🎁 Black Friday: 50% off Unlimited and 30% off Pro annual plans
ISC Co., Ltd. (KOSDAQ:095340)
South Korea flag South Korea · Delayed Price · Currency is KRW
127,100
-19,800 (-13.48%)
Nov 20, 2026, 3:30 PM KST
← View all transcripts

Earnings Call: Q3 2026

Nov 9, 2026

Summary

Q3 revenue and operating profit grew YoY, while post-process bottlenecks deferred orders into Q4. AI demand, price increases and capacity expansion underpin record Q4 guidance and a 2027 growth target of at least 30%.

Speaker 1

Good morning. This is the ISC Investor Relations team. We would like to sincerely thank all analysts and shareholders for joining us today. During today's conference call, we will review our financial performance and key business developments for Q3 of 2026, followed by the outlook for Q4 and 2027. This presentation is based on the company's public disclosures and internal management data. We will begin approximately with 15 minutes of prepared remarks, followed by a Q&A session. Before we begin, I would like to make a forward-looking statement disclaimer. Today's presentation includes forward-looking statements based on management's current views and assumptions. Actual results may differ due to various factors, including changes in market conditions, the pace of technological development, and customer demand. Following the conference call, the presentation materials will be available in both Korean and English on our IR website.

An audio recording of the session will also be available for replay. Please note that simultaneous interpretation is provided during today's earnings conference call for our international investors.

Speaker 2

First, let me walk you through our key financial results for the third quarter. In 3Q 2026, revenue reached KRW 71.8 billion, while operating profit amounted to KRW 21.8 billion. Revenue increased by 11% year-over-year, and operating profit rose by 25%. Although revenue declined slightly quarter-over-quarter, we achieved an OP margin of approximately 30%, supported by increased orders for high-value-added products despite the impact of weaker U.S. dollar against the Korean won. There are three key takeaways. First, our AI-focused business mix became even stronger. AI-related revenue reached KRW 62.6 billion, up 46% YoY, accounting for 87% of total revenue. Notably, revenue from AI data centers exceeded KRW 50 billion for the first time in a single quarter.

Second, we expanded our customer base and product coverage. In the data center segment, we broadened our business beyond GPUs to include CPUs and ASICs. In the mobile segment, our supply, which had previously been focused primarily on R&D, began expanding into production setting. We also secured new customers in robotics and automotive. Third, we are maintaining our full-year targets of KRW 300 billion in revenue and KRW 100 billion in operating profit, and we expect to achieve record annual results since the company's founding.

We are preparing to meet customer demand in Q4, supported by increased mass production shipments for AI data centers and mobile applications, as well as the ramp-up of expanded production line in Vietnamese Fab 1. Next, I will discuss the performance of our key business segments. Let me begin with our AI revenue. AI-related revenue reached KRW 62.6 billion in Q3, up 4% QoQ and 46% YoY.

The AI share of total revenue increased by 20 percentage points compared with the same period last year. Notably, revenue from AI data centers exceeded KRW 50 billion for the first time in a single quarter. By product, growth was particularly strong not only in our core AI accelerator business, but also in CPU-related products. We secured a 100% share of the test socket business for all CPUs supplied by a North American customer. Alongside the growth of AI data centers, our application coverage has also expanded. During the quarter, we secured new customers in robotics and automotive, broadening our supply base into the physical AI market. In AI memory, orders increased for HBM4E and SOCAMM2 test solutions, and a substantial portion of these orders is expected to be recognized as revenue in Q4, and orders also increased for equipment.

For the full year, we expect AI memory revenue to reach approximately KRW 30 billion , representing an increase of around 200% YoY. Next, let me discuss our revenue mix between memory and non-memory products. Non-memory revenue reached KRW 61.1 billion in Q3, up 19% YoY and 7% QoQ. It accounted for 85% of total revenue, and growth continued as we expanded supply across key product categories, including GPUs, CPUs, and ASICs, primarily for AI data center applications. Memory revenue amounted to KRW 10.7 billion , down 20% YoY and 33% QoQ. However, profitability improved following ASP increases for legacy memory products. I believe that additional price increase is expected for memory products going forward. Next, I'm going to go to application segments. First is data center. Revenue from AI data centers exceeded KRW 50 billion for the first time in a single quarter.

