Phison Electronics Corp. (8299) Earnings Call Transcript & Summary

Thank you. My apologies, my video isn't working. But this is Randy Abrams from Credit Suisse. I want to thank everyone who's joined the line today. We're pleased to host the English session for Phison's Third Quarter 2021 Financial Results and Outlook Call. Format for the call today, we'll do to a presentation. So KS Pua, Chairman and CEO, Phison, will go through a presentation on the business. And then he'll start with some common questions that have come in the meantime, and then we'll open up the floor to Q&A. So with that, I'll turn it over to KS and then we'll come back to the line for Q&A later. Thanks, KS. You can go ahead.

Thanks, Randy, and thanks to, ladies and gentlemen, to attend Phison's 2021 CQ3 result announcement. In order to keep the good quality voice, I'm going to shut down my camera to make sure voice quality is good. Phison, basically right now, we are in silence period because of we issued a CB by August 6. Even though we are not -- we are not ready to raise the CB, but anyway, this is a silence period. So I'm not able to give any guidance in number. Okay. So this will start Q3's results. I think we all believe 5G is first starting year at 2021. And we already can see the 5G trigger the demand of storage is unlimited. We believe in coming few years, the demand in the storage, basically is a NAND storage is not any ceiling. Every systems from the edge such as smartphone, the IoT device and the data center for every kind of apps, social media keep looking for more storage to make sure the quality and also the speed, everything can be good. So Phison, we believe, because of 5G is going to trigger a lot of application in the NAND flash. At Phison, we cover, looks like almost every segment, which is including the auto, gaming, embedded ODM, industrial and also the enterprise data centers. In first year 9 months -- of '21 first 9 months, Phison's revenue almost is similar as the last whole year. So we are going to have a very good achievement by 2021. I would like to take this as 20 years history to explain what -- why Phison's first 10 years growing while we're flat for the other 5 years. Now we are trying to picking up. We started the company by year 2000. We developed the first -- the SoC for the USB drive. This is a totally new application for external storage. That nothing happened in the market. So Phison enjoying from nothing to mature. So we hit the $1 billion in 10 years, okay? Phison, is like the time we are local king in the external storage. But going to 2011, the external storage, even though the volume increased, but price dropping 50%. So we are keeping flat. And NAND flash start to go to the embedded system from iPhone by late 2000. Then start penetrate to a lot of application, but unfortunately, Phison was not able to get any business for embedded NAND. So what we try to do is from 2012, Phison started to make investment, heavy investment in the R&D. Start to develop the eMMC, the SSD industrial. Finally, from 2016, we're able to get some embedded business. We have a steady growth in the coming years -- in the coming 5 years. But after 2020, Phison already covered most gaming platform applications. We are now going to touch the enterprise data center application, SSD. We're already in the automotive, but because of the ASP in controller is low and the gigabit use in automotive is still low. So in coming years, we believe the auto drive -- autonomous drive will be the mainstream and they are looking for terabyte SSD. And we are now already developing for our semicon partners. And also we are going to build the SSD for our auto partners. And industrial is -- keeps slightly improving, but the most big growth is embedded ODM, mainly for the smartphone and their PC. So we believe after 2020, Phison was able to have the other pickup in the revenue and profit. This chart is CQ2 breakdown. Consumer in the past were like 70%. Last year, like 30-plus percent, 30% to 40%. Now we lowered to 21% from 26% last quarter. Consumer -- even though consumer is not welcomed by a lot of investors, but we still need to keep the consumer business, try to move our more -- when the NAND is getting more, we need to use the consumer market to cut budgets. But when market is tight, we can move the consumer component, go to the industrial, gaming and embedded ODM. But on the other hand, controller business, we are growing 2% from the second quarter. But you can see the big improvement is embedded ODM. In early this year, I already announced embedded ODM in coming 2 years will be the big growth because we are getting more design wins in the PC and the smartphone business. And gaming is the other one, we are more getting the more investment into that. Phison most -- we cover most of the gaming console. And also, we are implementing different kind of gaming SSD modules for different applications. So this kind of combination to Phison, I believe, is really healthy. Okay. Our third quarter revenue you knew already is close to $ 17 billion, and the gross profit is $ 5.35 billion, which we hit the all-time high in the gross profit and also in the revenue. The reason we are able to keep the all-time high by CQ3. Again, just because of our different kind of product business model and the supply is too tight, we're able to get the full support from our supplier in both NAND and the 5G. Okay. This is some breakdown. ASIC is 22%, we increased 2% versus Q-o-Q. And the controllers growth Y-o-Y in the SSD 34%, but we need to highlight the PCIe Gen3 and Gen4. We keep almost 100% growth Y-o-Y. So next year, we already got customers design win focus. We still expect we're able to have a big Y-o-Y in next year. In module-wise, we have 71% of our total revenue, which including not only consumer, but also automotive, industrial, gaming and embedded ODM. And we believe next year by the total content, content means that the device capacity keeps increasing. We're able to keep driving our revenue by the module business. And the BGA SSD is kind of a niche products. We already successful built this into the smartphone, gaming smartphone. Now we also got the program from the Tier 1 PC OEM to put this into the notebook. And next year, they have a Chromebook opportunity. So Phison is going to play a really certain role in the BG SSD development. And by revenue wise, you can see that industrial Y-o-Y is a high percentage, Even though we see the unit in industrial, by unit, the growth is single digit, by unit. But by the number is almost double. The reason is industrial by capacity increased a lot, triggered by video and also the 5G applications. And gaming module, we keep trying to penetrate to more platform. Gaming console, handheld gaming device and also the gaming cloud. We have already got a program from gaming cloud SSD and start to ship by this month. And we announced by early this year, embedded ODM will be our main business, okay, mainly in the computing and the mobile applications. And we expect this growth will keep continuous for coming 2 years. So revenue-wise, we are hitting a record high. And gross margin rate is slightly declining. The reason is by CQ3, foundry increased the wafer price. But Phison, we keep long-term partnership with our semicon partners. We are not willing to 100% transfer the increased cost to them. The reason is controllers cost per module, the ratio is increased too much, which is not healthy. Phison willing to absorb a little bit of cost increase from the foundry, we are not willing to 100% transfer. This is to keep long-term partnership. And I don't mind to doing this because on the other way, our main controller customers are also our NAND suppliers. This is for mutual benefit. In order to help Phison to secure more supply with a much reasonable price. So we are happy with this gross margin, and we are -- keep 2 quarters above 30%. If by the earnings per share, the CQ3 '21 