Good afternoon and good morning, everyone. Welcome to Advantech third quarter 2026 earnings result call hosted by Daiwa Capital Markets. This is Shang Yuan, covering tech hardware analyst at Daiwa. It's our pleasure to have Advantech management team to be with us today. On the line, we have Eric Chen, President of the GM and also CFO, Miller Chang, President of Embedded Sector, and also Grace Liao, theInvestor Relations Manager, to be with us today. For today's meetings agenda, we will start with Grace's presentation, followed by Eric's outlook update. Lastly, we will have the Q&A session at last. Without further ado, I will pass the microphone to Grace. Grace, please go ahead. Thank you.
Thank you, Shang. Good morning and good afternoon, ladies and gentlemen. Thank you for your time today. This is Grace Liao, the Senior IR Manager of Advantech. Regarding our third quarter 2026 financial results, third quarter revenue reached TWD 32.44 billion, increase 83% year-on-year, 24% quarter-on-quarter. Gross margin rate reached 36.7%, mainly impacted by material cost up. Operating profit reached TWD 6.8 billion, with OP rate reported at 20.9%. Third quarter non-operating net income totaled TWD 1.1 billion, mainly driven by dividend income and one-time disposal gain of Japan office building. Q3 effective tax rate was 21.5%. Net income reached TWD 6.22 billion, increase 125% year-on-year. Earnings per share in third quarter 2026 was TWD 7.15. Both third quarter top line and the bottom line reached record high of the company history.
Accumulated in year to third quarter 2026 performance, year to third quarter sales revenue reached TWD 79 billion, increase 49% year-on-year. Gross margin rate was 37.6%. Operating profit rate reached 19.8%, improved from 16.5% from the same period last year. First three quarters net income reached TWD 14 billion, increase 88% year-on-year. Year to third quarter 2026 earnings per share reached TWD 16.21. For regional performance, in terms of U.S. dollars, year to third quarter 2026 revenue reached $2.5 billion, increase 46% year-on-year. For regional performance, all regions enjoy high double-digit growth increase. Total major three markets accounts for 70% revenue contribution. North America increased 58% year-on-year, driven by semi equipment, robotic, and cloud infra projects. Europe market increased 36% year-on-year due to automation, medical equipment projects.
China market increased 55% year-on-year due to strong channel sales. From application perspective, China semi equipment, automation, and smart energy projects remain strong. For Taiwan market, increase 72% year-on-year, driven by semi equipment, automation, and also transportation projects. For sector performance, most of the sector enjoy double-digit growth year-on-year. IoT Automation increased 25% year-on-year, driven by factory automation, especially in North America, China, and Taiwan markets, and also transportation projects. Intelligent Systems sector increased 49% year-on-year, outperform in semi equipment, cloud infra, robotic, and AI projects. Embedded Sector increased 40% year-on-year, driven by medical equipments, especially in North America, Europe, China, Middle East, and Africa markets. Machine automation and mission-critical projects also strong. Intelligent Service increased 7% year-on-year. Strong growth in healthcare projects, however, decline in retail and hospitality.
For balance sheet, this is also my last page. Cash and cash equivalents accounts for 18% of the total assets, slightly lower than previous quarter due to cash dividend payment in third quarter. Generally speaking, inventory remain well controlled. Inventory accounts for 23% of the total assets, in line with the previous quarter. Inventory turnover days reported 95 days, while cash conversion cycle, CCC, was 86 days, both within the comfortable range, reflecting healthy and stable financial structure and capital efficiency level. Now, I am handing over the time to President Eric. Thank you.
Thank you, Grace. Good morning and good afternoon, everyone. Thank you for joining today's meeting. This is Eric speaking. Let me share some comments on our third quarter and year-to-month results. In the third quarter, our revenue reached $1 billion, a record high for Advantech. Gross margin decreased by 1% from the second quarter, mainly due to the higher cost of key components such as SSD, DDR4, which make up a significant portion of our product cost. Though we passed this cost onto our customers, the markup rate on these capital items was lower than the company average gross margin. This diluted our gross margin performance, and third quarter gross margin was 36.7%. Year-to-month margin performance was 37.6%, sorry. Operating margin rose to 20.9%, and the year-to-date operating margin was 19.8%. The operating expense rate was 17.8%, down 5.4% year-over-year, indicating Advantech has well-controlled expenses.