In addition to our established AI GPU business, we expanded our product coverage to include CPUs and ASICs. In the CPU segment, we secured a 100% share of the test socket business for all CPUs supplied by a North American customer. We also launched WiDER-Max, our modular socket solution, and this is going to take up around 10% of our total revenue next year. This is becoming a major product. In Q3, we also took the first initial order for WiDER-Max, expediting the supply of our products for clients, which is good news. In the ASIC segment, a development project with a North American customer transitioned to mass production. We also expanded our market share in ASIC mass production testing for global big tech companies. In GPU segment, demand for sockets used in system-level testing, or SLT, of next-generation products increased.

Going forward, we will continue supporting customers from the development stage through mass production stage while broadening our product and customer portfolio in the data center market. Next, let me turn to mobile business. The key development in the segment this quarter was the expansion of our supply into the production testing market. We had previously supplied test sockets primarily for R&D for a global smartphone customer. Beginning in Q3, we started supplying these sockets for production testing as well. In particular, orders for SLT sockets used in flagship smartphone application processors based on WMCM technology began to ramp up. Our market share at a North American mobile customer also increased from the existing 20% level to 40%. We expect that this expanded supply to contribute to our Q4 results, and we also secured order for the first half of 2027 as well.

Going forward, we will expand our supply of final tests and SLT sockets while addressing R&D testing demand for on-device AI, further broadening our presence in the mobile market. Next, let me discuss our memory and equipment and materials business. In the memory segment, we are improving profitability through ASP increases for legacy products while expanding the development and supply of high-end AI memory test solutions. Orders are increasing for HBM4E die-level testing and SOCAMM2 production testing, and we're also broadening our customer base. Particularly, HBM4E and SOCAMM2 mass production test, die-level test, and test equipment solutions are all coming as different orders, and I believe that this is going to contribute and become the major products that we supply to our customers from next year.

In our equipment and materials business, we began supplying products in September in line with the major customer's investment schedule for its semiconductor cluster in Yongin, Korea. We expect the scale of supply to increase gradually in line with the customer's production line construction and ramp-up schedule. Next is CapEx investments. The expansion of Vietnam Fab 1 was completed in the second quarter of this year, and the exact production line has been operating at full scale since October. Vietnam Fab 2 is targeted to begin mass production in Q2 of 2027, while Songdo Fab is targeted to begin mass production in Q1 2028. Through these investments, we plan to increase our annual production capacity in stages from 2 million units in 2026 to 3 million in 2027, 4 million in 2028, and up to 5 million in 2029.

On top of this, the automation rate, including pogo and rubber, in the past was 45% level, but it is going to go up to 85% level. Because our cost is heavily labor cost-driven, we should be able to save costs, and this will also in turn contribute to better profitability. However, the timing at which expanded capacity translates into actual revenue may vary depending on customer product approvals and mass production schedules and the ramp-up of new production lines. Vietnam production capacity addition was all finalized in Q2 this year, and because we started the operations in October, I believe this will positively impact our revenue recognition. Depending on the demand of our customers, we have been spending our money and making investments in stages, and this is turning into positive results.

Next, I am going to talk about the reasons for the revenue deferral in Q3 and the subsequent supply progress. In Q3, we secured significant order volume. CPU, ASIC, memory, GPU, major applications, we were able to secure significant orders. However, due to our production capacity constraints at the time, delivery of some orders and the recognition of relevant revenue were deferred to Q4. That is around KRW 11 billion worth of orders, and this was recognized in October. When it comes to operating profit, given that 95% of all the transactions are in dollars, because of the weaker U.S. dollar against the Korean won in Q3, we saw a decrease in profitability. We are currently making sure that we make investments so that production capacity is well-prepared, and meeting the deadline is also well-kept for our customers.

