and mass production of the 238-layer. Our plan is to start mass producing 238-layer by the middle of 2023.
The last question will be presented by Ricky Seo from HSBC. Please go ahead with your question.
[Foreign language]
Thank you for taking my questions, without ending the conference call. Now I have two questions. First is that now we see that the demand deceleration is much more than expected, and likewise the reaction by the memory companies is getting more and more aggressive. Now, looking at the current picture then, of course, we did talk about inventory earlier. Looking at the overall picture and especially looking at the cycle then, in what part of the cycle does the company believe that we are in now? And in other words, when does the company believe that the market situation will begin to recover? And the second question is about the valuation, inventory valuation loss.
Now given the sharp decline in memory prices, I presume that there would have been recognition of inventory valuation loss. Can the company share the size of the valuation loss if possible?
사실 그 macro economy의 불확실성, 거기에 더해 geopolitical issue 등으로 인해서 사실 유래가 찾아보기 힘들 정도로 serious한 형태로 지금 저희 모두에게 다가오고 있습니다. 이 상황에서 DRAM과 NAND를 나누어서 생각을 해볼 필요가 있는데요. 고객 입장에서도 어느 정도의 다운턴, 가격이 내려가고 뭐 그런 측면에서 그 상황을 즐길 수 있는 상황이 있을 텐데 지금은 supplier인 메모리 업체들도 고통스럽지만 사실은 많은 재고를 갖고 있는 고객들 입장에서 그 고객들이 갖고 있는 재고로부터 두 번째 질문과 연결되어 있는데 재평가손이라든가 뭐 여러 가지 측면의 고통을 생각해 보면 사실 이제 더 이상 고객도 현재의 상황을 즐기기에는 쉽지 않은 상황이라고 보여집니다. DRAM 같은 경우에는 뭐 사실 이 supplier 수가 제한되어 있고, 그런 측면에서 보면 이번 고객도 힘든 이러한 상황을 조금 더 빨리 복원할 수 있는 힘이 조금 더 강하다고 볼 수 있겠고요.
NAND 같은 경우는 사실 최근에 geopolitical issue도 더 들어와 있기 때문에 더 예측하기는 힘듭니다만, supplier 수나 또 가격 탄력성에서 오는 여러 가지 변화 요인들 등으로 인해 DRAM보다는 시장의 복원, 시장의 예측 가능성을 이야기하기가 조금 더 어려운 상황이긴 합니다.
I take it as a larger question about where we are as an industry and where is the memory industry headed in the future. Looking at the memory cycle now, then obviously, yes, it is highly cyclical. Of course, the memory industry itself has gone through many upturns and downturns. Looking at where we are standing today at this time in 2022, I would say that the current downturn is quite serious for everyone involved in an unprecedented manner because it is mostly caused by the macroeconomic uncertainties and geopolitical issues. We need to take a look between DRAM and NAND. For the customers, when there is a downturn, then there is a certain benefit to be gained for them in the form of lower prices.
Looking at the downturn today, then of course, obviously it is painful for the suppliers, but also for the customers, they are no longer in the position to just enjoy this downturn and lower prices because they are saddled with a high level of inventory. Obviously there would also be some inventory valuation loss, which is also related to the second question. Now looking at DRAM, the number of suppliers is limited and also the customers are going through a hard time as well. Given the market dynamics here, I do believe that there would be perhaps a stronger force or stronger power to try to restore the market. Looking at NAND, it is a bit different. I mean, on one hand, the geopolitical issues are more complicated and there are many more suppliers, and there's also the price elasticity to consider.
It is much more difficult to predict when the NAND market is going to recover.
결국 이 macro economy의 불확실성이 어느 정도 해소되는 순간이 그 메모리 그 업에 있어서도 새로운 upturn을 바라볼 수 있는 시점이 될 텐데요. 물론 조금 그 시차는 있겠습니다만. 그런 관점에서 보면, hopefully 내년 하반기 정도에는 어느 정도 시장이 안정화될 것으로 기대하고 있고, 또 그리 되기를 바라고 있습니다만, macro economy의 상황이나 여러 geopolitical issue 등이 해결되지 않는 상황에서 그보다 더 다운턴이 길어질 가능성도 배제하고 있지 않습니다. 다시 말씀드리지만, 생산을 축소하거나 Capa를 축소하는 것은 메모리 사업자 입장에서 굉장히 painful한 일입니다. 적절한 생산 bit growth, 판매 bit growth를 통해서 원가를 낮추고, 이러한 것들이 고객이나 전체적인 IT industry의 성장에 기여하는 것이 가장 건전하고 건강한 것이라고 생각을 하고 있습니다. 다만 내년 같은 경우에는 그러한 생산 bit growth가 심지어 DRAM 같은 경우에는 없을 수도 있는 상황이어서요.
결코 건전한 상황이라고 볼 수 없겠고, 다시 말씀드리지만 DRAM은 그러한 복원력이 조만간 작동해서 새로운 new norm으로 건전한 상태로 돌아갈 것으로 보입니다. NAND는 조금 더 시간이 걸리겠습니다만, DRAM과 같은 track으로 조금의 시차를 두겠지만 그런 복원력을 가지고 있을 것으로 믿습니다.
That means the timing of when the macroeconomic uncertainties and geopolitical issues ease would be when the memory market or the memory industry can also look forward to an upturn. Of course, there is going to be some time lag between the two events, but we would have to wait and see whether the macroeconomic conditions and geopolitical issues ease in any way. Given this picture then hopefully the market would start to stabilize in the second half of next year. That is our expectation and hope. Again, due to the macroeconomic and geopolitical uncertainties, we cannot rule out the possibility that the current downturn may last a bit longer than that.
For the memory supplier, to cut back on production and capacity is indeed quite painful because like I said, what we do is try to gain production and sales bit growth and then achieve cost reduction, then return this to the industry in the form of a lower price. That I believe would ensure a healthy and wholesome development of the industry. Bit growth is going to be limited next year and even perhaps none, from the DRAM side. This is by no means a healthy picture that we want. Again, for DRAM, there will be a strong enough force for restoration, sometime next year, so we hope that we will be able to go back to normal, healthy industry in the near future.
Likewise for NAND, it would move, I believe, in the same track as DRAM, although it might take longer.
Yeah. This is the CFO speaking. With regards to your question about inventory valuation loss, in the past quarter, it was around KRW 200 billion. Thank you very much.
[Foreign language]
That concludes the SK hynix earnings release conference call for the third quarter, 2022. Thank you very much for your participation.