Nanya Technology Corporation (2408.TW) Q4 FY2025 Earnings Call Transcript & Summary

[Foreign Language] Welcome to Nanya Technology's 2025 Fourth Quarter Earnings Conference Call. [Operator Instructions] The conference will be held only in English for investors around the world. Today's conference will be approximately 60 minutes. Nanya Technology's President, Dr. Pei-Ing Lee, will summarize our operations in the fourth quarter 2025, followed by our guidance for the next quarter and key messages. Then Nanya Technology's Executive Vice President, Dr. Lin-Chin Su; Vice President, Mr. Joseph Wu; and Financial Executive, Mr. Philip Jao, will join us as we open our Q&A session. Today's presentation materials are available for download at Nanya Technology's website at www.nanya.com. And as usual, we would like to remind everyone that today's discussions may contain forward-looking statements that are subject to significant risks and uncertainties, which could cause the actual results to differ materially from those contained in the forward-looking statements. Please refer to the safe harbor notice that appears in our presentation materials. And now I would like to turn the call over to Nanya Technology's President, Dr. Pei-ing Lee, for the summary of operations and current quarter guidance. Dr. Lee, please begin.

Ladies and gentlemen, welcome to Nanya Technology Q4 2025 Investor Conference. I'm Pei-ing Lee. The content of my presentation to you first start with Q4 '25 revenue and results and followed by year 2025 revenue and results, then followed by CapEx and bit shipment, market outlook and conclude by business review and outlook. Our Q4 2025, the net sales TWD 30.094 billion compared to Q3 '25 TWD 18.779 billion, Q-to-Q up by 60.3% compared to Q4 last year, TWD 6.575 billion year-to-year up by 357%. Gross profit, TWD 14.759 billion at a margin of 49% compared to Q3 TWD 3.465 billion at margin of 18.5% margin. And operating income, TWD 11.781 billion at 39.1% compared to Q3 TWD 1.119 billion at 6%. EBITDA at Q4, TWD 14.74 billion at 49% compared to Q3 TWD 4.731 billion at 25%. Nonoperating income, TWD 1.602 billion and income tax expense, TWD 2.3 billion. The net income comes to TWD 11.083 billion at 36.8% compared to Q3 TWD 1.563 billion at 8.3% and I'll explain the reason why in the next few for you. Earnings per share at TWD 3.58 per share compared to TWD 0.5 per share in Q3. Book value per share, TWD 54.99 per share compared to Q3 of TWD 50.78 per share. For quarterly revenue result comparison, the revenue Q-to-Q up by 60.3% and year-to-year up by 357%. For ASP point of view, Q-to-Q increased by 30s and year-to-year increased by 80s. Shipment-wise, Q-to-Q increased by low teens and year-to-year increased by 1.5x. Exchange rate, favorable mid-single digit for Q-to-Q and unfavorable mid-single digit for year-to-year. A little detailed comparison for Q4 '25 and Q3 '25 results. And net sales in Q4, TWD 30.094 billion versus Q3 TWD 18.779 billion, up by 60.3%. The reason is ASP increased by 30s and bit shipment increased by low teens and exchange rate favorable mid-single digit. Gross profit for Q4, TWD 14.759 billion at 49% margin compared to Q3 TWD 3.465 billion at 18.5% margin. And the gross profit increased by TWD 11.3 billion, mainly due to higher ASP and higher bit shipment. Operating expense at TWD 2.979 billion compared to Q3 TWD 2.346 billion. Operating expense increased by TWD 0.6 billion. And operating income at Q4 comes to TWD 11.781 billion at margin of 39.1% compared to Q3 TWD 1.119 billion, at margin of 6% and operating income increased by TWD 10.7 billion. Net income comes to TWD 11.083 billion at net margin of 36.8% compared to Q3 of TWD 1.563 billion at margin of 8.3%. The net income increased by TWD 9.5 billion due to exchange rate favorable of TWD 0.2 billion and income tax unfavorable of TWD 1.8 billion and also operating income increased by TWD 10.7 billion. For operating expense, SG&A comes to TWD 855 million and the R&D comes to TWD 2.123 billion. And the R&D mostly is a normal range and SG&A expense marginally increased due to -- we had recognized employee bonus distribution. For cash flow, Q4 last year, beginning balance, TWD 52.313 billion, with cash from operating activity, from the right-hand chart, you can see that the cash from operating activity comes in TWD 15.169 billion, minus capital expenditure of TWD 2.24 billion and then minus the financial activity, basically, we pay out the debt repayment of -- at the bottom note, TWD 8.9 billion. And the end balance comes to TWD 57.738 billion with free cash flow of TWD 12.929 billion. If you pay a little bit attention to the bottom notes, you see that we have paid back debt payment of TWD 8.9 