Even though there was not a favorable condition for foreign exchange rate, we will be able to maintain higher than 30% of OP margin level. Now, let us talk about the fourth quarter and full-year outlook. Every quarter, whenever we have earnings call, we are going to provide guidance for the next quarter. The reason why we are giving a next quarter guidance is because we want to increase visibility for our investors. Even if our preparation is not perfect, please kindly understand. We believe that the Q4 revenue is KRW 87 billion-KRW 97 billion, and for operating profit, KRW 28.3 billion-KRW 38.3 billion. Based on this outlook, I believe that we are going to achieve a record quarterly revenue and operating profit in Q4.

For the full year, we are maintaining our targets of KRW 300 billion in revenue and KRW 100 billion in operating profit. There are three key drivers supporting this outlook. First, the expansion of mass production for new AI data center products. We are increasing test socket shipments in line with the supply schedules for AI GPUs, server GPUs, and ASICs at major global customers. Second, increased market share with an existing customer base and a higher proportion of high-value-added products, along with expanded shipments for mobile mass production testing. Price increases are also expected to contribute to the results. Third, increased supply capacity following the ramp-up of the expanded production line at Vietnam Fab 1. This is enabling us to meet demand for Q4 while fulfilling orders deferred from Q3.

We are going to continue this trend so that we can make sure that we supply products to our customers according to their demand. We will also make sure that we will focus on delivery management, production management, shipments, and high-value-added products to make our full-year targets achievable. Now let us look at the outlook of 2027. Supported by growing AI demand and a more diversified business portfolio, we are targeting YoY growth of more than 30% in 2027. This target is underpinned by secure supply volumes and new orders across our key product categories. In the CPU segment, supply volumes for 2027 have already been confirmed in response to customer requirements. In the CPU segment for specific customers, the order volume for 2027 full years is already finalized.

In ASIC segment, TPU products are in mass production, while existing R&D projects are expected to transition to initial mass production beginning in Q4 of this year. In the mobile segment, orders for mass production of application processors for new flagship smartphones from a North American customer have increased, and we have also secured supply volumes through the first half of 2027. In the GPU segment, we have also secured R&D and mass production orders for new products for a North American customer.

In the memory segment, we are pursuing further business expansion supported by increasing orders for HBM4E and SOCAMM2. On the supply side, we are preparing to begin mass production at Vietnam Fab 2 and will expand production capacity in line with the product approval and mass production schedules. To improve our profitability and yield improvement, we will focus on increasing production line utilization and leveraging AI in our manufacturing process.

In 2027, growth target takes into account potential changes in capacity expansion plans and customer production schedules. We will provide further updates on our outlook as order visibility and production ramp-up plans become clearer. To summarize today's remarks, first, in Q3, we recorded revenue of KRW 71.8 billion and operating profit of KRW 21.8 billion, representing YoY increases of 11% and 25%, respectively. Despite the impact of foreign exchange movements, we achieved an operating margin of approximately 30% level. Second, AI-related revenue accounted for 87% of our total revenue. We broadened our growth opportunities by expanding our supply of CPUs and ASICs for data centers and entering production testing in the mobile segment. Third, we expect to achieve record quarterly results in Q4, supported by the ramp-up of our production line at Vietnam Fab 1 and the expansion of mass production for new products of key customers.

In 2027, we will continue to grow, supported by secure supply volumes and new orders. We will improve profitability through phased capacity expansion and enhanced manufacturing efficiency. We will continue to keep our investors fully informed of our business progress and financial performance. Thank you. This is everything, and we will now move on to the Q&A session.

Operator

Now, let us begin the Q&A session. If you have a question, please press asterisk and number one on your phone. If you want to cancel, please press asterisk and number two . The first question is from CLSA, Mr. [Hyun Sung Kim].

Speaker 4

H ello, thank you for the opportunity to give me a chance to ask a question, and thank you for your great audio communication. I have a few questions, and then I am going to move on to the next one. First, it seems like CapEx is being invested much quicker than before.

