Good afternoon, everyone. Welcome to Phison's second quarter earnings call. This is Jason Tsang from CLSA Taiwan. Today, it is our honor to have Pan Jiancheng, the CEO at Phison, here with us. Mr. Pan will share the recent update and outlook for both NAND flash industry and Phison company itself. After that, we will, we will open the floor for Q&A section. Now, let me turn this call over to Mr. Pan. Thank you, Mr. Pan.
Yeah. Thanks to Jason, again, thanks to ladies and gentlemen, to participate Phison's our CQ2 earnings call. Next, we go to today's the file we are presenting. Okay, again, we, we keep repeating to talk to the investor to define what's so-called the memory. Phison, we are in the storage business. We are only cover the NAND flash industry. We are not in the NOR, we are not in the DRAM. This is what we call it till yesterday. I want to share with you, we, we, we share this in the our CQ1 earnings call. By CQ1, the total loss in this five memory company, exclude YMTC, is NT 10 billion. CQ2, Kioxia not announced yet, CQ2 accumulated loss is by NT 8 billion.
YMTC, what the rumor saying, they keep still in the big minus. This just want to show you, if the NAND price is not bouncing back, memory company, they are not able to make money because the NAND cost is never come down unless they move to the next generation. Move to the next generation, they need to spend big CapEx to get the next generation flash ready. Unfortunately, right now, everybody cut CapEx due to the poor income. CQ2 keep losing, I believe, accumulated maybe like $9 billion. CQ3, if price not bouncing back significantly, CQ3 is still around $7+ billion. This is not healthy, but the market is not turning back. The NAND company, memory company, will still suffer.
We believe, we believe, since the flash keep getting cheap, cheaper, we also observe the contemporary device actually have a significant increasing. To Phison-wise, what I'd like to say is, if memory company keep losing money, actually, Phison, we are buying much cheaper material from them. We don't see this is bad to us, but we can take much lower component price to make to the products, to the market. Of course, we wish the market back to the healthy NAND company able to make money. This is only way, is the NAND price have to bounce back. Okay. Supply side, I believe right now, according to many news, Korean, U.S., also Japan, they keep cutting their production in NAND.
Hopefully, Q3, the production reducing, and Q4, demand coming back, the market will return to normal. This is Phison's our breakdown. I need to highlight here, Phison, we are putting the enterprise module revenue into our pipe. We in CQ2 around 4%, and back to this record, actually, we have a certain 2%, 3% at the past by quarter. Recently, we start to get some significant order. The revenue in the portion is increasing. By 2024, you know, we are doing a lot of design-in activity. We are accumulate the, the, the products, the component. We are doing a lot of design-in, so we believe, again, when market back to normal, we are able to increase our revenue in the enterprise modules.
In controller side, I believe this is in the 4 several quarters, is the first time lower than 30%. The reason is, actually, since last year, CQ4, the NAND company... Actually, Phison controller, we are only selling to NAND company and to Kingston. The NAND company, they are digesting their inventory. Actually, they cut down a lot of forecast and order. This lower demand in controllers happened by last year, CQ4, CQ1, CQ2, reflect to our revenue. This, the number is first time below 30% of our total revenue. Since May, from May, the NAND company start to ask the new forecast, and we are start to order our new wafer into our foundry, both for TSMC and UMC.
We believe the forecast demand will back as normally every quarter. We've been suffer for around 2-3 quarters, a low demand in controllers, but now looks like the things back to normal. Not normal as of 2021, but at least the new demand coming in with a forecast, we start to release our wafer to the production. I believe this will help our gross margin in the coming futures. The revenue we already published to the market. The revenue by CQ2 is NT 10 billion, and the gross profit is NT 3.2 billion. Gross profit down YoY, but up around 1.0%-1.5% QoQ.
Revenue trend looks like the number is similar as CQ1 due to the price dropping too much. Control demand actually very weak, the overall demand, even though by content itself increasing, but ASP is still very, very weak. Anyway, it's a two quarter we keep in the NT 10 billion revenue. We need to improve our revenue, but this we have to wait until the demand back to the normal. U.S. market looks like gradually getting better. China looks like not yet, we need to see October, September, if China can be a little bit stronger than the last two quarters. Gross margins by our CQ2, actually, little bit surprised. The gross margin rate is 32.48%, which is a history high.