Regionally, the U.S., China, and Taiwan markets were strong, growing 58%, 55%, and 72% respectively. In the U.S., a strong demand was observed in the medical and semiconductor sectors, with a higher BB ratio of 2.12. In China, growth was mainly driven by the semiconductors, factory automation, and medical sectors. The Taiwan market saw 72% growth, mainly driven by the semiconductor sector. Japan and Korea grew in line with corporate average. From a product perspective, AI system, embedded, and Advantech Service+ posted growth above 40%. Strong demand for video streaming and semiconductor drove AI system high growth, while the medical industry drove growth in the Embedded Sector. For Advantech Service+, the growth was mainly fueled by increasing demand for SSD, memory, and hard drive and shipping cost. Now, let's look at the BB ratio trend across regions. As shown, the BB ratio for the third quarter was 1.64.
The U.S. BB ratio rose to 2.12, while in Europe and China, it increased to 1.52 and 1.54 respectively. It is worth noting that Advantech booking order has grown for five consecutive quarters since third quarter 2025, while shipment has increased for four consecutive quarters since fourth quarter 2025. Next page, please. As we look ahead to the fourth quarter, we expect revenue to range between $870 million and $880 million, based on the exchange rate of $1 to 31.8 New Taiwan dollars. In terms of margin, we expect gross margin in the fourth quarter to be between 36% and 38%, with operating margin projected to be between 18% and 20%. This concludes my comment on third quarter and fourth quarter guidance. Thank you.
Now we will hand over the time to Shang. Thank you.
Thank you. Now we are Thank you, Eric and Grace. Let's move on to the Q&A session. We will start with the questions that we collect from the investors before the call, and then open up the floors later. Before we jump into the Q&A session, just want to confirm with the 4Q guidance on the revenue side. For the current of the guidance coming from the Advantech management, I would like to confirm whether the revenue will be $ 970 million- $ 990 million.
Yes, yeah.
Okay. Thank you. I think the very first question has to do with the financial BB ratio and outlook. The first questions that we collect will be, how should we think of the near-term demand BB ratio in the quarter and in October? Any update on the latest of the design-in and design-win numbers?
Yeah, let me reply these questions. Maybe Miller can add some color. The quarter three BB ratio is 1.64, driven mainly by strong semiconductor demand in U.S., China, and Taiwan. These three regions account for 80% of semiconductors demand. The year-to-date semiconductor sector BB ratio was 2.62, higher than the YTD ratio of 1.61. It's very high, yeah. In addition to the semiconductor sector, the medical sector in the U.S. and Europe, and the automation sector in China also showing a promising BB ratio. In contrast, the BB ratio in retail was below average. This is regarding the BB ratio. As for the October BB ratio, it was 1.60. Please note that our shipment in October hit a record high, and the total revenue was around $352 million. We have a record high in October. We will have a formal release this Friday, right? Yeah.
Regarding the design-in and design win figures, through Q3, design win in U.S. reached 687 million EARs, with 161 design win deals. EAR is short for expected annual revenue, and it increased 262% year-over-year. The design win indicator in the U.S. is quite strong. This is my answer for the question one.
Thank you, Eric. For the second question, how should we think of the price trend going forward toward the current cost inflation, and how they would impact the company's gross margin trend in the 4Q and beyond?
Yeah. Recently, AMD, along with Intel, increased its selling price by about 10%-20% across different SKUs. PCB pricing is also raising by around 15%-20%. By contrast, SSD and memory price are more stable. Internally, we have adjusted our selling price to reflect the CPU cost increase. We expect fourth quarter gross margin performance to be close to the third quarter, just as guidance. As mentioned earlier, our gross margin fell by 1.0% versus the second quarter, mainly because SSD and memory price rose sharply. Although we passed the upside to our customers, the markup rates cannot align with our corporate GP. I just give you a real example. In our server products, the CPU, memory, and peripheral used to account for 30% of the total product cost.