Are you just preparing to make sure that you can respond to the order from your client? Can you also further expedite the CapEx implementation? Number two, you can see that margin is going to go up by 9% or so. Is it operational leverage that you are enjoying, or is it because of the price increase that is scheduled?

Speaker 2

Well, thank you very much for your question. When it comes to the productivity and the capacity additions that you mentioned, we have a few projects that involve large volumes, CPU, memory, and ASIC clients mostly. Because of the capacity restrictions, basically gave us a hard time throughout Q3. We had to adjust some of the deadlines and lead times with our clients. Because of the situation, we expedited the process of the ramp-up of Vietnam fab.

Especially right now, we are at the initial stage of implementing manufacturing AI solutions. We are working on the training of the workers who we employ there. That was the focus of our manufacturing side. When it comes to Vietnam fab two, we finalized the land agreement in July and August. Instead of building a new building, we were basically using the existing building space so that we do not have to think about utility or water supply of our plant. What we are doing is because we produce and manufacture all the equipment to come up with test sockets. Compared to other companies who are trying to add more production capacity, we are much more quicker when it comes to how we can respond to inquiries of supply.

I know that in the past, we should be able to start the operation around Q2 next year, but we believe that it could be even quicker than that. Because we are investing money in automation too, if we need to respond to more supply and orders, then we should be able to make it even quicker. We are also seeing 30% or over more productive equipment for new facilities, and many of the process is already automated. Let us say in the past, per equipment, there were two workers we needed. Now we can have only one worker per equipment and with better productivity. Please refer to this.

Secondly, when it comes to operating margin going up, I think because it is around Q4 where we see the benefit of price increase for sure, and for our memory clients, we increased prices in first half, but second half we did it again. So from Q4, you can see ASP going up for our memory clients. In addition to this, related to CPU, we have higher margin levels. Related to CPU segment, operating margin is over 55% at this moment. Obviously, this is based on our great market share, and when we are on the table for price negotiation, we have an advantage and an upper hand. For ASIC segment, right now there are some projects that are turning to mass production level from the existing R&D level.

Because some of these projects are all ours, 100% supplied by us, because of that monopolistic status, we have a higher negotiation power, and that is why our operating margin is going to go up in Q4. Q4 and 2027 full year growth, this is all based on the calculation that Korean won is going to be around KRW 1,300 per U.S. dollar, so there's going to be minimal impact about foreign exchange rate difference going forward too. Now let's move on to the next question.

Operator

The next question will come from Ms. Mink yung Kim from Hana Securities.

Minkyung Kim
Analyst, Hana Securities

Hello, this is Mink yung Kim from Hana Securities. Thank you very much. I have four questions. First of all, if you look at Q3, this is below the consensus because of the production issue, so please give us more detail. Number two, what's the growth engine for 2027?

What's the biggest application that's going to contribute the most to the growth? Number four, what is your CPU update?

Speaker 2

Let me answer your questions. For Q3, in general, there was a constraint in our production capacity, not so much about assembly. If you think about sockets, there's preprocess and post-process. Preprocess is processing. Post-processing is what you do afterwards, and the post-processing was problematic. We have to put in a lot of labor and workers during the post-processing, and that was the biggest bottleneck for us to complete the products. I can see that by adding this more capacity, things are improving a lot from Vietnam Fab 1. I know that it's regretful, but I believe that if our revenue had been recognized in Q3, I believe we would have had a greater quarter.

Instead, we should be able to reflect that in Q4, and we make sure that we update our customers and investors as quickly as possible. Because we have full order of volume from our customers, and we have to ship these products by January and February, once we reflect this in Q4 performance, I believe things will look much better. Especially January and February, there are some of the volumes that were deferred because of the big holiday, the Chuseok holiday in September, and we were able to finish that earlier and focus on adding production capacity. I believe Q4 and Q1 of next year are going to look very rosy. For CPU, it's hard for us to reveal a name, but on top of our existing customer, max type CPU client, I believe our wallet share is going to go up further.