Okay, it's all-time high in our company's record. Definitely you are going to ask why, right? Later, I will explain why. We in CQ2, we also write down inventory around impact our - 0.5% of gross margin. Means, Phison's overall, the module, the business from our controllers, module design service, we are getting much, much healthy. The overall 32.4% record high, and we are happy about that. We hope we are able to improve our gross margin rate in the futures. Right now, what we are trying to do is to increase our revenue. Okay, the earnings per share. CQ2 on the book is NT 2.28 EPS per share. We still suffer the investment loss from our China investment subsidiary.
If without this loss, NT 2, we are keeping 4.2 EPS, overall growth a little bit. On the book, EPS went to the 2.2, is better than the last quarter. Of course, it's still not enough to us, but gradually, we try to back to normal. This, again, we suffer the Hongxin loss for already 3 quarters, so it's around close to NT 10. So it's a NT five someone capital. Anyway, their loss in the GAAP is reducing from NT 3.9, now loss NT 2.7, $2. I hope CQ3, they are able to get even, but they are still fighting very hard. At least, the GAAP is getting smaller. Income statement.
I think you are concerned more in the gross margin rate, I think this all-time high is good to us. The OP margin, OP margin is 7.2% because we the overhead in the engineering and management, we increased it a bit, okay? The net profit improved from 2.4%-4.4%. This is a first half income statement. Overall, the gross margin, 32.1%, I still remember this is also all-time high in the first half of the year. We keep working hard, try to increase our gross margin and our revenue. Balance sheet. The cash flow, until today, Phison still running our business with our own cash.
We still have a NT 13 billion cash on hand, which is much healthier than any other module house. Looks like we are still actually capable to buying the wafer, NAND wafer, and the semiconductor wafer. Inventory-wise, actually, we keep, like, NT 20 billion. We have write off, like, 10% of total our inventory. If the price still keep weak in CQ3, we are going to increase our inventory. The reason is the price is so cheap, and NAND supplier selling NAND to us, they are suffer - 20%, 30%, 40%, 50% gross margin. Why not just buy it? With our enough cash, right, we hope to increase our petabyte inventory on hand.
Overall, the price per gigabyte is much, much lower, which is helping us to get a more bounce back when the market turn to the normal. Overall company, still doing healthy. We still have enough cash on hand. This is also some supplemental information about T-IFRS and Non-T-IFRS, our analysis to you. By CQ2, we have the share-based payment, around NT 66 million. This impact to margin is quite small. Anyway, it's like 0.6%. By the net income, I think the impact is more. Again, we, Phison, we consider in the coming future, we are going to exercise some more options, stock options, and some treasury share.
If this happen, I believe the OpEx in this may increase. We like to let shareholder to know how we use it. Okay. By July, revenue is NT 3.3 billion, MOM is down 1%. July still is a summer holiday in Europe and in the U.S., so China is still weak. Even though we try to push more business, more revenue, but unfortunately, we're still not able to improve a lot. We hope by August, September and October is a traditional season, back to school and also Christmas. We hope we are able to gain more revenue there. For business highlights, since COMPUTEX, Phison announced our first E26 controllers to the market, and looks like now the market, they have a Samsung.
Other than Samsung, looks like Phison. We enable many our partner here, and these products, the margin is still quite comfortable, and the demand looks like for gamer, the demand is increasing. We are happy to bring our partner to this market together. By this Computex event, we also have a partner to get us award. Our partner AnandTech , to get us some award in this event. We are happy to see that. Yeah, this is by Patriot, okay? The good review from Tom's Hardware. Okay, by Q2, Phison also announced we got a award certificate of automotive ASPICE certificate Level 3. Looks like Phison is the first independent house to get this certificate in the automotive.
Actually, including the name company, I think we may be still the only one to get a CL3. Okay? We invest this for 2.5 years R&D to develop and to study. Finally, we're happy we get this. Recently, the automotive market has some changing. Okay? Phison, since we got this CL3 announced to the market, we got so many inquiry from the U.S., from Europe, Japan, and also China, to ask Phison to be one of partner in their EV and the traditional auto storage partner. We are busy on that. We are working hard. I believe this, hopefully by 2024, Phison revenue breakdown, we are going to put the automotive storage into our breakdown. Now, this quarter, we put enterprise. Next year, we are going to put automotive.
