Good morning and good evening. First of all, thank you all for joining this conference call. Now we will begin the conference of the Fiscal Year 2019 Second Quarter Earnings Results by SK hynix. This conference will start with a presentation followed by a divisional Q&A session. If you have a question, please press star one, that is star and number one on your phone, during the Q&A session. Now, we shall commence the presentation on the fiscal year 2019 second quarter earnings results by SK hynix.
Good morning and good afternoon to those calling in from abroad. This is Park Sanghoon, the head of IR at SK hynix. Welcome to the SK hynix 2019 second quarter earnings release conference call. Before starting the conference call, allow me to introduce the executives present here with me today.
First, Executive Vice President, CFO, Cha Jin-seok, Vice President Kim Seok, in charge of the DRAM Marketing Group, and Vice President Kim Jung-tae, in charge of the NAND Marketing Group. 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 2019 second quarter earnings release conference call. Let me now turn over to Executive Vice President Cha Jin-seok to present the earnings for the second quarter and the company's market outlook and plan.
Good morning. This is CFO Cha Jin-seok. I will first report on the company's financial performance in second quarter 2019. Amidst rising external uncertainties, consolidated sales in the second quarter was KRW 6.452 trillion, d own by 5% from the previous quarter.
This is despite the higher unit sales in the second quarter, owed to the continuous faster than expected ASP decline. For DRAM, rising concerns over demand amidst intensifying global trade conflicts resulted in conservative purchasing behavior by some customers. The company responded with mobile and PC DRAM that have relatively higher demand growth. DRAM bit shipment increased 13% quarter-on-quarter, close to the original expectation. ASP fell 24% from the previous quarter, with computing applications showing continued price weakness. Demand for NAND Flash continued elastic recovery as price went down. The company actively responded to high density mobile solution demand and temporarily increased sales of discrete products as a way to relieve inventory burden. NAND flash bit shipment grew 40% quarter-on-quarter, performing significantly better than expected.
ASP was down 25% quarter-over-quarter, as price competition continued among suppliers and sales of the relatively low priced discrete applications increased. In MCP, increase in demand volatility of Chinese smartphone customers led to sluggish shipment growth and steeper ASP erosion. Revenue decreased 11% quarter-over-quarter, while revenue mix was maintained at 23%. Although cost per unit was reduced across all applications, it was not enough to completely offset sharp price decline. Incremental costs recognized from the low efficiency in initial operation of the new fabs and higher unit sales led to rise in COGS. Operating profit in the second quarter was KRW 638 billion, down 53% from the previous quarter, and operating profit margin was 10%. Depreciation and amortization in the second quarter was KRW 2.108 trillion, slightly up from the previous quarter. EBITDA was KRW 2.745 trillion with EBITDA margin of 43%.
There was a net non-operating profit of KRW 38 billion in recognition from foreign currency related gains as the Korean won depreciated at the quarter end as well as others. Corporate tax in the second quarter was KRW 139 billion, lower than in the previous quarter. Net profit after tax was KRW 537 billion with a net profit margin of 8%. Consolidated cash balance at the end of the second quarter was KRW 3.11 trillion, down by KRW 4.082 trillion from the previous quarter. There was a large-scale cash expense occurring in the second quarter for the payment of corporate tax and dividends for 2018. Interest-bearing debt was KRW 8.747 trillion, up by KRW 2.6 trillion. Debt-to-equity ratio was 18% and net debt-to-equity ratio was 12%. Let me now turn to the company's market outlook and plan.
DRAM market in the second quarter showed signs of recovering from the demand slowdown in the previous quarter, the pace was not up to expectation due to unanticipated external uncertainties. Unlike the projection for recovery from late second quarter, demand from data center customers remains sluggish. Despite the decline in their DRAM inventory level, data center customers are expected to maintain conservative purchasing policies for now. Purchasing by enterprise server makers is expected to partially normalize in the second half to maintain stable inventory level as their business continuity is important. Overall, the growth rate of server DRAM demand will slow down considerably from the previous year.
[Non-English content].
In the global smartphone market, high-end smartphone sales were below expectations in the first half, while demand volatility grew due to escalating trade tension. As some also forecast robust demand growth for low to mid-range models and easing of demand volatility, we are keeping a close eye on market trends.
[Non-English content].
Aside from the uncertainties from external conditions, supply of smartphones supporting 5G service will begin in earnest next year, driving up the adoption rate of high-density products from 8 GB up to 12 GB . It also means that demand for mobile DRAM will keep growing.
[Non-English content].