The existing customer we have, our market share is around 60%. By Q1 or Q2, I believe our wallet share is going to be around 100%. What's even more positive is that ARM-based CPU, for that particular segment, our wallet share is going up, and relevant orders are going to come in in Q4 and Q1 of next year as well. Going to your third question, when it comes to contribution level, in 2025, the big thing was GPU. This year, it's going to be CPU, and next year, probably ASIC projects that are transitioning to mass production level. Global big tech, their in-house chips are coming up, and then we're supplying products for that.

AI-related big tech companies, there are two of them, we secured some of the order volumes, and I think mass production is going to take place at around Q1 or Q2 of next year, and these are going to be the biggest contributor for our growth. Last but not least, so many are asking about CPO, to be honest, but we're supplying mainly two products. One is transceiver, and the second is silicon interposer. For transceiver, we're doing pins, but customers are asking us to produce rubber-related ones, so we're doing R&D. For silicon interposer, silicon-based interposers are what we supply, and very well-known CPU customers are who we have. I think mass production revenue recognition is going to take place Q1 next year at earliest, Q2 next year at the latest, and probably around then, we should be able to talk about it fully.

Related to CPO test sockets, it's actually Link Edge that we talked about in Q2. It works as a connector, but it also serves as an interposer too. That is why that Link Edge connector-related business and CPO business are all happening at the same time for us. Thank you. Let's move on to the next question.

Operator

The next question is going to come from Mr. Dongkwan Kim from Meritz Securities. Please go ahead.

Dongkwan Kim
Analyst, Meritz Securities

Thank you for the opportunity to ask questions. This is Dongkwan Kim from Meritz Securities. I actually have three questions. The first one is related to price increase. You said you increased it in the first half, and you did it again in the second half of this year. Why did you do that? Next year, what do you think is going to be like for price level next year?

My second question is ASIC, CPU, you can share some of the positive outlook, but what about your Chinese business? Do you have a specific outlook about the Chinese market? My last question is related to 2027. I think you shared that your growth rate is going to be around 30%, very conservatively, but what's an ambitious target? If conservative is 30%, what's your ambitious target?

Speaker 2

Well, let me share my answer. Related to price increase, well, for non-memory products, when chip generation changes, when there is a generational change, then 30%-50% of price increase follows. For example, either pin count is going to go up, or pin layout inside the socket could take place, or pitch could become more sensitive and detailed, and you can charge more.

If there's not much change, then I would say we can charge around 15%-20% more from the existing price level. This year, for non-memory products, there was a significant increase for non-memory, and for memory, once in first half and once in second half, and we have a plan to increase price further. Especially for non-memory products, for our ASIC clients in particular, we have raised around 30% of price, but it's going to be even bigger because given the characteristic of WiDER-Max product, it requires much more sophisticated technology behind it, and of course, that requires price increase. Our major ASIC customers do understand this logic. I would say for ASIC clients, 35%-40% level price increase is going to go up, or CPU, 40% level increase.

For memory clients, I believe we're still going through the price negotiation with them for next year product price, so we can't reveal all the information. But I can say for sure that our price is going to be higher than this year's level going forward. OP margin is around low 10% level now, but our target is to grow this to mid-20% level OP margin. We want to make sure that we leverage this negotiation power so that we can improve our profitability in the legacy business too. Related to AP, for mobile AP, compared to this year, we're going to increase our price by 30% or above next year, and our clients also agreed to our plan. Next year, I think our profitability is going to go up. Our OP margin as a result will also go up.

Related to price increase, how this much is going to contribute to our OP margins? Right now, by application, it's all different, but for non-memory, I believe Q1 next year order volume, we should be able to apply the higher price. For memory segment, by Q2, we should be able to apply next year's new price level. For the Chinese market, I know that many are asking questions. You know that we have changed our strategies from the past, and right now, for Chinese client server, test solutions, or what we have been providing, but from Q4, Chinese memory test sockets will be supplied by us. Especially for those Chinese customers, we are charging a very favorable price level, and the clients were okay with the price level that we would love to charge. Probably, from next year's revenue, we'll see the contribution.