is second in the record high. But if we take out the 4Q of 2020, we have the asset size of our China subsidiary. So basically, in the business-wise, CQ3 this year, we hit the record high, all time high in the earnings. And income statement, we have many index hit all time high. I think just because of we have a different product portfolio, our business model, different applications and also we secure enough of NAND flash and the foundry supply. Even though the foundry, was still falling behind, okay? The fulfill rate is only like 60%. But this is helping Phison to hit the very good quarter. By balance sheet, I think, you are most concerned about inventory. Since early this year, we already talked to everybody, we are going to increase our inventory. The reason is -- for example, the foundry is still tight. We need to keep at least a 3 months run rate of controllers, okay? And flash memory itself, whenever we get the wafer from supplier, we need at least 8 to 10 months to make these 2 products, they can turn to the revenue. So the turnover rate is almost 10 months to 12 months overall. Without the inventory, we are not able to generate the revenue in business. So by the way, this is -- we keep talking to every supplier about next year's NAND flash LTE long-term supply. Unfortunately, we are asking more what the best commitment from them is freight. But next year, we make too many design wins. We need a lot of NAND flash. So at this moment, the NAND flash market looks like it is a little bit soft, so we try to accumulate the flash for next year use. By the way, we also already got a lot of forecasts from the smartphone, from the PC OEM, from the gaming and industrial. We need to secure the NAND flash for next year's CQ2. So our strategy is we need to increase the inventory. But I know I had to try to buy more, but I just worry I'm not able to buy as many as I wish. So we are doing very hard, trying to see if we can get opportunity, okay? And overall, the ROE in the CQ2, I think we are still doing good by 24%. And our goal is to hit between 20% to 25%. Overall, with our CQ3, the total financial statement. Personally, I feel we did a good job, and we have a very good result to the shareholders. October, we -- by monthly revenue, we just announced is close to $ 6 billion. It's the other all-time high by monthly. The reason by October, we hit the all-time high because they have a lot of consumer products pushed to the CQ4, a lot of transportation, the shipping issue okay? But on the other hand, we still get a very strong demand in the smartphone and the PC OEM and also controllers. I'm also able to expect by November, we're still doing quite good. Okay. So overall, the whole year till the October, we are already in the $ 51 billion. So this year, I think, overall, we're potentially are able to make the $2 billion in the revenue. Okay, business highlights. In retail side, we have the semicon using Phison controllers, our Gen4 controllers, build the modules for the retail business. And we also have semicon taking our external controllers to build their USB 3.2 SN SSD. And in the whole retail market, Gen4, if I count by brand, they have -- over 75% of brands are using Phison's controllers and solutions. At Phison, we also cover the high-end camera. Okay, all the -- most likely the storage card for high-end camera are covered by Phison, proprietary supplier. And in OEM, our fastest controllers, this controller [indiscernible] already put into many systems. We also got award from the Tier 1 PC OEM to put these controller SSD modules into the high-end gaming PC and also the workstation PC. So we are successful to have a big breakthrough to the Tier 1 PC OEM in both notebook and the workstations. And in eMMC UFS, we also get a few design wins in the China Mobile, the android phone. And our redriver, the new IC start to get the order and start to ship by coming January. So from design to shipping is 10 months. So I'm proud of my engineers. In gaming, actually Phison, we take a lot of advantage from gaming because we have early Gen4 development to support AMD ecosystem. So we are in a few console platform. Now we also, again, the award for cloud, gaming cloud SSD modules. In the data center, in legacy, data center SATA (2F). We get extra 2 million piece plus demand forecast by 2022, which is a bit surprise to me, okay? So from this, I can see the demand in the data center looks like keep growing. And the capacity itself also is a high capacity. So we are happy, even though we are the small player right now, but we can see rapid growth in the data center, especially our Colorado office already have a project start to make design in the qualifications. Industrial is a solid growth, but the content of industrial we can see is almost 2x growth per year. So we are happy we cover most of different industrial applications, and we are trying to put more into the automotive application. Okay. This 4Q, we can feel that gaming is one of winning mainstream. We got the award and start to ship the modules for the cloud SSD, which is first time from Phison. So a very big breakthrough. And coming years, we are going to try to design more enterprise gaming SSD for the cloud applications. And for handheld gaming platform and console, I think, we have a very good coverage. And by coming CES, we are going to announce our Gen5 SSD controllers and modules. And we are going to MP these modules by next year capital. If you [award it] by 2019, Phison is the only one Gen4 SSC supplier to support to AMD ecosystem by 2019 Computex time print. after 3 years, Phison is going to make the same story, but this time, it's moving to Gen5. So we are trying to build and shift the Gen5 to support AMD systems by coming Computex. So overall, we still suffer the foundry undersupply. We are trying to keep asking the supplier from them, but we have a very good extra support from our partners, foundry supply. So fulfillment rate, next year basically was like 60-plus percent now we're able to get to close to 70%. So we're able to get extra revenue and also the extra profit and extra market share. So this is a good sign. We keep asking more support from foundry. I wish we're able to get 100% fulfillment next year. Three, 4 years ago, Phison already keep emphasizing. Phison is going to put more resources into R&D. We make it. Now we can see we have a very good return. We have a different kind of application controllers, module systems, all are coming from the heavy investment in R&D. So in Taiwan, basically every company in the IT in the high-tech keep saying not able to get extra engineers. But you see Phison in 9 months, we are -- gained 1/3 of number. So we have a very good incentive program to get a lot of engineers. And I believe after the other 9 months, we're able to hit 2,500 which is our goal by 2025. We pull in 2.5 years. The reason is we keep hiring engineers just because we have so many projects, requests from semicon we are not able to fulfill. So the only way is to make the expansion. So R&D spend, as you can see by this 9 months is almost hit the last whole year. So overall, we can see in R&D expense, we also hit the all-time high. Phison's strategy, you can see we are saying that the so-called 5+5, left side is kind of technology from IP, system, ASIC integration design service. And left -- right side, you can see it's a different application segment. Each segment needs a lot of NAND flash. Phison, we are going to cover each one by ourselves. So imagine each one has a big demand growing in the storage, Phison will cover everything by ourselves. So in the coming years, we have confidence. We're able to enjoy the growth in this business. Okay, above is our CQ3 summarized. Now we collect some Q&A from investor. After I answer these, we directly go to the online Q&A. Okay. Last Friday, Taiwan Hsinchu District Court rendered judgment for my personal case. No Phison, okay. It is a case personal. So no