Some business highlight by CQ2, we put our high-end Gen4 into the notebook OEM. We start to ship the products to the some gaming notebook. In server wise, Phison's cooperate with the Seagate, our dual port single port products are already in the market, and we get more than 35, 40 design win AVL in the system. This will helping us to gain some more revenue in the coming futures. Okay. Automotive, we have many, many projects to work with our partner, Europe, U.S., Japan, and China. Okay, this will deploy our revenue by 2024. For industrial demand, we made a lot of customized design into our partner. Okay.
IoT, there's a lot of IoT special for edge computing, and recently, the AI application, we put our storage into the many single port, CPU port with our SSD. Phison, since three, four years ago, we announced we are going to have a design service business, and every quarter, we have some income through the NRE design service. Our customer mainly is in the PCIe 4, PCIe 5, UFS, eMMC, and also we are working with the phone makers to customize the storage controller and firmware. More customization happen in the automotive. Okay? We have many, many project ongoing with the automotive players. Looking to the recent, the industrial semiconductor industry, what happened to that?
We believe the NAND maker, you know, they are trying all their best to lower their production. Since the demand still not bounce back that much yet, only way to push the price up is cutting production. 2 Korean and 2 U.S., 1 Japanese company, they already announced they are going to cut against certain percentage on the NAND output. The output after 3 months, means by end of CQ3, the output is going to decrease again. Every NAND maker also keep announce to their customers, module customers, or wafer customers, saying that they are going to raise the price. Okay, there was some news in the media saying that Korean company raise how much, or this raise-- but they are trying to raise the price by themselves. Okay?
Their CapEx in the next generation NAND looks like most likely pushed out again. Japan, U.S., and some Korea, they pushed out their CapEx in the next generation NAND. Means in the coming years, one or two years, their NAND output in petabyte may not be increased that much. Okay? Overall, we also observe the U.S. market is slightly back. China, again, still very weak. We need to see what happen August, September, since their government start to inject some policy into their market. What Phison trying to do is, we agree right now, supply side, we already clear on supply side. They want to raise price, they want to cut production. Demand side is a little bit complicated. We believe, again, the NAND market, the demand storage will come back again someday.
We keep invest into our R&D. We believe with our investment in R&D, we are little sure, since four years ago, with a heavy continuous investment, their gross margin. Even though the market is so bad right now, gross margin, we are happy about that. We hope when market turn to good, we are able to get more returns from our investment. Talking about the investment, little bit, you know, heavy by in the second half, is 20% of revenue, went to the R&D investment. Of course, the revenue dropping, but the commitment is here. We are not laying off. We are keep hiring the new headcount. We are keep investing the new project.
We migrate to several 7-nanometer projects, so this is a heavy expense, but this is to prepare for future. We hope our shareholders, you know, to agree what our decision. At least today, we show our gross margin all-time high due to last four years heavy investment. Few weeks ago, Phison announced our activity in the AI. Of course, we are not trying to tell the market we are AI-related company, but for many years, our controllers, we implement many AI algorithm to improve our controller error correction. We have, like, 60-plus AI engineers in the lab, doing all the design for our controllers. Okay? By the way, Phison, we are outsource our module to the factory. In many years, we also suffer the using the AOI.
They have a lot of, fail, okay, fail, investigation. My engineers, AI engineers, they jump into the lab, they try to use the algorithm to train the AOI machine. The fail rate from 40%, now we improve to fail rate 1%. Means the pass rate from 60% now improve to 99%. This actually helping our outsource factory reduce a lot of, human headcount, and we prove our algorithm on AI is good to use. We also propose this to several SMT company in Taiwan, which is a big size. They have 20 SMT line in their factory. They are using Phison's, our AOI solutions, the AI AOI solutions. We are going to turn this to a new business.
As a service business, we are going to provide a total solution, including our SSD, our, the, server with our enterprise SSD, then we provide algorithm and service. Hopefully, this can be. You know, this is a service fee, so the gross margin is 100%. We hope after this, 10 +, our partner, they are satisfied, comfortable, we are able to push this solution to worldwide, okay? To improve Phison's, our revenue and our gross profit in our coming futures. Anyway, this is a proven solution, and we're happy to see this. Hopefully next year, if we get some significant number return, then we are going to tell you more in this number.