Next year, we'll be able to see great growth in the memory sector in China. Last but not least, when it comes to the growth rate of next year, obviously our next year BP is not 100% finalized, but roughly speaking, I can say that compared to this year, at least 30% growth is going to be expected, and there are two main reasons. Number one, price increase, we'll see revenue growth, but on top of that, we are expediting the addition of production capacity of Vietnam Fab 2. The actual operation is going to be around Q2 of 2027, and as a result, if you add this synergy, and if you Because we have to think about that there is a little bit of gap between now and then, and the permitting issue and all, that is why we set this conservative level of growth target.

Given those risks, I would say we say 30% or over, but let's say our expansion plan goes according to our initial assumption. If permitting goes through also very smoothly and get done quickly, then probably this year's level of growth or even higher level of growth is something that we can expect. If we can secure production capacity earlier, then our growth can be even to 40% or so level. However, as I mentioned earlier, there are some reasons why we wanted to be conservative. Because we are sensitive to the semiconductor cycle in general, we want to be very prudent and cautious. That is why we came up with the number 30%. Q4 earnings call is going to be around the end of January or early February. We are going to share our full year guidance then, so please stay tuned.

Let me go for our next question.

Operator

The next question will come from CLSA's Mr. [Hyun Sung Kim]. Please go ahead.

Speaker 4

Thank you so much for the opportunity. I have two questions related to price. As you mentioned, this 30% or above price level, is it for like-for-like basis? Are you charging 30% more for the same product, or you are charging 30% more because it is a next level product? Number two, you have more than six LTAs. I heard that there is no price ceiling. You said this earlier. Next year, would you be able to negotiate price again next year because there is no ceiling?

Speaker 2

Thank you. Let me answer your question. As I mentioned earlier, because there was production spec change, socket price will change. The increase is going to be more than 30%.

Even if there is a product change, if there is not much change in socket, obviously we cannot charge as high as 30% more. As I mentioned earlier, around 15%-20% more would be what we can charge, given the material cost increase or inflation. If not, we will be able to charge just like the same price. Secondly, related to LTAs, price ceiling is not existent. There is no price ceiling. For clients who are going through generations, obviously we will charge more. Compared to the ones who did not sign LTAs, obviously we should give them some advantage. Instead of a drastic upcharge, there will be a moderate upcharge. I do not think we are going to drastically increase our price. We will be on the middle ground, and they will be happy, and we will be happy. Let me go for next question.

Operator

The next question is come from [Sang Yoon Kim] from CLSA Korea.

Speaker 7

Thank you so much for this opportunity. I also want to ask a question related to LTA. You said 50% of your total revenue is coming from LTAs. Is it the same? What is the change or update going forward?

Speaker 2

There is not much change from that level so far. It is the same level or proportion, I would say. I would say we will see bigger volumes. We are also reviewing signing LTAs with memory clients too. If we can secure that order, I think our memory market share is going to go up further, especially in the market. I know the market is thinking, "Our competition took over our market share," right?

But you can see that these clients are actually reaching out to us to sign LTAs, and once that is done, we will see a lot of recognition and contribution of that LTA. Many different offers are coming in, both from memory and non-memory players, which are very positive. I think some of the headaches that we used to have are also going to disappear. We call it either designated line or LTA. I think LTA-related revenue is going to definitely go up next year. As I shared earlier in the last call, this is going to be a big part of our revenue. If there is any major change, we will continue to communicate with the market. Thank you.

Operator

The next question will come from Mr. Jongwon Kim from Macquarie.

Jongwon Kim
Analyst, Macquarie

Thank you so much for this opportunity, and congratulations. I have one question.

When you talked about CPO, you talked about transceiver and interposer, and you said it is mainly Link Edge. So it is not necessarily a test socket, does it? Should we consider it as a new product? Because when you say CPO, it is basically a new product that requires new test socket. I want to understand whether this is a testing solution or a product with a new functionality.