billion. And then we also have a net cash by the end of the Q4, net cash and equivalent comes to TWD 38.5 billion. That means the cash minus the debt comes to TWD 38.5 billion, which compared to Q3, net cash is around TWD 24.2 billion. For 2025, the whole year revenue and result, for 2025, net sales comes to TWD 66.587 billion compared to 2024, TWD 34.132 billion, year-to-year growth of 95.1%. Gross profit, TWD 14.984 billion at the whole year of margin 22.5% compared to 2024, loss of TWD 421 million. And operating income TWD 5.243 billion at 7.9% compared to 2024 loss of TWD 10.555 billion. Operating income, TWD 2.639 billion -- nonoperating income, I'm sorry, TWD 2.639 billion and income tax benefit this year is the expense, TWD 1.279 billion. And net income comes to TWD 6.603 billion at 9.9% compared to last year a loss of -- 2024 loss of TWD 5.083 billion. Earnings per share for 2025, TWD 2.13 profit compared to 2024 loss of TWD 1.64 per share. Book value for 2025, TWD 54.99 per share compared to end of 2024, TWD 53.27 per share. The comparison between 2025 and 2024 financial results. Net sales, TWD 66.587 billion compared to 2024, TWD 34.132 billion, year-to-year improvement of 95.1%. And the remark there is that ASP increased by 30% and bit shipment increased by 50% and exchange rate unfavorable mid-single digit. For the gross profit, TWD 14.984 billion, a margin of 22.5% compared to 2024 loss of TWD 421 million and the gross profit increased by TWD 15.4 billion, mainly due to higher ASP and bit shipment. For operating expense, TWD 9.741 billion compared to last year is very much similar, okay? Operating income, TWD 5.243 billion, a margin of 7.9% compared to 2024 loss of TWD 10.555 billion. And the net income comes to profit of TWD 6.603 billion compared to 2024 loss of TWD 5.083 billion. The net income year-to-year improved by TWD 11.7 billion and exchange rate unfavorable by TWD 2.8 billion -- I'm sorry, income tax unfavorable by TWD 2.8 billion and exchange rate unfavorable by TWD 0.9 billion. A little summary of 2025 sales and EPS by quarter. And on the left-hand side of the chart, the net sales improved quarter-by-quarter very significantly and fourth quarter result is unaudited. And on the right-hand side of the chart, earnings per share Q-to-Q from the first quarter and second quarter loss of total near TWD 2 and the second half of the year, Q3 positive TWD 0.5 per share and Q4 positive TWD 3.58 per share. For CapEx and bit shipment, 2025, our CapEx TWD 13.4, okay, with around TWD 6.2 billion being postponed to 2026. For 2026, we plan to target TWD 50 billion capital expenditure. This is still subject to Board approval. Our wafer CapEx is about 30% of this total CapEx for 2026. The rest of 70% will be used for new fab construction and facility. And the bit shipment for 2025, bit shipment is up by 50% year-to-year and 2026, the bit shipment target to be up by teens year-to-year. For market outlook, overall market, the AI cloud server and general server continue to drive 2026 DRAM demand. Supply shortage may persist for multiple DRAM products, that including DDR5, low-power DDR5, DDR4, low-power DDR4 or even DDR3 due to capacity constraint. Bandwidth and high-density DRAM continue to expand across cloud and edge AI application in the next few years. On the supply side, we are seeing limited new supply capacity are expected for 2026 and first half of 2027. We also see that supplier has been optimized their product portfolio to meet valuable value customer requirements. Demand side, for server demand for high-bandwidth memory -- sorry, as a type layer and conventional RDIMM is increasing significantly. For PC and mobile continue to see DRAM content increase per box. For consumer, we see a wide range of consumer electronic applications, sustained solid demand. For business review and outlook. First, Q4 2025, gross margin, 49% net margin, 36.8% and EPS of TWD 3.58 per share. For the whole year 2025, gross margin, 22.5%, net margin 9.9% and EPS TWD 2.13 per share. Nanya has been getting 128 gigabyte DDR5 RDIMM ready for 5600 and 6400 speed and monodie meet 7,200 speed level. We continue to optimize our supply to DDR4 and low-power DDR4 to consumer market. Our third-generation 10-nanometer and fourth-generation 10-nanometer and also customer-made AI projects are going on schedule. And our new fab equipment move-in are expected in early 2027. Also, Nanya had received ESG recognition from CDP for A list in Climate Change and Water Security. We also received recognition from TCSA for Annual Best Corporate Sustainability Reporting Award, Top 100 Taiwanese Sustainable Companies and 4 Leadership Awards. With that, now we may move to question-and-answer session.