Now, it had risen to 70%, and the 40% cost gap was mainly due to the price inflation in SSD and memories. Normally, we will pass through this 40% cost gap to our customers, but the markup rate would be below our corporate gross margin average because this peripheral price are quite transparent. We cannot have same markup rate, corporate margin rate to markup to our customers. Basically, if key component price and allocation become more stable, we have a better chance to optimize our gross margin performance. Our goal is still return to our GP to 40%. It is our long-term goal. Thank you.
Thank you, Eric Chen. The next questions, given the concern that the cost inflation might be driven order pull-ins, could you provide more colors on the current order backlog and the extent of any kind of pull-forward demand?
Yeah. The pull-forward demand for shipment is rare, partly because of component allocation issues and partly because most of our products must integrate into their overall system before end use. Most of our customers don't have the intention to pull forward demand unless they have an ongoing project on hand. In terms of backlog, the backlog for November, December, and the year 2027 is around $2.5 billion, showing that customers are pressing order ahead to secure a higher allocation priority. This was due to some component allocation issues, and the fulfillment lead time is sitting at 20 weeks, more than 20 weeks. We have five weeks needed to review demand and supply trend. Order postponement and the cancellation rate are always under watch list. So far, the indicator are on the right track. So in short, we have around $2.5 billion backlog on hand, including November and December.
Also, the postponement and order cancellation rate is still on the right track. Thank you.
Thank you, Eric. The next question we had is with the AI contribution expected to raise in the second half this year and also in 2027, how will this impact the selling price, excluding the impact coming from the cost inflation? Also, the rising AI contributions, how would be the impact toward to the gross margin? Will higher AI sales mix be margin accretive or dilutive to your blended gross margin?
Okay. AI product revenue contribution through the third quarter accounted for 24.5%, up 7.3% year-on-year. In terms of ASP, the impact was significant in the high-end edge server product line. Let me elaborate on the details. In the past, the most popular IPC we sold to customers had an ASP of around $5K-$6 thousand, depending on configurations. For high-end edge server products, in some specific cases, the ASP will up to $80K-$100 thousand because the system is configured with powerful GPU cards, up to six Xeon CPUs, high bandwidth memories, SSD, and et cetera. So the ASP in a specific edge server project will be significantly higher. As for the Edge AI product margin performance, it really depends on the product and the product mix.
Edge AI products in the medical and automation sector have a higher than average margin, so as AI design projects. The GP will outperform their non-AI product line by 2%. However, in some high-end edge server projects, as mentioned earlier, even though the ASP is higher, margin will be dilute because some component, such as GPU, CPU, and memory, accounts for a large portion of product cost. Due to this, retail market price are transparent, so we can't sell the high-end edge server product with high margins. The average high-end edge server GP is around 28%-30% below corporate GP performance. So to the question, the margin of our Edge AI product really depends on the applications. It's very diverse. Thank you.
Thank you, Eric. So that conclude for the first part. For the second part of question is on the component and supply condition. The second questions will be, could you provide an updates on the current component shortage, particular in the memory chips and also the PCBs? I think you previously mentioned about the price hike from the AMD is around 10%-20%, PCBs around 15%-20%, and then stable in the SSD. Then how about the current of the supply condition? Can you also elaborate a bit on that? Thank you.
Okay. Thank you. I just got a brief back from our procurement heads. The overall component supply remains challenging, although condition varies across product categories, and let me explain more details. For memory, the supply condition have gradually improved for mainstream memory products, while server-grade SKUs remain constrained. Pricing pressures continue across the market, driven by strong demand from AI and data center applications. The mainstream memory product is quite stable and had gradually improved. For the CPU and chips, certain CPU and chipset products remain under allocation, particularly selected legacy and embedded platforms, which impact our ELD revenue a lot, with extended lead times. This is regarding the CPU and chips. For the PCB, supply constraint are primarily concentrate on certain upstream material, particularly CCL, we call Tongbo Jiban, rather than overall PCB manufacturing capacities.