In addition, the CPU undersupply that affected the PC and graphic DRAM market has begun to be eased, and demand started to pick up from the late second quarter. As the trend continues into the second half, it is expected to offset part of the decline in server DRAM demand.
[Non-English content].
In the NAND market, price decline is leading to elastic recovery of demand across most applications. As PC CPU undersupply situation starts to be eased, adoption of high-density SSD of over 512 GB is rapidly increasing, and the mix for the MCP products that adopt 64 GB or over is approaching nearly 50%, continuing the content increasing trend. Internet data center clients who have been passive in NAND as well as in DRAM, are expected to begin gradually increasing their purchase of server SSD in the second half.
[Non-English content].
From the supply side, NAND suppliers are reducing wafer capacity and wafers input. Due to the production disruptions of some NAND suppliers, the pace of inventory reduction across the industry is likely to accelerate. This is heightening the possibility of easing demand, supply, demand imbalance in the second half of this year, and thus price decline is expected to slow down considerably.
[Non-English content].
To effectively respond to such changes in the market environment as well as other externalities, the company plans to adjust its production and investment.
[Non-English content].
First, to reflect the changes in DRAM demand environment and to strengthen competitiveness of the CMOS Image Sensor business, from the second half, we plan to start converting the DRAM capacity in M10 to mass produce CMOS image sensors. Consequently, DRAM wafer capacity that stayed relatively flat compared to last year will begin to decrease from the fourth quarter of this year. Adding the wafer capacity loss resulting from tech migration, next year's DRAM capacity will be reduced compared to this year's.
[Non-English content].
Likewise, for NAND, we will accelerate the reduction of 2D wafer capacity, for which demand is quickly declining, bringing down NAND wafer input by over 15% this year Year-over-Year, compared to the originally planned reduction of 10%.
Going further, with less visibility in demand growth down the road, the company intends to revisit the schedule for opening an additional clean room in M15 fab and equipment installation in M16 fab for efficient and stable management of the company's capacity and cash flow. Overall, next year's CapEx will be reduced considerably compared to this year's. Not only such flexible CapEx execution and reduction in capacity, the company will also focus on developing next-generation process technology and deploying stable volume production. We will also run a product mix heavier on high-density, high-value-add products to help ensure profitability.
For DRAM, we will reduce the portion of 20 nm class products while increasing the 10 nm class products in total from the mid 40% at the end of second quarter to 80% by the end of the year. We also plan to market computing products with 1Y nanometer technology from the second half of this year. Product mix will focus on high density mobile products that offer relatively strong demand and profitability. We will prepare for NAND, the 72-layer 3D NAND will be the mainstay for now. We plan to gradually increase the portion of 96-layer starting in the second half. We will focus on sales for high-end smartphones and PCIe SSD market. We will prepare for volume production and sales of the 128-layer 1Tb TLC product, which the company became the first in the world to successfully develop in June.
Meanwhile, DRAM bit shipment growth in the third quarter is planned at mid to high single-digit percent in light of the base effect in the previous quarter and the heightening demand uncertainty. NAND bit shipment growth is likely to remain flat quarter-on-quarter, at most varying by low single-digit percent. For the year, in light of the slower than expected server DRAM demand, DRAM bit shipment growth target has been adjusted downward to low to mid-10%. For NAND, given the higher than planned shipment in the first half, bit shipment growth target has been adjusted upwards to high-40%. The emergence of a number of external factors in the last quarter has dramatically lowered visibility into both supply and demand in the memory market. The weaker than expected demand recovery has kept suppliers' inventory levels still high, while further driving down the price of DRAM and NAND.
As has been explained so far, the company will effectively respond to the current downturn with the strategy of flexibly adjusting capacity and CapEx in both DRAM and NAND according to the changing demand environment. At the same time, we remain convinced of the mid to long-term growth potential of the memory market. The company will maintain a long-term viewpoint as it tries to turn this downturn into an opportunity for strengthening our core competitiveness in our products and technology. We ask for your continued encouragement and support. Thank you.
Now Q&A session will begin. Please press star one, that is star number 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 number 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. The first question will be provided by Sei Cheol Lee from Citigroup . Please go ahead with your question.
First of all, I would like to thank you for the explanation on the very good strategic direction of the company. I have two questions. About the export control by Japan on the semiconductor materials, I'm sure that the Japan's move has also affected SK hynix. What is the current status of supply and demand, what is your outlook down the road? The second question is about the NAND price, which appears to be picking up recently, as you have also mentioned in your presentation. Can you also share with us your outlook on the NAND supply demand in the future?