impact to find Phison. Phison is nothing to do on this. This is all my personal issue. And myself still running my role, keep running Phison's operations, strategy, everything and Phison is still doing good. I'm still here. So don't worry. And talk to our -- my layer, we are going to appeal to get a fair judgment. So basically, we -- personally, I made some mistake. I don't know how to manage the past. But anyway, I take this, it's my responsibility. I will 100% cover by myself. Phison will be totally free about this lawsuit. So basically, Phison is doing good. I personally, eventually, I believe, I will be fine, then keep running Prison to hit our goal, NT$100 billion revenue as soon as we can. Phison, we issued a CB early in the August. We have an internal target price. The stock market recently, our stock market price is not doing good. We are not going to issue if the price is not going to our target. So potentially, this CB life cycle is 6 months. We are maybe going to expire this CV if the price is not able to hit our internal target. So don't worry, we are not going to issue the chip price based on CB. We insist to keep our internal decision to keep the price high. Yes. The freight is very tight. This year, we have LTA with one of our foundry, we have a prepayment to secure our few years supplies. On the other hand, our customers, we also signed -- we just signed an LTA with our key -- one of our key customers. The value is $100 million, and it's happened this week. We also have a few other customers, including semicon asking Phison to commit in the 3-year supply on controllers. So we are keep talking the other LTA with our other customers. So Phison not only signed LTA with supplier. Now we are going to sign LTA with our customers because the supply is already very tight and it looks like in the 1 to 2 years no solution. So Phison's long term, our goal, our goal is to grow our business, including every segment in the auto, industrial embedded, gaming and enterprise. So top line, we need to aggressively grow our top line. And eventually, we need to increase our bottom line, okay. To have a better ROI to our employees and the shareholders. So our guidance -- actually since last quarter, we increased our gross margin target from 25% to 27%. So I think we're able to make it. And in the net consumption, Phison consumed between 2% to 4%. We believe we're able to hit 6% to 8% if we bring the data center and enterprise SSD business into our portfolio. So we are aggressively to grow our top line. and eventually increase in bottom line. A few months ago, analysts keeps saying that memory, memory is weak, memory is in trouble. But I'd like to say memory, they have 2 kind of memory, DRAM and NAND. I have no study in the DRAM, no interest to the DRAM, but to NAND. Personally, if NAND price is dropping after 6 to 9 months, system content will increase almost double. This will cause the NAND going to tight again. So personally, I don't worry when NAND price going too weak. But NAND only in declining, only reason is price rests too much, okay? If price keep flat, keep slightly dropping, the NAND demand is always increasing. So this will be good to us in this industry. Then we try to consume more NAND, develop more different applications with our in-house IP controllers to get extra ASP. Phison, we are good to cover 2 car business. I mean the first one is called unit, unit means controllers. Don't expect the notebook can be from PC from 300 to 600 million immediately, no way. So if 300 pieces of PC equal to 300 million pieces of controllers, this is so called units. Controller is good in the gross margin rate. But in the ASP-wise, usually, if you take out the foundry shortage, ASP is slightly declining 20% in the year. But Phison, we also cover the content business. Many years ago, Phison, we are -- module, USB module with a 16 gigabyte is like a few dollars in ASP. But now with the SSD modules for gaming, we aim to the $400 per piece. Data center, we are aiming to the few thousand dollars per piece. So we're able to enjoy the content business with a higher ASP which can help Phison to increase our influence power in the NAND ecosystem. So we are happy with our business model. And I'm for sure only Phison business model able to grow the revenue in this NAND industry. So where's the price impact to the -- NAND price to the revenue and profit. If Phison still stuck in the consumer, yes, we need to worry. But now we are diversified to many different applications. So the freight price impact to the -- our business is getting less. Because, for example, for the gaming industry price is by kind of a quarter and half years. We're able to keep the price stable. So we are not much worried about the NAND market price anymore. And recently, I think, everybody keep talking about the metaverse. No matter what, I believe this is going to mature only after a few years -- 3 to 5 years, okay? But these kind of applications, they need a lot of storage, no matter in the edge or in the data center in a server. So this is good to us when this keep developing, Phison able to get more opportunity, develop the storage for either the edge or the data center and the cloud in the storage demand. Why Phison developed the Redriver and Retimer? Basically, Redriver to Phison is a small revenue. Retimer is a high ASP. Now we are developing by end of next year, we are able to get some business. The reason we are going to this business, our goal, Phison's goal is going to have a data center enterprise SSD business, okay, which is a high speed and high density. But on the other way, the motherboard, they also need to have a high-speed ASIC such as Redriver and Retimer. So Phison, we get into this. We like to have supply tower, full turnkey supply, full service supply to increase our influence power in these ecosystems. So this is helping us as a strategy products to bring Phison into the data center storage business. You may have heard from every design house that they are not able to get enough engineers. Yes, the answer is yes. But Phison, we working hard, at least we get 500-plus new engineers by this year, okay? Why? We keep talking to a lot of campus students. NAND is a big growth. NAND has no constraint, no ceiling. The human keep -- need a lot of storage. Phison is the only player after semicon able to provide the full turnkey. It looks like Phison is only player. We don't see any newcomer to provide the full turnkey solutions. And Phison, we look to the business, I think we are doing good, the profit doing good, bonus doing good. So overall, we got more attractive to get the more students to join Phison. But on the other hand, the costs in R&D, in OpEx, definitely will increase. So don't surprise year-by-year, the R&D cost is going to increase. But we have nothing to do on this. The only thing we can do is keep -- earn more profit to support these R&D OpEx. So I think Phison, we are in a very good decision-making. We get more people, develop more products, getting more profit then we invest into the R&D. And we make it -- you can see our -- this year our performance just because of 2, 3 years gross R&D investment. So we need to keep heavy investment now to support our coming year's revenue and profit. Okay. Talking about the renewable, the energy, Phison already start to adopt the renewable energy, starting with below 10% by 2021. The goal is a 2023 -- we already signed a contract. By 2030, we are going to hit 20%. But 2050, Phison is going to be 100%. But it depends on if supply of renewable energy in Taiwan is full enough, we're able to pull in potentially by before 2040 will be 100% to use the renewable energy to save our earth. Last is about ESG. Phison, we keep big improvement in the ESG, even the index from OTC and index from the MSCI. So we keep improving. And I believe sooner or later, we can get the other upgrade in the ESG. So above is today's update. Now we go to the online.