Advantech continues to proactively manage these challenges through long-term demand planning, strategic supplier partnerships, and early order commitment and supply chain diversifications. While some supply constraint and cost pressure may persist in 2027, our priority remains ensuring supply continuity, managing cost exposure, and minimize the impact on our customer deliveries. This is regarding the PCB, memory, and main chip situation. Thank you.
Thank you, Eric Chen. For the second questions from this segment, how much of the first nine month of revenue from Advantech was driven by the volume versus the price hike?
We have a 49.1% top-line growth so far. Revenue growth driven by price increase accounts for 20.5%, while volume growth contribute 28.6%. 20.5% driven by the price increase and 28.6% by the volume contribution. Thank you.
Thank you. Our third part of questions is regarding the regional and sector performance. How should we expect the growth from each region, including North America, Europe, and also China, and their respective key growth driver in the 4Q this year and also in 2027?
Okay. Thanks for the question. This is Miller Chang speaking. According to the question 3.1, according to our Q1 to Q3 2026, the actual revenue result and also the result for the book-to-bill ratio, our Q4 2026 expects to keep the same strong growth momentum. For 2027, currently, we are ongoing the process for the budget planning and also the feedback from our product group and also our region, global region business unit. We are positive, and we will update you after we have final budget planning for 2027. Regarding the key growth driver from each key region perspective, Advantech North American, as Eric mentioned earlier, the semiconductor equipment, AI infra, cloud infrastructure, medical equipment, and new robotics segment are very strong and will be the key growth driver from North America continuously.
Advantech Europe, industrial equipment, medical equipment are the main growth driver for this year growth and will continue for next year growth driver. Advantech China, industrial automation, industrial equipment, semiconductor equipment, and including the medical equipment and also new robotics segment are the growth driver for Advantech China region. This is the answer for the question 3.1.
Thank you, Miller. Regarding the 3.2 question, how should we think about the demand outlook from semiconductor in U.S. region, and what about the AI infrastructure globally? Could you provide more colors on that? Thank you.
Okay. About the question, our semiconductor equipment business this year has delivered exceptional growth in 2026. Looking ahead for 2027, the outlook remains very strong, supported by the investment, continued investment in semiconductor capacity and also advance the manufacturing. We see two area of the opportunity. One is the semiconductor equipment, and the other is AI infrastructure. The first about the semiconductor equipment, Advantech has built a long-term relationship with the leading global semiconductor equipment manufacturer. Our product support application across front-end wafer processing, and back-end packaging and testing. Especially for our product from the edge computing, industrial I/O controller, and also industrial server. As we can see, this demand will extend into 2027 next year.
Our North American business remain an important driver, and our customer, their investment across Asia and also Europe will create more opportunity for Advantech business as well. The second topic is about the AI infrastructure investment that bring the demand of the industrial equipment. Our goal is to provide the edge computing and industrial edge server platform using in the equipment supporting the AI infrastructure. The related relevant application will include the power system, energy management, cooling equipment, and also the industrial automation. This area, Advantech already has established the product and customer relationship. We will work closely with the equipment manufacturer and system integrator, SI, to address all those requirement from computing, control, and also monitoring and requirement. We are very positive on the semiconductor and also AI infrastructure business globally. That's the answer from my side.
Thank you, Miller. Now we move on to the part four, which is the HAI progress. The 4.1 question, I think this is also the question that we normally would be asked by the investors a lot of time. How does Advantech define the HAI revenue, and which products with solutions are included? Could you give us more colors on that?
Okay. Let me try to answer these questions. We calculate HAI revenue based on the following rule. First, from a hardware perspectives, the system products or all level products include a GPU, NPU, or TPU. It provides a certain level of AI computing power, around 10 to 200 TOPS or more, and can support HAI inference workloads. This is from the hardware perspectives. Then we come into the HAI revenues. We categorize the HAI revenue into two segments, HAI products and HAI components. HAI products use main AI chip from NVIDIA, Intel, AMD, Qualcomm, and NXP, and MediaTek, also Rockchip in China as well.