Yes. Thank you, Dr. Lee. [Operator Instructions] The first one to ask questions, Nicolas Gaudois from UBS.

In the current environment, do you actually see customers starting to despec as they may not be able to get exactly the volumes they need on the DRAM side. And if so, in which markets are currently seeing any signs of despecking. And I've got a couple of follow-ups.

Okay? [ Despecking ] case, we do see some, but not a major of the customer doing despecking. There are some [ despeck ] maybe in the TV section. Some of the TV users may [ despec ] from their use of DDR4 and low-power DDR4, down to DDR3. But this is not a major quantity. This is a limited quality.

And on your side, so you just talked about mid-teens bit growth for full year. Going to the first quarter, I mean, what could be a reasonable range for bit shipments for Nanya Tech?

Bit shipment wise, Nanya would not be Q-to-Q increase substantially because our -- we are in full capacity already. Okay. And also with the depleting inventory, likely we will not be seeing Q-to-Q shipment increasing substantially. However, year-to-year, we will be seeing some bit shipment increasing, mostly due to beginning of last year, the bit shipment is low, okay? And that is make up in this year.

Okay. Great. And lastly on my side, considering the environment we're in, I know initially the plan was to roll out new capacity with a new fab in 2027. Is there anything you can do to pull in the time line potentially as clearly you're effectively capacity constrained in the current environment?

Yes. As I described, we currently -- our new fab, the [ shaft ] is almost ready. And we will be doing facility clean room for this year and start to bring in equipment beginning of next year, okay? And hopefully, by that, we can have some output by the end of first half and gradually ramping up in the second half. Sorry. Maybe Nick's question, I have a little bit of additional comment is that in the beginning, we are not targeted full fab loading, we are targeting for -- between 15,000 to 20,000 of first phase of capacity loading.

Okay. Now the line is open to Charlie Chan from Morgan Stanley.

Also, congratulations for very strong results. My first question is that, do you plan to change your price quotation behavior because one of the major Korean fab since this year, they changed to monthly quotation. I know you kind of honored all those contracts with your customers, right? But reality is that there are very strong demand, double booking, et cetera. So I just want to double check, do you want to change the contract negotiation to a monthly update.

As a matter of fact, no, we don't. We don't change our contract behavior. For long-term customer, we will stay with the current price quotation, okay? And as a matter of fact, with the uptrend, most of customers want to secure longer term of the negotiation. We are working with the customer case by case to see their need and what we can do for our customers, okay? So the answer is no, we don't change our behavior. And the second question is double booking. That's a very good question under the current business environment. It's very possible that a customer would like to secure more of the -- more deliveries so that they can secure -- making sure that their production will be going on well. And therefore, we have to work closely with our customers. Our sales has looked into this very, very closely for many years with our customers. So we will be looking into customers' typical normal running rate and going from there to arrange our location to customer. And we want to make sure that customers do have parts to run their fab, run their business. We want to prevent customer putting their parts into our parts into the inventory into the [ shelf ], okay. So that is all been going around worldwide, okay? And our sales individual has been working on that closely.