We also figure out this not only good in the AOI, but also in the AXI, which is, AOI is optical, AXI is X-ray. Now we are implement from AOI to AXI. It's the same algorithm, same application. If we can make the AOI good, we believe AXI and others, application, we are able to gain some share in the market and helping Phison to gain more profit from this business. Okay, above is our presentation about CQ2, some Q&A. Let me finish, we'll go to the live Q&A. Okay, Today, we have our board meetings about our first half dividend. The EPS by first half is around NT three something, we decide to have a NT 4.5 cash dividend to the shareholders. Okay?
The reason is, actually, the loss on the investment didn't impact our cash flow. It's just on the book, the China investment loss. If we don't count the China investment loss, I think we are capable to pay 4.5 as a dividend. We did our best, and hopefully, we are able to make company much profitable, and we are able to pay more to shareholders. In today's board meeting also, they have a resignation from our GM, Ouyang, and due to his personal career planning. He decided to step down from the general manager. Board members appoint him still as our chief of staff. Okay? Still, providing the strategy and advice, and consult of the company development.
The board members appoint CS Ma, our R&D VP, as Phison's general manager. Phison company policy, the general manager must be engineers, because Phison is an engineering company. With the new president, GM, CS Ma, I believe we are still capable to improve our R&D activity and efficiency. Asking about, since keep losing money, is that the demand will bounce back? I think demand is nothing to do with the supply side. Demand is about more in economy. U.S. looks like gradually turning to better. China, we still need to wait and monitor. Overall, the demand content per device still increasing. Still increasing. The only problem is the price is not bounce back that much yet.
We need to wait until the new signal, if the demand start getting strong with the production cut, I believe the unbalance is going to happen. Okay? The market looks like still consume the inventory, okay, store in the NAND supply side. When inventory went down, inventory going down, we believe the market definitely will deflate. If the NAND price is not bouncing back, since the NAND cost is not going to cost down, means the loss every quarter still will be the same number. Say, $10 billion, $8 billion, $7 billion. I don't know how many quarter memory company able to suffer. By this July, June and July, we already read that from news, several Samsung saying cutting, SK cutting, Kioxia delay, Micron cutting. I believe with this cutting, right, the impact will happen after three months.
Means by end of Q3, the output will decrease again. Okay, this asks about the flash, its potential to go to tight again. We back to the history. The flash, a big drop by 2008. After 2008, big bounce back. Okay. 2010, the other drop, 2011, bounce back. 2018, 2019, big drop. 2021, bounce back. Since there were almost maybe like 18 months, the prices are dropping. Next bounce back definitely will be much, much stronger than the past. How many months market drop means how many months the market will bounce back.
Anyway, what we are trying to do is keep ourselves ready, keep all the product line ready, keep the future investment ready, so when the times happen, hopefully we are going to get more ROI from the market. We also been asked by the, how the margin come to the 32.48%. I think there are a few reasons. Phison, since for many years, we have a lot of customized NAND module, including for auto, industrial, security, encryption, any kind of firmware development for every kind of applications. We build a lot of this design in business. Right now, the material cost is going so cheap, but the system price still able to maintain. That's why the gross margin, we are able to increase. Okay? What we need to do also to improve our revenue.
Phison also, we have engaged many our partners to doing the design service. Every quarter we have some good amount of revenue coming, okay, from our partners. This also improve our gross margin rate. I also believe with our heavy investment in the R&D, and also the NAND controller company is very challenging in the investment and also the gross margin. If we are able to hold the development, I believe we are able to gain a share in the business. Means we are able to improve our gross margin rate again in the coming futures. Same question about Phison gross margin. Phison, since we transformed ourselves after 2019, engaged with AMD CPU, after that, we are able to manage our gross margin higher, 27%. Okay?
Even though the market, the market, the, the, the gap in supply, demand still is oversupply. Even though oversupply, I don't see the material price dropping anymore, okay? The NAND company, they want to increase the selling price. I think by gross margin-wise, we have the confidence to keep, you know, much higher than 27%. We hope market return to normal. Inventory, we still have an inventory write-off of 10%. When market bounce back, this 10% able to write down, write back to our book, and this is a significant amount of EPS. Hopefully, this coming soon. During the AMD visit to Taiwan, we have some event in the Chiao Tung University, and we've been asked what type of AI Phison participate.