Edge AI component, our GPU card or AI chips are sold as standalone orders rather than as plug-in or bundle into our edge products. Year to date, Edge AI component represent only 5% of the total Edge AI revenues, while Edge AI products accounts for 95% of total Edge AI revenues. Right now, the total revenue was $603 million, up 109% year on year. Let me clarify. Edge AI product plus Edge AI component accounts for 24.5% of Advantech total revenue. But in terms of Edge AI revenue, it was $ 603 million year on year, grew by 109%. For this part, Edge AI component, standalone, are for sale, but only present 5% of the total Edge AI revenue. Second, from a software perspectives, use case are many in AI applications such as retail, factory automation, and robotics.
In a normal case, a customer adapt Advantech WISE or robotics software factory, combined with Advantech hardware and third-party solutions to enable their AI use case. So far, the pure software revenue we generate from AI use case is quite small. So we calculate from hardware perspective also include the software perspectives. This is my answer.
Thank you, Eric Chen. For the second question, what new applications are enabled by the high-performance embedded boards, and how does Advantech differentiate yourself from other IPC vendors beyond the hardware performance? Could management share some examples on the progress within the, I would say, the Edge AI side? Thank you.
Okay, let me try to answer this question, 4.2. Actually, many industrial applications require the high-performance computing, such as the medical imaging processing, AOI automatic optical inspection in the factory. They are all good example of the application that require both high-performance computing and AI acceleration. In medical imaging processing, as an example, computing power is essential for the task such as the CT, MRI equipment, the image recognition, noise reduction, and also 3D visualization. So AI add the capability such as image enhancement, abnormality detection for the clinical interpretation, right? In AOI factory automation, the high-resolution camera, multiple inspection angle, 3D imaging generator are much larger volume of the data. So system must process this image and identify a very small defect within a production cycle time. So the high-performance computing will be a requirement for the upgrading of the edge computing in the factory automation area.
Those applications will create a demand for more powerful computing platform across to the equipment. Advantech will expand the opportunity from embedded computing platform to the GPU, NPU acceleration system, and also the industrial edge server depends on the workload and system architecture. Regarding the differentiation to our peer IPC company, I would say a few points. The first number one is our global and local services. Most of the Advantech product are developed from Taiwan, but service global and locally. We have a region team cover more than 100 cities and 27 regions. We do provide a very comprehensive edge computing platform across all the world-leading silicon partner and the collaboration. This is the first point. Second is the software value added.
Our corporate WISE and IWS integrated WISE solution for factory, for energy, for retail, and healthcare. We put the hardware and software together to provide an industrial agentic AI solution to our service industry that we do this can create a great advantage than our competition. This is the answer from this question. Thank you.
Thank you, Miller. We move on to the fifth part, long-term strategy. The first question will be, what will be Advantech's key growth drivers over the next decade, and how significant could Edge AI and robotics become?
Okay, it's a long-term question for next decade. We can see the industrial computing become more and more important as the equipment grows more intelligent and AI move into the real-world application. Our established market, that means our existing big service market, including the semiconductor equipment, factory automation, medical equipment, provide a very strong foundation. Customer increasingly need high-performance computing for inspection, imaging, real-time control, and AI inference. They could add opportunity to Advantech. This is the first, number one, regarding our existing and established sector and industry. We also see the opportunity in emerging applications such as robotics, autonomous system, energy infra, and mission critical. Those applications require a reliable computing close to the equipment together, and the software integration, and also industrial spec expertise.
Edge AI is a core technology that we can extend our value to deliver to many of our service business and industry. Our focus is to help our customer integrate AI into their equipment and move from initial development to the deployment at a scale. Robotics represent a long-term growth opportunity. It is also the good example for Advantech to continue to expand the market to the emerging territory sector. We have established a dedicated robotic business unit since three years ago. The revenue contribution is still small, but with more than 200%-300% growth this year, 2026. We see the potential for this robotic market become a meaningful business for our expansion. Our goal is to provide edge computing, for sure, is our core business, and support software to help our customer to accelerate their robot application development to the market.