Okay. Thanks, Lee. My second question is about, do you see that a potential supply chain disruption because of lack of DDR4 components. If that's the case, which area do you see the urgency. So we are hearing automotive, networking, even AI server, general server BMC, whatsoever, right? So lots of critical tech device needs your chip. So how are you going to make sure that wouldn't cause the supply chain disruption? And is that the case, do you think your competitors no matter Korean or Chinese competitor to extend their DDR4 offering? I know some of them since want to do ELL, but do you expect any extension of the offering of those DDR4?

So far, we don't see our supply chain disruption yet, okay. For the critical market sector, as you mentioned, BMC, NIC those is the sectors that are likely to be priority for [ field ], okay? And then you mentioned that the supplier end of life, okay? And they may extend their residual capacity. That's well known, okay? And actually, it's very good. They do extend that to making sure that customers do have sufficient parts, okay? So point is that for those production line has been relocated, we locate it to different production. It's very difficult to moving it back and forth. So the best that most of suppliers can do is to continue to extend the lifetime of residual capacity.

Great. So one financial question, if I may, to ask CFO, Joseph. So can we get your latest depreciation schedule for 2026 and 2027. So that's why you can calculate your future gross margin trend?

Yes. Regarding the depreciation rate, as we mentioned already that in Q4, we are comparison in lower depreciation results. And moving on to the 2026, I think we will keep the similar trend continues this year. So that's the check, as we mentioned already.

Because the existing 2 fab, [ 3A and 2A ] fab, there's quite a few equipment already depreciated and more and more will be more depreciated. So we're seeing our depreciation going to be improved gradually also.

Can you quantify like how much year-on-year in '26 and '27.

Yes. In '27 because of the new equipment moving and the new fab, we need to recalculate in detail later on. Regarding the 2026 is lower than 2025 maybe 10% around.

Okay. So 10% decline in '26?

Yes.

Okay. Okay. Great. Yes, we will ask other questions in the second round. Thank you.

[Operator Instructions] Next one to ask questions, Simon Woo from Bank of America.

Okay. Great. Congratulations, Dr. Lee and Joseph. Great results, right? The key contributor, obviously, the ASP strength for 2 consecutive quarters. So 40% of last quarter -- I mean, the Q3 and then another 30% of this quarter. So what's the next -- any rough idea, the price trend these days, Q1, Q2 for 2026?

That's a tough question. The reason why the price went up on Q3 and Q4 substantially, one of the key reasons is that price when it started to recover is already way, way, very long, okay? And I mentioned that a couple of times to many gentlemen here is that because of the price downturn since 2023 -- actually 2022, 2022, '23 and into 2024. And 2024, by the time of HBM -- middle of 2024, by the time of HBM becoming good. HBM pricing recovered quite nicely, but the rest of the market other than HBM, which is over 90% of the product continue to go down until middle of 2025, which means that ASP-wise, for this 90% of product that we've been talking about HBM going down in pricing for almost 3 years, okay? So it's at a very low point, start to climb. So you see that significant improvement on Q3 and Q4, okay? And on top of that is that, of course, this is all driven by demand and supply. Okay. And moving on from here, we are seeing that price in Q1 may continue to improve. As to which extent of it going to be improved. We cannot be describing that now because there are only a few couple of months to go, okay? But based on the long-term contract, we can see that the price will be upturn in Q1 as well, okay? Likely will be some exceeding 10% at least, okay, price uptrend.

Compared to the December quarter, right?

Yes. Could be in the 20% plus/minus, yes. But again, this depends on 2 more months before we can be sure of what is the actual percentage on Q1.

As you pointed out, the U.S. big tech companies, they are buying HBM more annual contract basis, but not sure for the overall the conventional memory. So any rough idea for your sales, what percentage for annual or multiple year based contract versus quarterly versus monthly or a little bit spot trading or any rough idea maybe December quarter revenue mix?

For multiple years, the deal is very limited, very small number. And as of the market uptrend increasing, we've seen the quarterly deal become more, okay? And, however, I cannot give you a specific number because this is very much on each case-by-case negotiation, okay? And we still have quite a few percentages relying on month by month. Consumer side, they are percentage of month by month is still higher than quarterly.

The consumer more monthly, but nonconsumer still quarterly.