Okay. Yes. Thanks for the remarks. And we hope everything goes well with the case as well. So now we're in the Q&A. Just for this to raise your hand, if you go to your name on the right and hover over it, it will bring up the raise hand button. So yes, I'll start with a couple of questions and then we can go to the line. I think just 1 comment you made was about NAND. And if I heard it right, I think you're having trouble getting more than flat. You're trying to build up inventory because the makers are having -- having trouble giving you more than flat procurement for next year, if I heard that right. Is the supply that tight? Or are they prioritizing other applications? And because there's a market concern if the price is soft, there should be plenty of supply. So maybe what you're seeing on ability to procure NAND flash through next year.

Yes. We keep talking to suppliers to ask for a stable LTE next year. But unfortunately, we are not able to get enough number coming from them. And I believe they have prioritized the NAND to other key customers coming from next year mid-CQ2, okay? Usually, in the past, when time goes to the CQ3 we are always, in the last priority to get a NAND. I mean, there are a lot of big guys there, right? So what we can do is the reason I have confidence to build inventory because we've got so many design wins in different kind of applications, which we already can see we are going to build a lot of modules for the design win projects. But unfortunately, we talk to the supplier to make sure they're able to ship to us to commit, to ship to us at that period, but the answer is they're not committed. So what can we do is we need to take this period to make sure we're able to build out some inventory to -- also we design win into a lot of systems. We need to have at least a 9-month commitment to keep continued supply. So we have no choice, but we just try to build inventory.