Speaker 2

Let me answer your question. Not just us, but other players and companies would have a hard time calling CPO a test socket because, I know that we are providing test solutions to North American clients, but I think transceiver or interposer demand are going to be bigger, and interposer demand specifically going to be much bigger because on the substrate, the transmission speed is very important.

Because you have to be connected to optical communication, silicon-based products are what is favored by customers. As a result, I believe more so than test sockets, relevant orders are coming in much bigger volumes. So test sockets, of course, we are making them, but transceiver or interposer order volumes are going to be greater, but we will see. Let us continue.

Operator

The next question is going to come from [Yi Sung Shin] from JP Morgan. Please go ahead.

Speaker 9

Hello. Thank you so much for this opportunity. So, I want to ask a question about demand of Q3. I heard that GPU Rubin is being delayed. You said that you achieved lower than consensus performance because of the bottleneck in production, but maybe growth is demand is slowing down. That is my first question.

Number two, you said you are going to have a higher revenue in Q4, and you said that mobile growth is going to be around 30% next year. I want to ask, is mobile segment like KRW 30 billion level? If it is going to be 30%, that means your revenue in general is going to go up really significantly. That is my question. Number three, not just you are increasing prices, but other players are increasing prices too. So what is your long-term OP margin target?

Speaker 2

To answer your first question, Q3, I would say the main reason why we had a slower Q3 is because of the bottleneck in our production and manufacturing process. I would not say there was a slowdown in demand. There was a delay, but that does not mean there was a slow demand.

Like, the launch schedule was delayed, and design change was there, and there were changes before launching the product. Obviously, sockets have to be redesigned, and that's actually better for us because we can have a higher volume of revenue. Besides the delay of the product, I wouldn't say there's a slowdown of demand. The existing ones recognized and whatever that went to change, that part, that section of the revenue is going to recognize in Q4. I wouldn't say the delay of new product of our customers was the main reason. Of course, it's one of the main reasons, but the biggest reason was our constraint in production capacity, especially the post-process bottleneck that we talked about earlier, like the inspection and processing treatment, especially the inspection side. That's where we saw the biggest bottleneck.

We should have investigated and inspected the products quickly and then shipped them quickly so that all the products are recognized as our revenue, but there was a delay in that particular stage, which is quite regretful. We didn't think this was going to happen, so we're making sure that we make up for this going forward. Related to AP, there's going to be a drastic increase of orders. If you think about the current promise that we talked about with our customers, 30% or above is the right amount. Even some are going to be deferred to next year performance. Still, we'll be able to see a huge growth in mobile segment. I would say, the biggest quarterly performance was like KRW 20 billion once, but I think this Q4, we'll be able to see similar to that level of positivity.

Internally, what we view very positively is because we took away our competition's wallet share, a lot of them. We are not just doing rubber, but pogo too. For our pogo sockets, I believe that our pogo sockets are quite highly viewed, and based on that recognition, we took away our competition's wallet share, which is quite positive. On top of this, not just pogo, but there are projects that are trying to transition from pogo to rubber. If you add all these positive elements, I think next year mobile is going to contribute a lot to our growth. In the long run, our long-term OP margin rate is compared to, based on socket, it'll be 50% or above. For our total OP margin, 40% or above, or up to 50%. When is this going to happen? Not next year. I would say around 2028.

In achieving these targets, of course, there are challenges and constraints. I would say that if you look at only socket business that we have, we have extremely high margin. For rubber, over 40%, and if you combine rubber and pogo, it's around 40% OP margin. That's our average. I would say if it can increase high value-added product proportions, if it can take over some of the existing pie of our competition, we should be able to see a very healthy OP margin rate, and we're working hard to expedite this process and make that happen quicker. In the mid to long term, socket revenue is going to be around KRW 650 billion- KRW 700 billion, and that's going to be around 2029 to 2030. I believe both revenue and operating profit are going to grow accordingly.

So in the mid to long term, please stay with us. Let me go to our next question.

Operator

There is no one who asked or raised a question yet. If you want to ask a question, please press asterisk and number one. The next question is coming from Mr. [Hyun Sung Kim] from CLSA. Please go ahead.