That's still a percentage-wise again.

So not yet many the annual multiple year contracts with auto OEMs or some industrial guys?

That is, of course, is case-by-case could be treated differently. But just speaking, yearly or multiple year is limited.

Yes. Very quickly. So today, you mentioned the new equipment moving for the new fab early 2027. That means your shared that already their infrastructure power, water was already there. I mean, to be done through this year in the new fab?

New fab, the power and water already prearranged okay?

So if you fill out the new shell fab, what's the total wafer capacity there 700,000 possible? Or the 50,000?

The new fab will be 45,000 wafer per month. And our first phase is target for 15,000 to 20,000 per month.

You must have at least 1 EUV layer, EUV will be there, right, for the first time?

No, our new fab will come -- will start to manufacture with 1B, 1C and 1D second, third and fourth generation of the outset developed 10-nanometer process technology that we are sure we do not need EUV. Although we are developing EUV technology as we speak, using current generation to do the work, okay? And we have proven that we can do it, okay? And the matter is that when it's suitable for us to bring in equipment to produce.

So also the question from some investors they are asking to Joseph, CFO, whether you are considering dividend payment with a very good second half result. And then another investor question is whether you can break down your revenue, DDR3 versus DDR4 and a little bit DDR5, thank you, appreciate.

Okay. Regarding the product ratio, right now, we are more focused on the DDR4 and the low-power DDR4 is the major. But the DDR5, we will keep the running ratio is around 10% or a little bit be higher. And the remaining will be the DDR3 as well. So that's the product mix. And regarding the dividend payout ratio, our guidance is like 50% around, we will depend on our shareholders hoping and we will propose to the Board a meeting and get the final approval.

5-0, 50% of the maybe...

45% to 55% is our guidance, general guidance.

Even for the 2025 results?

That we're still trying to figure it out. Okay. It could be a little more on the upside or we are still discussing on that. But important point, Simon, is that as we're going to be building new fabs, we're going to be spending a lot of money in the construction as well as the equipment for the future, okay? Likely, we will be preserving some cash next year. But this year, could be a little bit more favorable for this distribution, yes.

The market is already getting bigger, so investors should be happy. I'm joking but basically Dr. Lee, so you are saying DDR5 portion is only 10% or a little bit higher in Q4 sales, DDR5 10%?

Yes.

But DDR3 may be around 40%.

20%.

sorry?

20%. Yes.

Yes. So the rest, around 70% is LPDDR5 or regular DDR5 or together, right?

DDR4 and low power DDR4.

So DDR3 20%. DDR4 should it be what percentage, DDR4 only?

DDR4 plus low-power DDR4 around 70%.

Okay. Okay. Okay. 70%. Okay, makes sense. So 10% for DDR5, DDR3 20%, the rest 70% should be the DDR4...

So that's including DDR4 and low-power DDR4.

This is a revenue basis, right?

Well, that's a shipment basis.

Shipment basis.

Yes, shipment.

But DDR3 price premium very high. So the DDR3 revenue portion should be much higher than this then?

Yes. This is -- DDR5 is a strategic product and technology that we have to move along prepare for the future, okay? So that's -- we'd like to grow DDR5 if the market is favorable, okay? And DDR3 is more of the residual legacy consumer supply.

Next one to ask questions, Jay Kwon from JPMorgan.

Thank you, Dr. Lee for having an opportunity to ask a question. I'd just like to ask building up on previous the question. About the DDR5 mix, you said it's around 10%. And I believe the pricing for the probably DDR4 on average is currently higher than that of the DDR5. Does it mean with also your stronger DDR4, probably [Technical Difficulty] will be largely unchanged throughout '26? Or do you have any specific target by year '26 that you are targeting? That's my first question. And also second question, it's been a while since the DDR4 pricing momentum has been stronger than the DDR5. By when do you expect the DDR4 pricing to increase momentum to be still stronger than that of DDR5? Any guesstimate or any range of probably price increase between the 2 for the first quarter, as you pointed out, it will be roughly 20% plus/minus range?