Let me first respond to your question about the export control from Japan. Regarding the export control, the company is trying to make sure that there would be no disruption to the production.
On one hand, we are trying to secure as much inventory as possible, and we are also trying to diversify the vendors while trying to minimize the input that is required for production. If the export regulation is prolonged, then we cannot rule out the possibility of the move causing some product production disruption. Given this fact, we are going to keep a very close eye on the situation. Please understand that, for us at this point, it is difficult to share more details because of the current situation and because of the fact that it is highly uncertain and unpredictable.
[Non-English content].
Now, regarding your second question about the outlook on NAND supply and demand. Now, on one hand, it is true that the volatility in demand and supply continues because of the external uncertainties.
[Non-English content].
Looking at the NAND market itself, the inventory level of the suppliers, which had been the main cause of the supply-demand imbalance so far, is going to fast decline starting in the third quarter, and we believe that that is going to help stabilize the supply-demand dynamics as well as the price.
[Non-English content].
Such stabilization in the supply-demand dynamics and the price are expected to continue into the fourth quarter.
[Non-English content].
[Non-English content].
The next question will be provided by Hyun-woo Doh from NH Investment & Securities. Please go ahead with your question.
[Non-English content].
First is about the production disruption by Toshiba and the Western Digital fab. What is the current situation as understood by the company? The second question is about the outlook for demand for smartphones, especially the Chinese smartphones, because the U.S. appears to be easing some of the regulations against Huawei recently.
[Non-English content].
Regarding your first question about the disruption to the production, our understanding is what we have read up on the media reports.
[Non-English content].
One more thing is that because of this change, what I have mentioned earlier, the inventory level of the suppliers, which had been the major cause of the supply-demand imbalance so far, is now going to decline faster than anticipated.
[Non-English content].
Now to your second question, rather than specifically mentioning a particular customer, I would like to just give you a general description about the mobile market in general.
[Non-English content].
Demand from the mobile market has continued from the first quarter up until the second and the third quarter. As you have understood, the demand coming from the Chinese smartphone makers is also on the rise.
[Non-English content].
The situation is different by customers. For some customers, they are increasing demand in order to prepare for a possible additional upside. Whereas for some others, they have turned around from their conservative purchasing to a more normal purchasing behavior.
[Non-English content].
The next question will be provided by Ricky Seo from HSBC. Please go ahead with your question.
[Non-English content].
My first question is about the inventory levels, which you did mention earlier, but then could you be more specific about the current inventory level? What is the current inventory level for both DRAM and NAND at SK hynix, and what is your outlook for the inventory level? Also in relation to this, can you also share with us the inventory level of the customers or by application as understood by SK hynix? The second question is about the pricing of DRAM. We see that there has been some segregation between the spot price and the fixed price in the DRAM. On the other hand, we have been observing a sharp rise in the spot price of DRAM recently. Do you believe that this rise in the spot price is also going to affect the fixed price down the road?
[Non-English content].
To separate between the DRAM and NAND, I would like to first answer the question regarding the spot price.
[Non-English content].
Now to your question about whether the recent spike in the DRAM spot price is going to have an impact on the contract price with the OEMs, is too early to tell at this point.
[Non-English content].
The OEM customers currently are maintaining a wait and see approach. Yes, some customers do seem to have some upside demand, but then we see that currently the signals are still mixed.
[Non-English content].
Going back to the first question about inventory level, first about the DRAM inventory.
[Non-English content].
Given the demand situation of late, by the end of the second quarter, the company's inventory level in DRAM has risen.
[Non-English content].
As a result of this, the inventory reduction to occur in the second half is likely to be delayed than expected.
[Non-English content].
On the other hand, from the customers, especially from the server customers, their inventory level, as we understand it, has gone down from the average eight weeks to nine weeks by the end of last year to around six weeks by the end of the second quarter.
[Non-English content].
We believe that the reduction in the inventory level will continue into the end of this year, although it is going to vary by customer.
[Non-English content].
Meanwhile, for the mobile customers, we believe that they have maintained normal inventory level, although in this case as well, it also varies by customer.
[Non-English content].
Let me respond to the question about NAND.
[Non-English content].
For NAND in the second quarter, because of the continuing mobile demand, and also thanks to our efforts to diversify the channels, we were able to increase sales more so than planned. As a result of this, the inventory level has decreased quarter-on-quarter.
[Non-English content].
For the products and applications for which demand is falling or which competitiveness is falling, we have also made some adjustments to the production and the sales plan in order to improve the inventory status.
[Non-English content].