For the next decade, we expect the growth to come from the both established business and market, and also develop the new emerging market and application. This is the answer from my side.
Let me try to add some color for this, the first questions. One of our core provision for 2030 is to become the global leader in Edge AI platform solutions. We are dedicated for five key industrial sectors, factory automation, energy and utility, Intelligent Systems and AMR, including robotics, healthcare, and smart city service. By concentrating on these sectors, along with the major silicon vendors, we aim to drive significant growth and secure Advantech long-term success. It is quite important we have four focus areas. Beyond hardware, we offer the WEDA Core WISE Edge development architecture platforms, enable our solution partner to build, scale, and involve AI application from prototyping to productions. Our objective is to foster an Edge AI ecosystems that empower end users to deploy AI solution more quickly, efficiently, and with clear measurable ROI.
We have the vision to become the global leader, especially in the Edge AI solution platform. We also provide comprehensive software platforms to our users to enable the Edge AI ecosystems. It will accelerate the Edge AI use case with a measurable ROI. This is my add-on for this question. Thank you.
Thank you, Eric Chen. The last questions from this long-term strategy is that, which industrial applications are driving demand for higher computing power, and how do you expect this demand to evolve?
Okay, the question for 5.2 actually is very similar to 4.2. I try to give a short feedback. Actually, from the medical equipment, medical imaging processing, also the industrial equipment for automatic optical inspection, these two area segment are currently our big business foundation, and also the growth engine for continuously for Advantech. They can drive the demand for high-performance computing power. Customer initially, they came to Advantech looking for a single edge computing, and then we can support them to upgrade to the high-performance computing with the AI acceleration computing power together. This included the high-performance computing power and also AI acceleration module accelerator, and then we also provide the industrial edge server and solution to support our customer. These are all about the business and demand evolution.
Yeah. Let me add some color for these questions. I think AI in the semiconductors, medical, and robotics sector play a very important role to drive the demand for higher computing power. Semiconductor manufacturing process are moving toward a size smaller than 2 nanometer, and the packaging technology continue to advance. Semiconductor process such as testing, inspection, and attachment will require increasing precise measurements and computing power to ensure the higher yield rate. Semiconductor is one of the key drivers. Furthermore, AI-based image interpretations and comparisons are becoming increasingly common in medical applications, just as Miller Chang mentioned earlier. All this requires high-performance edge computing devices, especially for the medical sectors. Last but not least, humanoid robots move and act by integrating perceptions, we call the input, decision-making, we call the AI brain, and actuations, including motor and joint, into a seamless, synchronized loop.
It also requires a high-end computing brain at the edge side. I think the semiconductors, medical, and robotics sectors are the main drivers for high computing power at the edge side. This is my add-on. Thank you.
Thank you, Eric, Miller. That will conclude the session for the first collected questions. Now we will open the queue for the online participants to ask questions. If you wish to raise a question, you can press the raise hand button and then unmute yourself to ask questions, or you can type your questions in the chat box. I will read it out for you to the management.
Sorry. One of the topics we didn't mention today is regarding the expense for our RSA. We issued RSA. The board already approved RSA as a means to restricted stock awards last Tuesday, but I'm not sure.
Thursday.
Thursday. Oh, sorry. Thursday. The total expense is around TWD 1.42 billion, and we will divide and accrue the expense from the year 2026 to 2030. The EPS dilute for each year is around 0.14, 0.8, 0.46, 0.2, and 0.03. The total EPS dilute will be 1.61 as a limit and won't impact our earnings per share at all. This is just information shared with other investors. Thank you.
Thank you, Eric. I think we got the first questions on the chat box. Clive would like to ask you, given the record bookings and 1.6 BB ratio in the quarter, Clive would like to know why is 4Q revenue guide down quarter-over-quarter? Is this a supply or capacity issue or a demand issue?