Your question is compared between DDR4 and DDR5 in ASP. And we've seen that DDR5 pricing also improved quite substantially for the past few months, okay? And as I indicated that DDR5 is our strategic product and process development that we will continue to pay attention to and we'd like to grow DDR5 if market is favorable, okay? And on top of that is for DDR4 we continue to support our customers and it's the best that we can do, okay? And if anything possible is that we can reduce DDR3 if the market size shrink to a certain degree, okay? So overall speaking, we're seeing that the market is tight on all markets, okay? DDR5, DDR4, low-power DDR4, [ SOAT ] DDR3 for now, okay? And the pricing will be very much depending on individual demand and supply sector product portfolio, basic product mix, okay? And our strategy is to continue to support our customers even though the margin may be vary from one side to the other side somehow. But strategically, our goal is to make profit, but as well as making sure that our key customer requirement is fulfilled.

If I may, just one follow-up question, then would you be able to share the key considerations of your -- the product mix strategy between the DDR4, LPDDR4 and DDR5, can we safely assume the mix will be largely unchanged throughout 2026? Or do you still put DDR5 as a strategic product up to increase the shipment mix?

As I say that we'd like to maintain at least 10 percentage or more. If there's a chance to grow DDR5, we will grow DDR5. And hopefully, it's at the expense of DDR3 instead of DDR4.

Next one to ask question, Charlie Chan from Morgan Stanley.

So definitely, you mentioned about you will grow DDR5 mix if the market is favorable, right? So I guess the 2 part of the question. So first of all is when do you think this DDR4 up cycle will peak? And second, I want to get your opinion about the DDR5 supply, especially Micron seems to send a contract to convert power chips fab to the future DRAM production, likely DDR5, right? So what do you think about the DDR4 and DDR5 cycle in the coming 2 years?

Yes. I think your question about the news on Micron and power chip deal. And I think that's quite normal, okay? And this historically in DRAM industry, there's always partnership, always consolidation as well, okay? So that in the market, looking for partnerships, developing new business opportunity, that's quite normal, okay? And One topic is that requires some time to build a new facility, new equipment, new capacity that we cut some time, okay? And also, the initial cost for that is pretty high, okay? So that's a factor we factor that in DDR4, DDR5 supply may not be substantially changed for next at least 1.5 or even 2 years.

Okay. Yes. So my next question is a little bit harsh because I feel like you honor all those...

kay. If it's harsh, I will excuse it.

Okay, let me try to be more modest. But anyway, because I feel like your industry peers, they try to monetize to maximize their profits as much as they can. Because given the down cycle, customers are like, I don't want to stay will be okay. It's your own industry problem. So this is 1 out of like 5 years or even 10 years, they can finally receive back all the investments. But based on your answer to your questions beforehand, you don't want to change the price negotiation behavior. So do you think you kind of underestimate your kind of pricing power. So your earning power may not be as great as other DRAM players in this super cycle?

I think to be fair, Charlie is that for last 13 years, Nanya had made profit for 11, okay? We lose money in 2023 and 2024, okay. And you go back, you'll find that the major supplier also suffered a lot on 2023 as well, okay? So overall speaking, that Nanya is doing okay, okay, even though we suffered quite a bit on 2024 when the market recovered with HBM and DDR5, okay, and we are behind on that, okay. So to be fair, I think that we need to improve ourselves instead of complain on the market or on the demand and supply okay? And from a profit point of view, of course, we are looking for profit for sure, okay? By saying that we want to meet customer requirement also that's important because a customer is the one who will buy our product, and that's the opportunity for us to do business and working together to gain profit. Customer gain profit, we gain profit as well, okay? So I don't think your question is too harsh. So I answered directly without asking Joseph to answer.

Okay. Okay. That's great. So just some feedback, right? I think industry average price hike could be 50% to 60% in 1Q based on what your industry peers quotation. So I wanted to clarify, when you said the 20% price hike that they refer to fourth quarter, not 1Q, right? 1Q, you haven't really came out with the price hike magnitude, right?

They are so many product portfolio, okay? So I saw some of you report actually you cannot take 1 product as a number and make it across the Board, okay? For instance, we're doing DDR5. We're also doing the DDR3. And for customer needs, we're doing some low power DDR4, some is very profitable. Some is not as profitable as well, okay? So for generality, if you take 1 number and bring it across the Board, your number will be far exceeding the reality that we may be able to deliver. Okay. I'd just like to remind you, when you're doing your model, you need to take into account of the product portfolio.