We are doing some POC with a several server company. This is all about the training model. A large language model is doing POC. We believe within four to six weeks, we are able to finish the POC, then we're able to see the efficiency and the performance of our SSD into the AI. Once complete, then we are going to make the announcement about this program. We still keep filing more patents on this architecture, and we are working very closely with the server makers, AI server makers, and we believe once this POC proof is good to use, we are able to gain the more AI server company to engage with us. Let's see. By end of Q3, we are going to announce more about this program.
Talking about our inventory, we still write off in the Q2, and we forecast Q3, potentially we will still may write off some because we keep buying the lower pricing component from suppliers. Since the suppliers calling Phison to buy more, so we may able to benefit little bit lower price. We need to write off the inventory. Overall, I, I don't believe the NAND price will have a big drop anymore. This is about China investments. Unfortunately, we keep losing for 3 quarters. I hope Q3, they are able to keep even or just a little bit minus in the finance. We still helping them to try to gain more share in the Ch-China market, and we try to make them more healthy in the finance.
Hopefully next year, when market turn to good, they're able to help Phison to make more money from the market. Look to the Since Longsys and BIWIN, they are public, and they announced some loss. Both company, CQ1 lost and the CQ2 lost more. I mean, it is what it is. No comment. If talk to the look to the NAND related market, Phison is one of several companies still able to keep profit, and profits still growing. This, again, this is all about other investment. Back to the ESG, we gather some new ratios, ESG rating. Overall, it's improving, okay? Every segment we are improved by 2002, 2022.
We still investing, we're still working hard to improve our ESG score. Okay? This is all about today's Q&A. We can back to online.
Thank you, Mr. Pan. We now can start our Q&A section. For online participants, if you want to ask questions, please use raise your hand functions, and I will unmute you to start question. Thank you. We now have a question from Tom Hsu, Tom Hsu at KGI. Please go ahead.
Hello, thank you for taking my question. Hello, thank you for taking my question. You have mentioned that some NAND flash suppliers has delayed their process migration into 200 layers and above. So I'm curious if Phison would also have to postpone our product development roadmap with the delay of migration of our NAND flash suppliers. Thank you.
No. Okay, let me clear back again, okay? Actually, development, development of a next NAND flash memory delayed it a bit, not delayed that much, okay? Development delayed, but keep going. The question is, once development ready, going to mass production, they need a big, big CapEx to buy equipment, they don't have money. In this case, right, actually, we follow... Now, they have a 5 plus 1 NAND flash maker. We follow the 5 makers. If they push out the CapEx, then I no need to invest on the resource today for that flash development because, you know, time to market, right? We are able to keep flexible inside Phison. We able to take the more project for recent NAND flash, just to improve our revenue and profit. We are making every judgment, almost every week, to follow the change of the CapEx.
Okay, that's my question. Thank you.
Yeah.
Thank you, Tom. We have a question from Donnie Teng. Donnie Teng, please go ahead.
Hey, Chairman, thank you so much for taking my question. My first question is regarding to the, the progress of r edriver and retimer, and wondering if you could kindly give us some updates on these two business and the potential opportunities, maybe in the AI applications in 2024. Secondly, is that, you know, there are two important mergers related to our business outlook. First one is potentially like Kioxia and Western Digital, and otherwise, MaxLinear and Silicon Motion . Because there are some uncertainties and lots of news coming out recently. Just wondering if you could kindly provide some insight to us. Thank you so much.
Okay. Talking about the Gen 5 driver in the motherboard, I believe Phison is the biggest. We gain the biggest share of the motherboard in Gen 5 gaming. Unfortunately, the total amount is not that big. It's a few million pieces a year. Gen 5 retimer, right now in the market, they have only one supplier, which the supplier service team mainly located in U.S. and India. Retimer customers, 90% in Taiwan, create a lot of service problem. Phison, we deliver our Gen 5 retimer samples to those Taiwan, Taiwanese server makers by May, June and July for testing. We deliver the samples. Most of them are very happy about our performance. Also they are much, much happy about our service because we are on the same island, same language.