It's a good question. We are conservative for the Q4 guidance because in Q4, normally, it's very close to the Christmas holiday seasons. Also, we will have a system upgrade. We will upgrade our ERP system from on-premise to cloud. The working day we have done from, and the working day for the quarter four is shorter than the quarter three. But actually, we have a great chance to surpass the third quarter's revenues that we expect. Ever since the buy, also the demand situation is getting better and better. We still have a great chance to surpass the third quarter's revenues, which means we will have a great chance to over $1 billion again in the fourth quarter.
Thank you, Eric Chen. Online, we have Eileen raising the check. Please unmute yourself and ask a question. Thank you.
Okay. Thank you, Eric Chen. My question regarding the top-line growth. Do we have enough capacity to support the revenue growth, like QoQ or even in 2027? This is my first question. Okay. Maybe let me finish. Second question is, how much as a percentage of total sales is coming from the semi-related, like semi equipment industry? Also, my third question is on OPEX. How should we think about fourth quarter OPEX? Typically, is it a higher QoQ in terms of absolute amount, but considering the higher sales, how should we think about OPEX ratio? Thank you.
Okay. Let me try to answer question one and question two. Regarding the capacity issue, actually, we already add on two SMT lines, one in Kunshan and another in Huai'an. Also, Japan's plant has upgraded their SMT capacity in the middle of this year. So right now, capacity is ready. The capacity rate for the China and Kunshan, in my opinion, is over 100%. It's around 103% or 108% for both sites. For Japan, it's nearly 100%. So, at least regarding capacity. For the semiconductor accounts for the total revenue, I have not exactly figures on hand, but in my opinion, it's around 15% of our total revenue regarding the semiconductor, generally from the semiconductor revenue. This include a front-end wafers solution, also the back-end solutions. It's around half and half.
Also, the major contribution from China, from U.S., and from the Taiwan accounts for 80% of the semiconductors revenues. This is my figures on hand. Okay. For the third question regarding of expense, maybe Grace Liao can try to answer these questions.
For the third questions, third quarter OPEX, quarter-on-quarter, I think would be slightly improved quarter-on-quarter basis because according to Eric Chen mentioned that we do have RSA expense will be starting to impact from the Q4 this year. So, I think in the operating expenses level, I think Q4 will be slightly higher than the previous quarter. Yeah.
Yeah. Thank you, management, and also Eric. We will go back to the chat box questions. The next one, client would like to know about the Chinese competitions in China. The client also would like to know your current strategy in China regarding to the positioning, manufacturing. Yeah, that would be the questions from the investor.
Thank you for the question. Let me answer the China market and also competition. Actually, the China peer company, the competition is always there, especially after COVID, after the China economy decline and also uncertain recent years. So, Advantech, yes, in the past two, three years ago, we really faced the price competition, cost competition in China market. That also shows that we have revenue decline in around 2023, 2024. However, some action we take that already happened and also have the great progress. Let me show you some of that. One is we established the China Kunshan R&D and product development team located in Kunshan, nearby Shanghai. We have more than 200 product manager and also engineering development team there already, which is supported our customer to provide the product roadmap, design, development, manufacturing, and also service locally in China.
That equal to make the China product for China local business. That result is very good ongoing and progress. Second, especially for China market, some long-term support and longevity services also require is most important for some industrial, the high industrial grade of the equipment, machine, and services. Customer will, they come back to Advantech to request a more solid product, longevity services, and also the quality as well, but not only concerned about the cost. That demand gradually happened since last year. That also you can see the result, that progress that our, whatever from our BB ratio and also our revenue growth this year in China are very significantly. Last point is about the China government. They put a lot of the focus and attention to the company, to the service industry, especially from the semiconductor, new energy, AI, and robotics.
Those area are all Advantech focus area. That also the reason that our regional team from the product development, from engineering development, business deployment are very connected with the China local government, their policy, and also their business direction. Yes, we still have the competition every day for the business in China, but we have quite confidence that we can have a good progress to support Advantech China business continue to grow.
Thank you.
Oh, by the way, there is another last topic is about, last question is about China manufacturing. Yes, now we are under planning to expand our China manufacturing capacity in Kunshan, Shanghai. This is under planning. Once we have the proposal ready, that we will elaborate to everyone to understand our plan. Okay.