Yes. Okay. And can I ask a follow-up on that. With the design win activity, would there be a -- do you think the mix will shift in a big direction, like your target long term is away from consumer. With next year like design wins, is it more embedded gaming, high ASP industrial? Sort of it will be more of a mix shift next year?

Yes. Actually, most design win is coming from the so-called embedded ODM, mainly it's mobile and computing applications. Then next is gaming. Gaming is going to add a lot of high-density drive. So these 2 actually is our main business in next year. Industrial, the growth by unit is small, but the content itself is almost double. So to me, these 3 segments is what we have to fulfill in next year.

Okay. And a follow-up on that with the very high ASP and data center, industrial, which you talked about, does that dilute your margin where it's a very high ASP? Or can you get a pretty close similar margins supplying into the very high module?

No, the answer is negative, okay? Usually, we talk to the high-density drive for data center, right? The first cost is like 99% of total bond. So I don't expect that we're able to get the same margin as a so-called gaming or embedded ODM, but we have to make it, okay? What we try to do is for the data center enterprise, we need to prove -- I mean, the first step, we need to prove we're able to make it. We're able to ship. We are able to get a design win. I don't care about that kind of market. But once prove we're able to make it. Then we can take a lot of our design service projects and customization projects from the customers, which can ensure we have a higher margin.

Okay. And one last and then I'll open to the line to see if any questions. That gross margin target that you took up -- you've actually been delivering above the 25% in the past few quarters. But for a mid- to long term, is it -- if you could go through the factor that gave you confidence to make that change.

Okay. Again, I used this target. Controllers. Controllers should contribute -- is a high gross margin rate and our controller portion by total revenue is increasing. And with thanks to our 2 foundry suppliers that give us good enough support to increase the wafer supply to us, okay? So this is helping our average gross margin. And going to industrial, even though industrial -- the total growth is small, but it's stable. Then embedded ODM, the margin is not as high as we expect, but we're also building some niche of modules to the mobile. For example, we built BGA SSD to the gaming phone, which is helping us to get extra better margin. So we need to keep investing a lot of R&D to make differentiation. We cannot just follow the standard spec, for example, Samsung go to the PC makers, the standard spec. We take that business. It's good in revenue, but the profit don't expect. So Phison had to fully use our internal strength, means R&D activity, flexibility to get the extra project with the differentiation, customization. If this keeps happening, Phison, we have confidence to keep the gross margin Yes.