They are pushing Phison to turn our samples to mass production. Our production version will be ready by CQ1 next year. Okay, hopefully when the CQ1 sample is ready after quarter, then we are capable to ramp up by CQ2, then second half next year, we are able to take the share and convert to the revenue. We are very optimistic about our retimer business. Okay? Talking about the M&A, KIC and WD, I've been asked more than 1,000 times, can they merge? Can they marry? To me, I'm very simple. Every M&A, you need to get the antitrust approved, mainly from U.S., Europe, China, Japan. Okay, these few big economy territory. Based on today's observation, I think it's hard to believe China is going to approve KIC and WD merger.
The reason is simple: China don't have any benefit from a Kioxia, WD. Since from the news, China, I mean, the U.S. and Japan still strengthened the equipment shipment to China. I don't think this is able to convince China to approve U.S. company and Japan company M&A. Again, I, I don't think this is going to happen, even though maybe they are going to engage, but the wedding party will never happen. Okay. Second, SMI, MaxLinear, I think is hard to comment. I mean, this is a very dramatic announce. A day, after a few hours, then they say, "No," I mean, a lot of our shareholders suffer. To my opinion, I don't know what's going to happen, and I'm hard to comment, but I believe the M&A potentially will carry, I believe.
This is my personal comment, so don't ask me, okay? I believe at CQ1, by March, in the call, somebody asked me about, oh, how about China? The SAMR is going to approve the antitrust. I remember that time, I said yes, because I visit to Beijing, February. Many investors call me to ask comment, but no, this is no comment, okay? I believe they are ongoing, but I don't know how to comment anyway. Overall, this is nothing impact to Phison. Phison, we still keep investing our, the activity. We're still getting the-- our, you know, customers, confidence. Imagine one thing, if you are telling the customer you are going to sell company, you fail, then you tell company-- tell your customer you are going to sell, are they going to believe you or not?
I think this is how customer trust to the company. I think they have to work hard to gain back the trust. Okay.
Okay. Thank you, Chairman.
Thank you, Donnie. We have several questions from online, online chat. The first one is, how do we look at the Chinese market competitions?
Okay, let me back to the question here, the 1st question, Maxio, Maxio and the InnoGrit.
Yeah.
I think for many years, right, we keep telling our investor, Phison, only selling controller to NAND company and Kingston. We never sell any controllers to the central module house, we are not willing to sell because bloody market, okay? They use a lot of inventory, they don't care about the quality, they care about only, like, fast food. We don't have any business there. Since the Maxio tech InnoGrit, YEESTOR going aggressive, the SATA controller went down to $1.10, $1.20, gross margin, only like 10%-15%. To be honest, I'm not interested, okay? I never take that. This kill Silicon Motion a lot. From a financial point, from a Silicon Motion, China is their biggest market. Their biggest market is coming from Central Module House.
This is hurting Silicon Motion controller shipment. This is a signal to them, not to us, okay? For China market, when China company coming to the market with a gross margin, 10%, 20%, 50%, what they're trying to get is subsidized from their government. They try to get money from capital market. They don't care the PNL. This business is not healthy. Why I'm spending my resources to take their business? Other than that, Phison trying to sell more modules to the China, customized module to the China market, because we believe with our good quality of controllers, we have a much better cost structure of a NAND flash, we still able to gain a share from China.
On the other hand, for those of China controller business, low-end market business, we let Hongxin to take their business, okay. If there's a broady market, then that Barry territory to participate. This question is better need to ask to the Silicon Motion, not up to Phison, because nothing to do with Phison. Second question, enterprise revenue. Revenue-wise, it's increasing, and 2024, we believe the demand in the enterprise SSD by gigabyte definitely improve a lot, okay. Unfortunately, right now, every enterprise SSD we caught, Samsung always can be lower than Phison because they want to move their inventory. I don't see now is a good incentive to heavy promoting, selling the enterprise SSD. I'm just trying to get more design in to the customer side.
I believe, again, NAND price, NAND costs never went down, never go down, unless new generation. Before new generation, you need CapEx. Before CapEx, you need make money. They are cannot make money. If NAND price cannot go down, to make money, only way is to raise the price. We believe if the price will back to normal, NAND company is going to make margin. When that happened, when NAND company raise the price, Phison costs still is our cost. Still, inventory cost still very, very low, so we are able to gain more share and profit when that happened, and that will turn to revenue. 2024, I'm very optimistic in our enterprise SSD revenue, but I'm not able to tell you how, how many percent is going to improve.
