Nanya Technology Corporation ($2408)

I'm not sure this group meeting for first 15, 20 minutes, we are -- we will be taking actually opportunity from Dr. Lee, the Chairman of Nanya Technology to give us his view on the memory sector outlook as well as Nanya business update for about 15, 20 minutes, and then we'll go into the Q&A. Just be reminded, this is not a video -- actually recorded call. This is only for the post recorded call to comply to the stock exchange request to actually unload the business meetings officially with the broker conference. So just 20 minutes, it won't be recorded. I will now hand over to [indiscernible].

In general, this is to protect all investors trying to make the information. So it's sort of fair treatment to general public -- general shareholders.. So I start with a couple of -- to share with you the general market situation and then the supply application, et cetera, and also get into a little bit of Nanya outlook and business view. So first, please. So first of all, this is the representation of the total DRAM application worldwide overall market, not one specific company, but 6 major applications here, including server, space, this is in the cloud and phone, 25 -- [ 31 ] going down to [ 25 ] and PC around 10-ish. AI HBM, this is also in the cloud. So the cloud is combining AI HBM and the server. As of last year, it is 35% plus 7%, 42%. And this year, likely going to 10% plus 52%. And the reason why is that the consumption there is increasing very substantially even with the phone, PC and others, maybe speedwise still increasing at much less percentage compared to the cloud area. Therefore, their percentage come down. So one of the very important information to justify this growth is on the cloud CapEx, the service provider CSP. Most of them are heavily invest in AI, in their AI server as well as the general server to support the AI operation, that's including training, inference, et cetera. And talking about AI in the cloud, AI is the inference becoming more and more important, likely AI will also penetrate into edge area, which means that in the phone, in the PC, in the general consumer may have more AI function built in when there's more GPS, some local AI function will have to build in local device. That's what we call edge AI. With that, that will also need to have more installed inside the local device. So with this trend going up that we may say that AI trend will continue for a few years, including in the cloud and edge development. And on the right-hand side here, what we try to describe here is the overall DRAM consumption worldwide in what type of memory. And this is the unit base. This one here is a bit base. So unit base-wise, DDR4, low-power DDR4, if we look into last year, 2025, combined of this DDR4 and low-power DDR4 is still in a good 30% range. So that's why in the middle of last year, when major supplier announced that the end of life, these 2 products becoming a big impact to the industry because the demand size is still relatively high in terms of units. And when they make an announcement, they mostly have made the transition on their operation, which means that they have to move the equipment, move the product, move the business sector operation-wise in terms of their marketing sales or do preoperation management before they make that announcement. So once they make an announcement, it's very difficult to switch back because once they switch back equipment-wise, it's going to be hurting the new production line and the old production line. It's very little business. So that the base that capacity can remain in that sector for those companies is to have their residual supply in this area in their capacity, taking longer time to completely phase out. However, if we're looking back on historical DDR3 time frame, DDR2, DDR1, SDR and even base stage time frame, always this kind of legacy becoming smaller and smaller market size, always the big supplier trying to phase out as soon as they can because becoming a key is operation for them. It's not a very sweet operation for them. They have -- may have to put in a lot more attention and getting revenue size-wise is not as much as they want to see. So if we're looking down on this chart here, '25 [indiscernible], '26 is still sizable, even '27, '28 still have some percentage in unit-wise. So the gap for DDR4 and low-power DDR4 likely will stay some time. And if we're looking back on this left-hand side chart, we can look into each of the application. For PC, it may be easier to go from DDR4 to DDR5. And that's actually mostly done, not 100% yet, but mostly done. In the cloud area, it could take a little bit longer time because there's still some CPU still built for DDR4 out there from the 2 big suppliers. And on top of that, in the cloud area, nowadays, the server rack, they require a communication chip, they require management chip, management board, communication for the whole, they still need DDR4 in most of those areas, okay? So cloud area will take a little more time for DDR4 to go down. And we come back to the phone, likely they will go to low-power DDR5 instead of low-power DDR4. And for automotive, likely will stay on low-power DDR4 for a longer time, even though some may transfer to low-power DDR5 already, but because of chipset reason, the reason likely that would stay longer for automotive and lowpower DDR4. Consumer will stay even longer for DDR4 and lowpower DDR4 for a reason that consumer, typically, they use 8 gigabit instead of higher density. If they want to go to DDR5, the minimum density they can get is 60 gigabit. So likely consumer will stay in DDR4 for a longer time. There are some consumers using DDR5 only on very high-end model and the quantity there is not that big. So in general speaking, that's the reason why the DDR4 -- the downtrend for DDR4 and low power DDR4 is already -- is staying pretty precise. And this is already considered that this is mostly a forecast, but already consider that when there are maybe some sector of business, they may be reducing the operation. For example, the PC, the low-end PC may be going lower in terms of because of end market may not be able to support. So low-end PC, maybe less quality, low-end mobile phone, maybe less quantity. So overall speaking, this is a general trend. And important point to make is that if we're looking into this situation for the AI development, likely AI development will go on for some time. And no matter who is doing it, NVIDIA or Google, GPU or TPU or LPU or NPU, whatever you want to call it, they all have to use memory, that is DRAM to make AI happen. So this chart is a very good reference at how AI going up. And in the future, when AI may be saturated we continue to look into this chart, may be able to see it as a good reference, as a good index. Now come back to Nanya product portfolio. We do PC, consumer low power, MCP, eMCP, this is combining DRAM and NAND Flash and automotive and industrial is about 10% of our business. We're also doing KGD. That means a longer time, we build high-quality wafer, that don't have to dies it wafer and high quality enough that we burn in that we sell wafer directly and our