Speaker 4

Sorry that I am asking so many questions. I want to ask a question related to automation. According to what I recall, current 45% level is going to go up to 85% level. What is the relevant timeline? I believe your cost structure is going to be better because you will have 2x efficiency. For example, your distribution cost is going to go down, logistic cost will go down. What are some details?

Speaker 2

Let me answer your question.

For rubber and pogo, both combined, I would say rubber has 16 processes and pogo has 15 processes, and rubber is much more automated. I would say rubber specifically 60%, and pogo automation rate is around 30% in my understanding. If you combine them, I would say our automation rate is at around 45%. For rubber, besides assembly process, we have a plan to automate everything else. For pogo, also besides assembly, we want to automate all the process automated. For pogo, I would say labor is intensive, and the biggest labor is required for inspection. When you do automation, there are around 30 people go for the process, but 30 people will be cut down to two people. So we can save a huge amount of labor cost.

According to our internal simulation, currently our labor cost takes up around 30% or so from our total cost. But if we go for extreme automation rate all the way up to 80%, the labor cost is going to be diminished to 10% level, meaning we will be able to see great profitability, which will help us in our OP margin. So, this is a positive trend. The target timeline is for the fab one, I think we should be able to conclude everything by February of next year, because for added capacity, it is all automated anyway, and we are just replacing the old lines with automated lines. For fab two, I think as we complete automation, we will already be there. So including Korea and two Vietnam fabs, by Q1 of 2028, everything will be completed.

Probably in October of 2027 or so, both in Korea and Vietnam, they will be almost fully automated. Thank you. Let us go for the next question.

Operator

There is nobody who has asked or raised a question. If you want to ask a question, please press asterisk and number one. The next question comes from Lee Sang Sik at JP Morgan.

Sang Sik Lee
Analyst, JPMorgan

Thank you for the opportunity to ask a question. I have a final question. If you look at memory revenue, I think the quarterly is quite bumpy throughout the year. You can see the memory market is booming, but Q3 memory revenue is diminished. So, what is the trend? How should I understand the memory revenue trend for you?

Speaker 2

Let me answer your question. If you look for memory, there are more legacy customers.

If you think about last year perspective, I would say we have cumulatively a KRW 30 billion revenue from memory. We have doubled that this year because we will see KRW 64 billion worth of revenue from memory. NAND is decreasing, but everything else is better. In Q4, I would say what is really good is HBM or HBM4 or HBM3E and SOCAMM-related orders are coming in. Especially, if you look at Q4 memory revenue of this year, HBM SOCAMM is taking up around KRW 12 billion. So, whether it is AI or legacy, I think we will be able to see a big growth. Actually, from legacy side, the volume of inspection itself is not that big. So relatively speaking, volume is not great and the pricing was not so great.

That is why you may think of it that way, but because HBM and SOCAMM are going to reflect it in the second half of this year and going forward, I think it is still positive. Next year, I believe that next year memory prospect or outlook, if I have to do it, I think it will be similar to this year or bigger. I think it is going to be different for memory because more so than legacy, before DDR6 comes out.

I think more so than legacy, HBM or SOCAMM-related revenue are going to be much bigger because we are talking about these projects with our clients more so than legacy products. So, I think it is still hopeful. In 2027, it will grow, and compared to 2025, we grew or double. Next year, volume could be similar, but profitability will be much better in our memory division too.

Due to time constraints, we will conclude as is. My last point is that, I mean, it is internally very regretful to see achieving below consensus level in Q3, the revenue issue in September. I think we should have done better management because it was not the assembly, but the inspection and shipping, the post-process for sockets are the bottlenecks that stopped us from seeing the revenue recognition. Still, Q4 and Q1 next year, we will be able to see even higher growth. So please continue to stay with us, and we will be able to get back to you with better results in Q4 and 2027 Q1. Also from this quarter, we are going to give quarterly guidance going forward every quarter to increase visibility of our investors. Please stay tuned. This brings us to the conclusion of today's earnings call. Thank you.