Thank you, Miller. The next question we had online is, the client would like to know, there is currently some delays in data center construction. The client would like to know whether that actually will impact the revenue recognitions from either, say, in the semiconductor side or within the AI infrastructure that you previously mentioned earlier. Thank you.
This is a good question, but so far, we don't have the message regarding the delays in data center construction, actually. We don't get feedback from the build side that most of the data center construction will be delayed. Allow us to do more investigate, then feedback to the journey. Okay?
Thank you, Eric. Eileen, please go ahead. Thank you.
Okay. Thank you, Shang. Thank you, management, for taking my question again. I just want to ask a little bit long-term view. The strong positive order momentum has been lasting for a while. How do management view how sustainable, how long this positive momentum can sustain? Thank you.
Okay. Regarding for the business momentum forecast, at least 2027 next year, because of the order booking and also our feedback, and also the feedback from our region business team, we are also still very positive for next year. How to sustain the growth momentum for longer term? I would say the first and number one still to keep our core business, our established business and market as our core business continue to grow, like industrial equipment, medical equipment, semiconductor equipment, those our core business, they can contribute our bigger portion of the revenue. We need to continue to expand the business for sure. Second is to gradually incubate this new emerging sector industry, such as the robotics market, new energy, mission critical market, for example. Those new emerging sector and market that can contribute and support Advantech growth continuously.
From the both side, we got to work hard and also recruiting more talent to support Advantech growth continuously from the engineering, from the product development, and also regional business deployment. Those are very important to support Advantech, the growth. I hope that answers your question.
Yes, that's very helpful. Thank you.
Eileen, I would like to add on some information that Advantech company policies that to provide short-term financial guidance, which is we already do, which we will give a quarterly basis financial guidance to the investors. However, internally, I think management team do have long-term growth path, and this is solidly supported, according to Miller, solidly supported by in region perspective, in business perspective. So, in the future, we will evaluate the suitable indicators to share with the investor for the long-term growth. But right now, we still will keep our short-term financial guidance as a policy, but maybe next year, we will provide more color and suitable indicator for the investor to have a more understanding of the long-term growth. Thank you.
Okay. Thank you. That is great. Thank you. That is very clear. Thank you.
Thank you, Eileen. Because of the time constraint, we probably may have another one or two questions. If anyone would want to raise a question to management, could actually click the raise hand button or type your questions within the chat box. Personally, I have two questions. I would like to consult Eric. Regarding your previous comments on the first nine months of the revenue growth, which relative at around 50%, and then 20% are coming from the price hike, and around 30% are coming from the shipment growth. I just want to note about whether the rising pricing due mainly to the high-end AI server that you mentioned earlier, I think the price hike might be like 10 or 20x higher than the original ones. Whether this kind of the contribution has been included into the 20% that you mentioned earlier.
Actually, we have a specific project on hand. The customer belongs to semiconductors. The high-end AI servers, ASPs, it is more than 80K to 100K, but this is a single K.
Oh, 100K.
We will watch this trend carefully and feedback to our investor if we have any insight or any indicators.
Thank you. My second questions will be on the target about the HAI contributions. I think earlier you mentioned about 25% in the first nine months, and that we now know that 30% within this year. I wonder if there is actually a target that we are going to see in 2027.
We will try our best to meet this target, let's call 30%. But right now, up to the third quarters, the figure is 24.5%. There is still a little bit to get to achieve this goal.
Thank you. I think that will be all the question I have on hand. Also due to the time constraint, I am thinking about if the management would have any kind of the closing remark before we conclude the call.
Yeah. Given the strong BB ratio and design-win indicator, we foresee year 2027 might have a great chance to secure high two-digit growth in both our top line, also the bottom line. Besides, we will have a formal announcement regarding our founder CEO position transitions planned by tomorrow. So please be patient with the message, and Michael will deliver message to all investor. Thank you.
Thank you, Eric Chen. Then I think we will conclude the call here, and thanks for the management to attending this meetings and also all the online participants. We will conclude the call here and see you guys next time. Thank you.
Thank you.