customers can take those high-quality chip directly bound together with their couple chip. And this is also a good size of business, 15% or our business in this area. And that's the customer like NPK, Realtek, Novatek, like Qualcomm, like mLogic, those customers, they require this supply, we will serve them with high-quality wafer, so they can directly make it a very [ credible ] time to market for them. And their customers do not have to requalify the DRAM. Everything is packaged together. So it's a very efficient business model for them, okay. And lastly, the customer memory, this is where we work with our customers, combine our ability in memory with customer ability whatever it is in the market space for others. For instance, we already have been working with customers on AI memory, high-bandwidth memory, where we design high-bandwidth memory. And they design their logic die and combine together, for AI application. And we already have some revenue generated, but not very big yet. It's in the beginning. It's just expanding, not great yet, but we're looking forward for that business to grow as well. So Nanya business is going also into AI gradually in the future. And this is, in general, the product area, the application area that we've been doing. But of course, this is just a typical example, we are doing more than this. So overall speaking, we have 800 customers worldwide. And the business area is mobile, automotive, PC server and consumer in relatively percentage. And what we call consumer is anything other than these 3, we will put it into consumer. For instance, SSD in the cloud, we put in the consumer. Networking is in consumer. So a lot of the consumer is relatively from this area. And our product is more than 40 product concurrent in the market, and we are working on the customer AI chip project as well, more than just this product. So this is a very quick chart to describe to you how Nanya performed in the last 13 years. And this chart give you an indicator that a couple of things. First of all, the top chart is the income. Over the past 13 years, Nanya was making profit for 13 for 11, out of 13. Two years, we will be -- we were losing money. However, if you look at this 2 years losing money, 2023 is a very bad year for memory industry. All companies losing money. As a matter of fact, we lose TWD 7.4 billion. The big 3 companies, they lost TWD 180 billion to TWD 340 billion if you convert to NT dollars. So it's much bigger ratio compared to Nanya in terms of size, in terms of capacity. During this year, Nanya actually suffering and losing money pretty bad, TWD 7.4 billion. The margin will go down to minus 45%. That means that we sell one chip and giving another chip free and customers still not happy during those times. But one thing I want to indicate here is that a couple of days ago, because our share price is fluctuating in the market, we are requested by TWSE, the Taiwan Stock Exchange Commission, they demand us to make disclosure on our February results. And because of that disclosure on MOPS, so I can disclose this to you today. Otherwise, I won't be able to say anything about our performance. But anyhow, by February, 1 month, we were making $7.5 billion profit, which means that last month in 1 month, we make a little bit more money than 1 year, the toughest year in the last 13 years. On March, we will make it up for the whole year. So that indicates that the market has been recovered. But also one very important point is that here, minus 45% margin, which means that the ASP has to grow 100% before we break even, okay, before breakeven. And nowadays, because it's getting into profit range, it means that ASP is coming up quite substantial. And that's one point I want to say, bottom chart here describes our cash situation. And this is for the past 13 years, our net cash -- net cash means our cash minus our debt, net cash and that situation getting better and stay positive for many years. At the end of last year, our net cash is TWD 13.8 billion. And this year, we will be spending TWD 50 billion CapEx. So with the market outlook, we are anticipating that we may be spending TWD 50 billion and net cash of about TWD 100 billion left at the end of this year for future growth. And that's the current outlook. That's the forecast. So the cash situation and financial situation for Nanya are relatively healthy. So next. This chart, I say a lot just now on the market. So if you have a question, I can come back to this chart. Next chart, please. So a little bit summary of our last Q4 results. As I described to you that last Q4, our EPS earnings per share is about TWD 3.58 in 1 quarter. That's a very remarkable recovery from beginning of last year to the end of the last year, Q4. And looking because we are able to disclose February 1 month result, 22.5%, you can image the Q1 profit will be much bigger than this. And I guess that probably give you a little bit of indication of how that's going to be. So that's one thing that I would like to comment. And the second thing I want to comment is that Nanya is also capable of doing DDR5. We are shipping around 10% of our output in DDR5. We are capable of doing a lot more than 10% in terms of our capacity readiness, our process product readiness is all there. But we are not doing that much for the reason as I explained the gap in the DDR4 and a lot of customers, they are in tough position in their production in their business. So we are trying to try our best to optimize our supply in DDR5 and DDR4, mostly to serve their requirement. And we are preparing for the future. We don't stop preparing for the future. So our DDR5 is now capable of going up to 128 gigabyte which is on top of what we're talking about high bandwidth that we're working with the customer. And we're also working on our next-generation process technology is a third-generation [ z-developed ] process technology. This is already in pilot and already have some early yield, and we expect that this will be maturing as we speak over probably by the end of this year and first -- beginning of next year, we should be able to do more third generation. And our fourth generation will be into pilot next quarter. And hopefully, we can make it more ready by the end of next year or 2028 when our new fab coming up, we can have more of this new generation is implemented in the new fab as well.

Is that 1C equivalent to Samsung's 1C?

1z maybe equal to 1A. We call it a long time ago, and they call it -- so everybody call it the a, b, c, small a, b, c, not capital a, b ,c. So everybody call it [indiscernible]. And for Nanya, we are able to use this technology and product portfolio to maintain certain competitiveness in the market, and we're able to generate profit within 13 years, we are doing okay for them. So for ESG recognition, Nanya has been working on ESG for more than 10 years. We have quite a number of international recognition. And here is only 2 examples, the CPP climate change and water supply and also the TCSA corporate sustainability is one of the couple of examples, we are recognized by Dow Jones and many others. We're doing okay in the ESG area as well. And thank you for your attention.