Okay. Okay. That's fair. So again, to clarify, 20% you were referring to last quarter and not the current quarter, right?

The 20% plus/minus. I'm referring to Q1 this year, okay? And maybe that will be a slightly plus or minus situation. I would say mostly will be more than 20 plus instead of 20 minus, most possibility.

Okay. Okay. Sure. Yes. So I think your pricing strategy will continue into 2Q, right? I mean...

Yes. We will -- as I described, we will keep our pricing strategy. We don't want to -- too much of opportunities. We need to work with our customers on their need, okay.

Okay. Okay. And thanks for your guidance about how we should do the real assumption for ASP. And lastly, as we ask like every quarter about your wafer-on-water, this AI opportunity. Can you provide more details?

Yes. That's a very good question. As I described that our third generation, fourth generation development on the process and product is going on schedule. So is our customer make AI projects. That's going on well and actually it's a little bit ahead of our schedule. Originally, I expect that second half this year, we may be generated revenue. Likely, we will be generating revenue first half of this year.

So again, it's both cloud and edge AI?

Yes, cloud and edge AI, depends on customers' application. And for that is the customer confidentiality and that I cannot comment more.

Okay. Let me try and feel -- if you don't feel comfortable to comment, then we can skip. Yes, because smartphone and also human [indiscernible] are very important future edge AI market, which one or both you think you will have a potential opportunity.

I would say both cloud phone and also even a smart [ robot ], smart PC, they will gradually increase their AI environment, okay? And in longer future, you will start to see that AI move into automotive, okay, as well as robotic, all have wonderful AI application.

Next one to ask question, Nicolas Gaudois.

Could you confirm to us that you -- when you'll be starting mass production for 1C-nanometer? And I think you also alluded to starting -- piloting on 1D-nanometer in the first half of this year. Is that still the case?

1C we're targeting by the end of this year, we should be able to start our other qualification, we are targeting for the end of this year.

And how about 1D for piloting?

1B, we are already fully in production.

No, no, 1D?

1D, maybe 1 year beyond.

For qualification?

Yes.

And the next one to ask a question Jay Kwon, JP Morgan.

Just one last quick question, if you don't mind. The Nanya Tech had talked about the 16 gigabit DDR4 model that with a very long life cycle for a while from consumer side. Are you still seeing the trends continuing? Or is there any trends of a customer migrating to DDR5 or even probably the very tight market, customers have no preference in trying to still secure as many DDR4 as possible?