The inventory level of the customers by application, we believe that, of course, that there would also be some variances by customer, but still we believe that they are mostly maintaining a normal level of inventory, which means an average of four weeks to six weeks.
[Non-English content].
Now the outlook for the second half. Now, given the current demand and the sales outlook for the company, we believe that the company's inventory level will continue to reduce to a normal level by the end of this year.
[Non-English content].
[Non-English content].
The next question will be provided by Jung Hoon Yoo from Korea Investment & Securities. Please go ahead with your question.
[Non-English content].
I have two questions. First is about the PC DRAM demand. By the end of the second quarter, it was explained that the demand for the PC DRAM has been recovering. Coupled with the demand for the PC with the Windows 10 cycle, then perhaps the DRAM demand for PC is going to intensify in the second half. Can you give us an outlook for the PC DRAM demand in the second half. The second question is: in the first half, because of the hyperscalers CapEx reduction, there has been some impact as well. Now the hyperscalers data center, the utilization rate is going to recover soon. What do you see as the current utilization rate of the hyperscalers data center?
Once the hyperscale data center goes back up to the normal utilization rate, then this is also going to increase the CapEx. When do you believe the timing of this?
[Non-English content].
First, about the PC demand. Looking at the company's performance as well, we see that, yes, there has been a sharp increase quarter-on-quarter between the first quarter and the second quarter.
[Non-English content].
For the second half, we keep receiving orders coming from the PC clients. We have also made our mix heavier toward the PC side.
[Non-English content].
Part of the reasons is, yes, on one hand, the impact from the Windows 10 cycle, but also the fact that the undersupply of CPU has eased.
[Non-English content].
[Non-English content].
Second question about the hyperscaler and their CapEx. Yes, it is true that their CapEx this year has decreased year-on-year. The problem is that their CapEx is mostly going into infrastructure. It's only a part of that would be spent on server sets or parts.
[Non-English content].
Against this backdrop, the utilization rate of the data centers, it hasn't really gone up sharply. We see that for the TDCs, it's about 30%-40% of the normal level. For the IDCs, it's around 60%.
[Non-English content].
Now, in terms of the recovery of demand coming from the server customers, we believe the key is their inventory level. As was explained earlier, the inventory level continues to go down, expected to reach the normal level by the end of this year. We are expecting the demand coming from the server customers to recover starting by the end of this year.
[Non-English content].
[Non-English content].
The next question will be presented by Do-yeon Choi from Shinhan Investment. Please go ahead with your question.
[Non-English content].
I have two questions regarding the DRAM and NAND cost. First for NAND in the second quarter, there has been the inventory valuation loss. Given the volume expected as well as the pricing expected for the third quarter, then do you believe that you will be able to reduce the inventory valuation loss in the third quarter? If yes, then what do you see as the extent of the reduction in the inventory valuation loss? The second question is, I see that in the second quarter there has been a bigger reduction in the profit margin compared to the reduction in sales. Do you see this situation to continue? When do you believe that this will come to a stop?
Also in relation to this, so given the DRAM and NAND bit guidance for the third quarter, what do you expect to be the cost cutting to be achieved by the company?
[Non-English content].
Yes, it is true that in the first quarter there has been significant inventory valuation loss. For the company in general, it was about KRW 400 billion in inventory valuation loss in the first quarter. In this quarter, because of the meaningful cost cutting that was achieved, the inventory valuation loss was reduced quite a lot, almost to a minimal level.
[Non-English content].
In the first quarter, one of the major causes of the inventory valuation loss was the cost coming from the initial operation of the M15. The initial operation cost of the M15 had been concentrated only on a small number of first products to be produced. Since then, as the M15 operation was stabilized and normalized, the manufacturing cost has also been distributed across a bigger volume, meaning that the cost has been reduced. As a result of this, there has been almost no inventory valuation loss in this quarter, and we believe that this trend will continue into the second half.
[Non-English content].
You also asked about the cost cutting as well as the profit and loss estimation for the second half of this year. In the second quarter, as I have explained earlier, there has been meaningful cost cutting. On one hand, the sales of DRAM have gone up, and also the portion of the 1X class of sales has also gone up, and this has reduced the cost per unit. For NAND, again, there have been increase in sales and also the portion of the 72-layer product has gone up. As a result, this has achieved about a double-digit reduction in cost.
[Non-English content].
For DRAM and NAND, in both cases, compared to the past, we see that the fixed cost is gradually increasing. This means that the extent of the cost cutting per unit that we can achieve is slowing down. Having said that, the company will continue to work on reducing costs by adjusting the mix for DRAM and NAND, and also continuing with an enterprise-wide cost-cutting effort. Such efforts will continue into the second half. By the end of this year, our target is to achieve a cost cutting that is higher than last year's.