Okay. And I'll see if questions, feel free to cut in, if I'm missing it if -- are there any questions on the line at this stage?

Okay. I have a question from chat room. What percent of NAND controller IC goes to an external customer or almost all NAND controllers are used for Phison own? Okay, of course, not. We have more than 80% controllers going to semicon instead of in-house use. So -- actually, in-house use is a less priority to be honest, our controller mainly to supply to the semicon partners. Okay. So this is answer to Simon. Any other questions online?

While we're waiting, I'll ask actually another on the margin. for the foundry, you talked about the price hike. You don't want to fully like to protect customer relationships not fully protect or not fully pass on. How do you see like into next year, there's been a discussion. There's a further raise going into early next year. I guess are you seeing up from your side like further price raise and how you feel about passing those on?

Good question. Basically I need to go to ask the foundry, ask the customers. But to be honest, I think today the wafer price to the foundry is happy, happy margin. If keep increasing, of course, able to get extra profit, but this sometimes, they kill some applications. Such as Phison, we get a wafer from supplier. I'm able -- because my fulfillment -- fulfill rate today is only below 60% today. Next year, I hopefully can go to over 70%. But that is still not enough, not enough -- of course, I can go in to raise the price, but I'm not willing to do that because business -- Phison prefer to run the long-term business partnership, okay? I don't like to see controller costs ratio in the total module increase that much. This will kill business in the future. So by the way, Phison, most likely our controller partners, customers, are also my suppliers, okay? So mutual benefit, mutual relationship is much important than there's a small margin to us. Yes.

Okay. That's great. Another question just on the outlook. Fourth quarter, we got the first preview with October. How do you see it holding up later? Like the -- because some of the builds have been deferred. Fourth quarter, even if you have enough visibility to even see first quarter, maybe it seems like a different pattern.

Okay. To be honest, in July, August time frame, personally, I'm pessimistic about CQ4 and CQ1, to be honest, okay? But October give me a surprise. Actually, November, I believe, also give me a surprise. Then, I figure out what happened. Or the reason is at the last 2 years, we did a lot of design-in program, which I don't expect is coming to real business but it's coming. So -- and on the other hand, you're talking about next CQ1, usually with CQ1 in the NAND market is terrible usually. But what I can share with you, we can see the focus is increasing with our new design win program. So next year, if I'm saying that I'm going to increase my inventory, of course, a supplier at this moment, market is soft, we expect them to take the flash. And I willing to take, by the way, and not only I ask, can you sell more to me? So they're also a little bit surprised. The reason is we can see, we have a very good visibility in the CQ1. This is what I can answer.

Okay. I mean I think your ticket is a share gain like they -- and it's more a smartphone or just for the client, it's a period of share gain in those markets.

I think eventually, yes, but we did a little bit slightly different with others. Phison, we don't like to go to a lot of mainstream business because mainstream, we are not able to compete with semicon, right? The first is coming for them. So at Phison, we are trying to use our strength, means R&D activity, differentiation, customization and flexibility feasibility to gain some extra project which potentially semicon may not interest. But to us, it's a big business and a much better gross margin rate. So we are willing to take that kind of business with the heavy design R&D activity. And this is a Phison's strength. Yes.

And maybe just one more on the outlook change where you originally thought NAND terrible. I mean you've gotten more market share design or I should say these like custom design wins, are you feeling different about the NAND market? Like better or still feel, let's say, tough overall NAND market and it's just your design is helping?

No, no, no. Okay. Phison is just a small guy, only 3% in the total NAND market, right? So CQ4, to be honest, we can feel that NAND is soft, is getting soft. Getting soft for many, many reasons, the component shortage, the smartphone cut down, PC looks already no any momentum to growth. But overall, to my business, again, we are just a peanut in this total temp. But luckily, we made a few design wins. This is helping us to grow ourselves. So that's why we are more willing when the supplier asks for help, we try to help. And this help we have confidence to fulfill and to use it. Yes.

Okay. And I think, Simon Woo had some questions. Let me turn it over and then we have one that came in on the chat, but we can start with Simon.

It is better to have a live discussion rather than email it. But Mr. Pua, can you hear me well?

Yes, I'm good.