Yes. They are a customers plan to move from 16 gigabit DDR4 to DDR5. But also they are customer decided to stay in DDR4. And that's also true for low power, okay. So we're seeing that the demand for 16 gigabit DDR4 and 16 gigabit low-power DDR4 are pretty strong. And Nanya is preparing to do some sampling towards the end of this -- first half of this year. Okay. So now can we move into online question. We have the online question from KGI Securities. Michael Shen. He has 2 questions. His first question is we have observed a meaningful improvement of variable costs per bit in Q4 compared to previous quarter. And he would like to ask what is the reason? And basically, we -- our cost has been coming down Q4 versus Q3 in depreciation, okay? Q-to-Q depreciation-wise is coming down by almost 20% and moving on as Joseph just described, moving on for now, maybe 10%, okay? So depreciation is a key and also product mix and also other minor cost reduction as well. And second question is regarding to the news about Micron transferring process technology to one of the DRAM peers, then do you believe this could have impact on the supply-demand dynamics of DDR4 and low-power DDR4. Now if so, at what point in time might such effect begin. And additionally, given that the competitor has sold its fab, but may receive Micron's process technology, how do you view potential implication of competitive landscape of wafer-to-wafer going forward? I guess it's not suitable for me to comment too much on the competitor or the previous partner, their business actions. However, I could share from our own experience, as I described just now, history in DRAM industry, there are many cycles of partnership as well as consolidation back and forth, okay? And Nanya do have experience working with few partners, including IBM, including Infineon, including Qimonda and including Micron, okay? And personally, I have been deeply involved with all these projects, okay? And all these projects from the -- from Nanya side, we do see that it's going to take time. It's going to take cost, take money to build the capability, okay? And likely, because of that, we're seeing that the impact may be 1.5 to 2 years from now. Okay. And then we have the next question from Capital Securities by Liyen Chen, okay? And the question is, what is the current year ramp-up status of the 1B process? And what is the target bit shipment contribution for '26, okay? And our 1B is already in full production, okay? And yield is already pretty reasonable at a pretty high range, okay? And our target bit shipment is -- will be around 40% of output shipments. We're looking for that percentage, okay? With the industrial-wide shift of capacity towards HBM how is Nanya Tech capitalized resulting shortage -- supply shortage on DDR4 market, okay? To answer your question, Nanya in on twofold, okay? First, we're not totally give our HBM. We are working with our customers on what we call customer make AI project that's going on well, okay? Second, Nanya has been optimized our capacity and delivery to DDR4 as well as low-power DDR4 market as we just described past few minutes. And your question 3 is what is the projected revenue mix for DDR5 by the end of 2025. How will this impact the overall branded average selling price? Okay. And I also commented that on my previous answer to the question is that DDR5 is Nanya strategic product line that we will continue to support and develop. If we see opportunity to increase our contribution to DDR5 market, we will do so, okay? And overall speaking, we are looking for at least 10% or more. And yes, it may be impacting somehow, okay, on DDR, overall ASP, okay. However, DDR5 pricing also improved substantially over the past few months. And your question 4 is regarding to customer -- customized HBM strategy, what is the status of collaboration and with the ecosystem partner for edge AI and automotive application. Our partner decide their application, either it's going to be in the cloud or in edge, in PC or mobile phone or any other automotive, okay? Our customer decides that, okay? And so far, that has been going on well, as I said, it's ahead of my previous expectation. Originally, I was expecting that second half, it may start -- this year may start to generate some revenue. It looks like it's ahead of schedule. We will be generating revenue in the first half of this year. Your question 5. Well, long list, what is the 2026 CapEx for new fab equipment move in? 2026 CapEx, as I described, will be 30% out of our total CapEx of TWD 50 billion targeted. And your question 6, given the return to profitability, what is the Board current stance on 2026 dividend policy? If you're talking about earnings from 2025, okay, likely will be a little different from earnings from 2026. And for 2025, first of all, our general guidance is between 45% to 55% of our earnings, okay? And that's the general guidance. For 2025, we may be looking for higher percentage. And for 2026 earnings, we're looking for lower percentage to preserve cash for our future CapEx requirement for new fab. And the next question is from Yuanta Securities by Michael Hsu. You have 4 questions, okay? Let me quickly answer one by one. Had the DDR4 16 gigabit start to ship to client? No, we target to ship towards the end of first half. Any update on mass production schedule about wafer-to-wafer? Yes, we had done -- we are moving the equipment, and we have do early engineering setup, and that is going on schedule very well. And your question 3, have company signed LTA? If so, what percentage of capacity booking for LTA? Yes, we signed LTA. However, I cannot give you the precise percentage of LTA, okay? And this will discuss case by case as needed. And Q4 -- just question 4 is financing [ plan ] for 2026 and 2027, looks at launching debt or new shares. Both are possible. First of all, looking into our current net cash situation is that by the end of 2025, we had net cash of roughly TWD 1 billion -- more than TWD 1 billion -- and we are looking for by this -- by the end of this year 2026, we will be generating another 2 billion from Nanya's net cash point of view. So likely, we will have around TWD 3 billion of net cash by the end of this year. So with that, we have at least large percentage of what we need for building a new fab, first phase. However, we will continue to consider that and new share, okay? And that will be our range in the future. Okay. With that end of all questions. Thank you so much for your questions and comments.

Yes. Thank you, Dr. Lee, and thank you, ladies and gentlemen. That concludes our conference call today. Please be advised that the replay of the conference will be accessible within 3 hours from now, which will be available through Nanya Technology's website at www.nanya.com. We hope you will join us again next quarter. Thank you for your participation, and have a wonderful day. You may disconnect now. Thank you, and goodbye.

Thank you. Thank you for your support.