The reason is, when market go to the tight, go to the under supply, we need to take all our inventory to support those design in business. Means industrial, embedded, enterprise SSD will increase a lot in revenue, and retail business will go down. Okay? Again, conclusion is I cannot answer you how many percent go to grow, but will grow a lot. Third questions. Approval for server-grade automotive SSD. Yeah, this is ongoing. This is a very complicated project. We are going to have CS to customer by next CQ1. We invest this over 4 years. 4 years, okay? This also is a commitment to show to the auto player, Phison, we are already committed into this business.
I am not, you know, eager to see the business revenue coming from this, but this is a significant, you know, it's a credit in this auto market. Once this happen, I think Phison able to get more business other than the server grade, but also the eMMC, UFS, Gen4 SSD from those auto player. This able to help to build our credit in the market. Call community auto no good for second half. Yeah, due to the smartphone. Smartphone is... Again, Phison, we have so many customized module, again, in the industrial, auto, security, okay? Many different form factor design. I think smartphone is not doing good, but the storage still improve in the content. Storage company, NAND company, losing money, they have to raise the price. We are waiting to gain the business here.
Mobile phone market, especially Android phone, is very, very terrible bad, but when market is coming back, hard to say, but I want to give you one observation. Since by middle June, middle June, we got several rush order from the eMMC and UFS module and controllers. Just a week ago, we got the other rush order, 20 million pieces from customers. For middle, low-end phone, looks like third world country, India, Africa, you know, Southeast Asia, looks like demand is coming back. Of course, the price is terrible, but demand is coming back little bit. For high-end market, I think it's a question, okay? I'm not much worried about myself in the smartphone business. The reason is, the same controllers, same module, we are also going to the automotive, going to the embedded, going to the industry.
I believe we are okay. This question is about Hongxin. We own Hongxin 36% of the share. We not do much business with them. By the way, the chairman is the major shareholder of the company. Why he willing to suffer the loss from Phison? Okay. I think the question from you, right, I'm totally disagree why you asking this, because we are didn't making more revenue with the Hongxin. Hongxin doing their own business, in China business. We are supporting them, controllers, supporting them their technology, and helping them to gain the trust from their auto, mobile, server makers in China. We never asked Hongxin to suffer Phison's loss, because the biggest shareholder is the chairman. He is not willing to do that. Okay? Hongxin H2 question.
Looks like gross margin is getting better. Of course, I hope they are able to keep even by CQ3, very challenging, but definitely the loss is a minus, will be the lowest in these last four quarters, and for sure, by CQ4, they will turn to pro. Okay. Their loss is mainly coming from the inventory loss. The inventory loss realized since last CQ2, and some will happen by CQ3. Samsung keep highlighting to Shenzhen they are going to raise the price, and their cost is low enough, CQ4, they will turn to profit. Hopefully, this able to help them to turn to profit and not become my burden anymore. When Hongxin turn to profit, I believe, you know, China, now they have a three public company in the NAND storage.
Hongxin, definitely the chairman want to bring company to public. Hopefully, then Phison will be forced to assess our share. Now on the other…
Thank you. Now we have the question from Simon Wu. Simon, please go ahead. Oh, Simon, you may start your questions.
You mean to me, Simon?
Yeah.
Okay. Thank you. Thank you very much, Mr. Pua. Your presentation always great. Appreciate. Number one quick question is top-line growth. The July revenue is still NT 3 billion. If the month-on-month, the revenue remain flattish, another kind of the NT 10 billion flattish quarterly revenue for September quarter. Question is, when do you expect sort of a meaningful top-line growth? Because near-term guidance among the, you know, memory makers or supply chain companies, not really bullish. Meanwhile, your July revenue is still flat. Any view for the inflection point to see the meaningful sequential revenue growth? Thank you, sir.
Okay, my policy, if I want to gain the revenue, I think I can easily make it. Just go to the Shenzhen, cutting the module price, taking the loss. I'm not willing to do that. The system business, so far still flat, little bit increased, but the SP slightly down 1%-3%. With their design in Phison, we keep the good growth margin. I am more willing to take their business, but the growth is not that much. Recently, in Shenzhen, they have many, many bid in the eMMC, UFS or smartphone. You know, the price is a - 30% gross margin. I'm not going to take it. If Longsys and BIWIN are happy, let them take. Okay?