Yes, yes, yes. Good. Number one, your competitor Silicon Motion were saying there is very strong, strong growth in the NAND controller area. I think one of the key reasons is NAND makers maybe from Korea or Japan, even the U.S., they cannot access well the foundry, sourcing right? Once they design the NAND controller chips. However, they have limited access to the foundry capacity. So do you agree with this kind of view? I mean the NAND makers cannot access well, maybe Taiwan, the foundry?

Disagree. I disagree.

In that case, do you think it's a temporary -- for example, it's kind of a structural change... [Technical Difficulty]

Simon, your voice quality is terrible. I cannot catch it. The voice quality, I don't know what happened there? Hello? No, no.

Good?

Now okay. Please try. Now okay.

Yes, my question is... [Technical Difficulty]

Yes, yes. Please continue. No, Simon, sorry again. Again I can't hear you. So what -- something wrong there? Hello, Simon?

Yes, I think I'm using the office phone that should have been okay.

Okay. Now it's good. Now it's good. Please. Now it's good.

Yes, the question is this trend is a structural change. I mean, the NAND makers need to rely on the Phison, Silicon Motion, the [indiscernible] companies for their NAND controller sourcing.

Okay. I can really tell you what is the root cause, okay. NAND is getting weak and complicated, so need more engineers, formal engineers to cover the weakness of NAND. And semicon itself, they may not be able to get good enough of R&D formal engineers to cover every product line. They rather prefer to use their internal R&D, from their resources, go to the high-end, high ASP so-called data center SSD. So for client SSD for the -- defined as a low-end SSD, they may prefer to outsource to third party. I think this is the only root cause. But talking about the foundry capacity constraint, again, disagree. Disagree. Talking about the size, right, the revenue size, Phison, Silicon Motion, just a peanut compared to semicon. So -- yes.

So you are saying the -- like Japan and Korean memory chip makers, they can offset the foundry capacity in Taiwan easily going forward or not?

No, no, of course not. Basically, in these -- since the last CQ4, CQ1, everybody struggled, everybody, not only semicon, I mean, everybody, okay? Even though today, the supplies are not enough. So everybody's struggles, means Phison struggles, Silicon Motion also struggles. But if talking about the bargain power, semicon, NAND supplier has a much big muscle to talk to the foundry rather than Phison or Silicon Motion. So I don't agree to this statement. Yes.

Yes, yes. And then lastly, I know someone already raising the question, but you are saying more than 50%. 5-0, 50%, you NAND controllers is now shipping to the NAND chip makers.

No, I'm talking about 80%.

8-0.

Yes, 8-0.

So you are using only the 20% of your NAND controllers design to fabricate it for the internal use.

Yes. Yes, you are right. Yes.

Only 20%?

Yes. So the other question is talking about Gen5. Yes, Phison, we are going to announce our Gen5 controller, the Silicon by next CES January and trying to make the modules MP by next Computex. Gen5, basically, I think, will be good next year in the data center enterprise, but for PC, I believe only the high-end -- super high-end gaming PC going to adapt Gen5. [indiscernible] until 2024 to me, I still put a question mark because the [indiscernible] itself is still trying to find a solution. So I believe '22, '23, the mainstream is still a PC Gen4. Okay. Yes.

Okay. If we have other questions. For the Gen 4, you've been very strong on AMD platform. Is there much of an Intel player or they'll just skip it and move to Gen5. Some of your business still off of gaming, AMD for Gen4?

Yes, we are the first -- what we can say, we have the first ZS50 with AMD. Of course, after Intel launched their Gen4, Phison, we are ready, okay? We are ready to go. But talking about the Gen5, the question is, Gen4 Phison already made the drive itself up to 7 gigabyte per second read and write. I asked a question to the PC makers. Actually 2 days ago, we have a dinner -- to have a party because we made the big shipment achievement to them, they're happy, so they buy me a dinner and drink. What they're say is who is going to use a Gen5 for 10 gigabit in read and write? I say, can win Windows 10, Windows 11 show the difference of PCIe Gen4 and Gen5, drive? The answer is no. So the only different able to show is in the data center. So PC, I don't see is a mask to use the Gen5 unless it's a high-end -- super high-end gaming PC.

Yes. Okay. And for your data -- I don't see a question, I'll ask another follow-up. For your data center SSD, I mean, you mentioned the bigger memory merchant -- the bigger memory vendors will focus firmware and development enterprise. I guess your market entry strategy and how far away the development for that. And it sounds like for Gen5, that might be the biggest driver or the biggest application. Is that where you're pushing to come in? And if there's kind of an updated time line on the R&D, how that's progressing?