[Non-English content].
[Non-English content].
The next question will be provided by SK Kim from Daiwa Security Group . Please go ahead with your question.
[Non-English content].
I also have two questions, mostly about the technology conversion. My first question is about the DRAM oversupply. You announced the plan to shift some of the M10 lines for the mass production of CIS. Are there any further changes to your plan for the production or the technology transfer or the technology conversion? Can you share with us the plan for volume production of 1X nano and 1Y nano? The second question is about NAND, for which SK hynix has also announced a plan to develop and mass produce 128-layer. You also have a plan to mass produce 96-layer products. What is your plan for mass production of the 96-layer as well as 128-layer?
What do you see as SK hynix differentiating point for the 96 and 128 layer products compared to the competitors?
[Non-English content].
With regards to your first question, as was explained earlier, for the DRAM, for the legacy technology, for example, the 20- nano. On one hand because of the fast dwindling demand from the customers and also for the sake of improving the quality of our inventory, we are planning to further reduce the production of 20- nano.
That means that there is going to be a corresponding increase in the percentage of the 10- nano production. As was explained earlier, we are going to increase the percentage of 10- nano from around 40% in the second quarter to around 80% by the end of this year. The most advanced technology so far for the 1Y nano, we are going to begin mass production by the end of this year for computing and mobile applications. About NAND.
In terms of the capacity by per technology, the company plans to have optimal operation of the capacity in consideration of the investment efficiency as well as cost competitiveness. By application, for the 96- layer, the qualification is underway by application. For the 96- layer, we will be expanding our sales in the fourth quarter and into the first half of next year. For the 128- layer, this will be targeting the next generation applications and also high density applications. For this, we are getting ready for qualifications in the first half as well as stabilization of the volume production and sales will begin in earnest in the second half of next year. For SK hynix, between the 96- layer to 128- layer, we would be sharing most of the major devices as well as the technology processes.
There is going to be a very smooth migration from the 96-layer to 128-layer. Also, the investment efficiency for the tech migration is going to be quite high. The company will be able to migrate to 128-layer in a very good timing so that we will be able to strengthen our fundamental competitiveness in NAND.
The last question will be provided by Nicolas Gaudois from UBS. Please go ahead with your question.
Hi. Good morning. Thanks for taking my question. Going back to your comment on DRAM reduction in production that you're planning for into back end of the year, next year. Could you perhaps help us quantify a little bit how much of a capacity in M10 you are considering to moving to image CMOS sensors, and what kind of output in terms of wafers you could therefore consider for CMOS sensors as a whole for you next year? Also clarify whether this is effectively 2Z nanometer capacity being converted to CMOS sensors or is anything else in terms of geometry trend. The second question is also on your plans for NAND flash wafer starts to decline 15% year-over-year this year. If we do the math, that probably means that you will exit the year below 200,000 wafers per month for wafer starts.
The industry, give or take, when we'll be back in full production at some point next year, will be at 1.5 million, 1.6 million per month. How would you handicap what should be over time your targeted supply share? Whether it is okay actually for Hynix to go below 10%, or if you intend at some point when the cycle gets better, to reverse that and go over 10%. Thank you.
[Non-English content].
[Non-English content].
Now regarding your first question about the capacity conversion from M10.
[Non-English content].
Regarding your first question about the capacity conversion from M10. As was explained earlier, we are going to convert the capacity, part of the capacity of M10 for the production of CIS starting in the second half of this year. We believe that the effect will be begun to be felt starting in the fourth quarter, and the effect will continue into the end of next year. In terms of the specific volume of the capacity reduction, it is difficult to give you the number at this point, given that we do not have the investment plan finalized yet, particularly for next year yet. For the M10 operation, because it was focused on 2Z production, if we are to convert the capacity to CIS, that means automatically that we will be reducing the 2Z capacity.
[Non-English content].
In terms of the NAND capacity adjustment, we would make adjustment to the legacy technology depending on the product competitiveness and the cost competitiveness, and it is a natural result of the legacy technology adjustment that the NAND capacity is being reduced.
[Non-English content].
Our investment decision-making will continue to be based on the judgment on investment efficiency and product competitiveness. As for additional changes to the investment plan, currently there are no further changes that are planned.
[Non-English content].
Thank you very much. That concludes the SK hynix 2019 second quarter earnings release conference call. Thank you very much for your participation.