Since they are keep losing money to turn to revenue, this will also force the NAND vendors has to selling much cheaper NAND to the module house, right? I'm able to use my own cash to buy more flash with a much lower price. With a certain China, today's market, I have no any interest to try revenue. With the U.S., with the Europe, yes, we have a many design win in the U.S. market, but June, July, their economy demand looks like slightly back, but there are most company executive, they are in the vacation, they're happy, okay? We have to see August, September, when everything back to office, back to normal, market demand turn to better, then we are going to take a more revenue. In China side, if the demand all coming with a big minus, let Longsys take it.
I won't do that.
Yeah, yeah. Great point. Very quickly, sir, you have a great point. The NAND chip price is very, very low, exceptionally cheap. You have a cash, so why don't you buy more and more NAND flash chips from Korea, Japan? Then why don't you try to increase your inventory if you are really confident, to see the NAND price recovery late this year or next year? It's a great opportunity to buy more and more NAND chips. Don't you think so?
Well, first of all, we did. Second, not volunteer, because every NAND makers came to us to ask, "Hey, please, can you?" If you look to our inventory, dollars a month is not increased that much. The ASP itself is almost double right away. We accumulate a lot of ASP on hand with much, much cheaper cost. I still believe CQ3, every NAND vendor will come to Phison to ask Phison to buy, because we are still have a cash without loan, right? Yep, we are happy to work with the NAND supplier, but also at this moment, since Phison, we need to support them in the NAND component buying. We also wish they work with Phison in the new controller project. Okay, this is always Phison's our strategy.
Recently, we also gained several controller projects from our NAND partners, due to we are able to buy the NAND from them.
Yeah. Yeah. You believe the NAND price recovery sometime maybe in December quarter, late summer, early autumn, or, or best case scenario is just a flat?
Yeah.
No photo, correct.
Let me call Samsung CEO later answer you, okay? It's all Samsung decision. If, if Samsung decide, you know, not too aggressive in the SSD, not aggressive in the modules, then price definitely increase. It's all about Samsung.
Oh, I see. Okay. Yeah, I think, great. Yeah, one last question: You think the Phison can do some extra PCIe Gen 5 for the AI server SSD going forward? Because, I'm sure that the lots of the AI server coming from even China, even without Nvidia GPU, I think we can see the AI server. Do you think, sir, AI server, PCIe Gen 5 opportunities?
Yes, we found some architecture, like what we show in this page. Now we are doing POC. Actually, we already proved our concept is workable, but we need to improve the efficiency, the performance. Within four to six weeks, if we work with the server makers in Taiwan, we prove, then we are going to announce and go to the GPU company to work together with them. Four to six weeks. Yeah.
This is a Gen 5, sir, or PCIe Gen 5, or?
Both Gen 4 and Gen 5.
Gen4 and Gen 5. Okay, appreciate. Yeah, all, all clear. Thank you so much, sir.
Thank you.
Thank you. I think we have another question from Donnie Teng. Donnie, please go ahead.
Oh, sorry, I don't have further question. I just forgot to cancel the, the hand raising.
Okay, thank you. I have a question in terms of your operating expense ratios for coming quarters. Do we have any outlook or guidance in terms of this kind of number? Thank you.
I think the total amount still will keep stable in the coming few months. First then, we are not heavily increased headcount like we did 2021, okay? We are still managing. We, we increased some new talent, but not crazy. Also for some internal R&D, which is less performance, we let go. No fire, we let go. By R&D, R&D expense, we believe we are to keep kind of stable in the coming few months. Until Phison turn to the good revenue with a good profit, we are going to increase more R&D expense as a bonus for R&D. If that happens, we are going to increase some expense on R&D.
We believe, someday we will be asked to exercise some share of Hongxin when they are going to apply to public, then Phison will turn to, you know, the good return, then we are going to write off some income to R&D expense. Yeah.
Thank you. In the interest of time, I think that's all of our ends call today. Thank you very much for your time and participations on this call. Thank you, Mr. Pua, for sharing your insights, for commentary and your view for the industry and Phison's. If you have any questions, further questions, you can reach or connect with IR team after the call. Thank you, everyone. We will see you next time.
Thank you.