Yes. Yes. But, okay. Yes, what you're talking about is what I'm doing today. But the most important thing today is Phison, we need to show we are able -- capable to build the drive for data center enterprise. Like 3 years ago, we tried to convince PC makers, Phison, we are good for CSSD, for PC OEM, but the answer from Tier 1 is negative because we don't have a trade record. Thanks to AMD, bring in -- bring Phison into the ecosystem. We make it. It's a miracle, a miracle from nothing to MP in 9 months. We get the full coverage as for everyone. So now you see we are good in the business. Next year, we have to prove our data center Gen4 SSD able to ship, able to get design win. Then the next Gen5, I believe we have a lot of opportunities there, okay? So this is my strategy.

Okay. And Simon has a follow-up question on the component shortage impact. I think when you -- okay, when you talked about the NAND market getting softer, and I have one follow-up, so ask both is -- so the softness in NAND, do you think it's mostly a supply like a mismatch? Or is there a demand issue? And then Simon's question is, is that an impact for the module business itself for us, like an OEM customer. I mean if you're there, you can clarify.

Yes. Actually, personally, I believe the component mismatch is also causing the system shipment, okay? This is a first -- I mean main reason. But second reason is look to the main population country, main populations. At the last 6 months, they suffered under COVID. So the income from those countries is a question. So those big population countries not able to consume the high-priced products. And by the way, the price itself was raised 20%, 30%. So this is -- they're the 2 reasons. But to Phison, I don't see the component shortage impact to Phison. Because Phison, we are not like semicon is the main supplier. We are just niche customization, differentiation products and we have the flexibility. So look to Phison's number, I think we are good.

Here is a question. The soft NAND, you mentioned about -- it's a little bit consumption mismatch. I mean everyone is 40% under fulfilled on logic. But do you -- I mean, I guess just looking at the industry, do you think there's an element -- like a potential for that to reverse quicker or you do see the structural driver for that sustaining. Like just we go from everyone trying to get more to if there is softness in the market that is reversing. And then all of a sudden, you don't have the supply issue anymore. Do you see much potential for that scenario?

Okay. To be honest, I need to make the way -- to answer, I need to go to as a fortune teller, right. But, yes. But okay, what -- I have confidence because I talk to most of the semicon supply, as Samsung here. We have no business with Samsung. But the rest, we talk to them, it looks like CQ1, they are able to supply what I'm asking, but start from CQ2? No, the answer is no. So you asked me what happened? Seriously, I don't know what's really going on there, but for sure, the demand and the capacity by unit, I don't believe it's growing, but by content itself, I can see strong growth because triggered by the 5G applications. So that's why if I'm not able to build inventory at this period, I worry after CQ2, I have to suffer a lot of business opportunity.

Okay. We have a few in the chat room. The -- I guess if we hit Simon's question, he had asked about the -- it sounds like it's impacting your margin. I just wanted to clarify if we caught that, it's impacting your module business but also OEM customers, the component...

No. The answer is not because we are not the mainstream supplier. We are not the mainstream so. And Simon asking about already asking about 4Q, the sales should be declining Q-o-Q. Good questions, but October looks I hit the all-time high, right? October. November, based on the systems, I can see we are doing good. We are doing good. So even though 4Q declining, but I believe it's -- potentially, it's a single-digit, potential. Yes.

Okay. Would the margin potentially be down a bit as well?

Yes, the answer is yes, yes. But I believe it's not a big impact, not big impact. Yes.

Okay. And then a question, it was an idea that on the second court ruling date if it's coming sometime -- if you kind of a time frame like whether it's second half 2022 or any estimate in court ruling.

Potentially, in next year or early '23. Yes.

Okay. A question, Simon also asked if you could give an update on the controller mix between the different nodes like 28 plus versus 12 slash 14-nanometer 16, like how your mix is and how that's migrating?

Now our main shipment is base covered by 20 and 12, all the new design today is 12 next year is moving to 7.

Okay. For the -- I guess for your volume production?

Today, 20 and 12.

Are both nodes equally tight or do you have better supply you're seeing?

12 by demand is big, but unfortunately, supply is still not enough to me. So if supply is enough, basically, it's equal. 20 and the 12 are equal.

Okay. But 12 just not enough supply?

Yes. Yes.

So any -- if there's other questions. What time -- we're up by 4:30, just how long do you want to extend the call because we can...

Actually, I have the other meeting appointment. Yes.

Okay, I'll just -- one last time any follow-up or final questions. Otherwise, we can wrap now.

Yes. Okay.

All right. thank you. Appreciate KS, also quoting for helping on the Q&A. And feel free to reach out to Phison as well for any questions or comments, and they'll post the presentation to their website.

Yes. Thank you. Thank you, Randy, and thank you to everybody to attend our CQ3, the call, okay? Have a good day. Thank you.