貿聯-KY 3665 — 2026-08-21 法人說明會逐字稿

日期：2026-08-21｜時長：1:05:00｜出席主管：Roger Liang（董事長）、Felix Teng（執行長）、Charles Tai（財務長）、Mike Wang（資深投資人關係經理）｜主辦：UBS（分析師 Ali Chen）

逐字稿為線上法說音檔 AI 轉寫後人工校正，數字以財報與簡報為準。本場以英文進行，逐字稿保留原文，重點摘要為中文整理。

重點摘要

財務（2Q26，CFO Charles Tai）

• 合併營收 NT$232.8 億（YoY +37%、QoQ +12%）。※ 轉寫原文為 33.28 billion，經毛利率／營益率反推應為 23.28 billion，見「待校正紀錄」。
• 毛利 NT$71.5 億，毛利率 30.73%（1Q26 為 28.77%）。
• 營業利益 NT$40.1 億，營益率 17%（1Q26 為 14.9%）。
• 歸屬母公司稅後淨利 NT$29.8 億、EPS 15.28 元，雙雙創單季新高。
• 1Q26 毛利率受產品轉換、客戶部署時程、匯率、業務組合多重因素影響，2Q26 部分因素已正常化；2Q26 改善不只來自量增，也來自製造效率、材料成本、生產力與費用效率的實質改善。
• 明確提醒不要把毛利率模型畫成逐季直線上升：產品組合、客戶組合、稼動率、匯率、專案時程都會造成季間波動；公司目標是「在放大營收基礎的同時維持健康且可持續的毛利率」，而非追求單季毛利率最大化。

訂單能見度與成長動能

• 半導體設備（capital equipment）已成為另一個結構性成長引擎：角色由零組件擴展到高階組件與系統整合，工程參與更深；季度銷售創新高。目前訂單能見度合理延伸至 2027 年下半年，與客戶的討論已開始延伸到 2028 年。
• AI 基礎設施仍是最強的結構性動能之一；公司同時參與多個技術世代，工程團隊已在與客戶做下一代之後（N+2）的方案。
• 較廣泛的工業業務出現逐步復甦，各終端市場速度不一但方向轉好。

策略定位（CEO Felix Teng）

• 核心命題已從「賣連接器產品」轉為「客戶要我們解什麼工程問題、我們有沒有能力解」。能力範圍由傳統互連擴展到電力傳輸、高速資料連接、光連接、工程、NPI、先進製造、系統整合。
• Deployment first（部署優先）：算力要能部署才產生經濟價值，而部署取決於電力、網路、冷卻、機構、廠務與公用設施是否同步到位。瓶頸會在不同時點於算力／網路／電力／設備／廠務之間移動，因此部署不會是線性的，即使底層需求依然強勁，採購與部署時程仍會出現變動——「底層需求與部署時點之間的落差正在擴大」。
• 電力架構不押單一路線：機櫃層級業界同時在評估更高電流與更高電壓等多種方案；公司認為不會有單一解適用所有客戶與資料中心，轉換期會是多種電力架構並行。貿聯的電力能力橫跨線材、匯流排（busbar）與更高功率方案。
• 銅與光都做，不視為單純的銅轉光：不同系統架構會要求不同的銅／光組合；公司在銅高速互連的地位隨更高速架構持續演進，同時（透過 XFS，※名稱待核對）大幅擴充光學能力。策略重點是「不論內容在技術之間如何移轉，都要參與這個轉換，而不是押注單一技術勝出」。
• 產能與資本支出紀律：不會因為需求預期高就蓋產能；擴產必須有客戶需求與預期報酬支撐。也會主動重新配置既有產能——把成熟業務的廠房、設備與組織資源移轉到成長性、策略價值或報酬更好的機會，必要時寧可讓成熟業務縮減，也不是單純為了保住營收而加產能。
• 併購以「能力補強」為主軸：評估標的看能力、技術、客戶關係、地理覆蓋等長期戰略價值，而非當下的營收貢獻；偏好能跨多個產品世代／客戶／終端市場保持價值的能力。

併購：Interplex Datacom

• 規劃中的 Interplex Datacom 收購案，補的是機械工程、精密製造與設計能力，與貿聯既有的電氣、光學、電力與製造工程優勢互補，特別是在機械與電氣需求交會之處。
• 本案尚未交割，Interplex Datacom 仍為獨立事業，未併入本季財報；交割前公司對細節保持保留。

資本結構與募資（CFO）

• 近期完成募資。公司表示原本有更依賴舉債的選項，但選擇以股權及股權連結工具支持成長，以維持穩健資產負債表——理由是客戶專案規模與投資機會持續變大。
• 公司對槓桿與流動性有內部財務框架（guardrails）；資產負債表可以暫時偏離偏好位置，但一旦超出護欄就會主動調整回來，「這個紀律不隨市場情緒改變」。
• 財務餘裕的目的不是花更多錢，而是確保資金可得性不會在對的機會出現時成為限制；更大的財務量能不會降低報酬門檻（hurdle rate）。
• 資本配置優先序：先投既有業務（營運資金、工程、技術、產能、自動化、NPI）；加產能前先評估既有資源能否更有效運用。ROE 與 ROIC 過去數年已顯著改善。

競爭（董事長 Roger Liang）

• 觀察到更多競爭者擴展資料連接、電力與架構產品線。貿聯的策略不是為競價而競價，而是在效能要求高、認證困難、客戶重視可靠度與上市時程的領域建立更深的位置。
• 高速銅纜是最好的例子：多個世代的開發與認證累積了工程 know-how、製造經驗與客戶關係。「這些市場不會每一個世代重新洗牌」——既有供應商帶著認證歷史、客戶關係與製造經驗進入下一個架構，後進者必須在技術持續前進的同時追趕。
• AI 基礎設施複雜度提高（更高資料速率、更高功率密度、液冷）要求電氣、機械、熱、製造能力協同運作，使既有地位更有價值。必要時以針對性併購補足能力，比全部自建更快。

光學業務（CEO）

• 光學視為資料連接業務的結構性延伸；機會不只是 800G → 1.6T 的速率轉換，密度同樣重要。
• 新建與既有基礎設施雙軌機會：新的 AI 基礎設施從一開始就會設計更高的光學含量；既有超大規模基礎設施也會隨網路架構改變而升級或演進。
• 公司已建立光纖組件（fiber assemblies）、線束（harnesses） 等相關光連接解決方案平台；策略是不預測哪一種光學架構勝出、何時轉換，因為「幾乎每一條通往更大更密 AI 基礎設施的路徑，都需要更複雜的光連接」。

產能與資本密度（CFO，Q&A 最後一題）

• 會持續投資，但不認為每一塊錢的新增需求都該自動對應一塊錢的新增產能；評估優先序、策略價值、報酬要求、機會的持續性，以及既有資源能否先被更有效使用。
• 自動化與生產力提升可以在不增加同等實體產能的情況下增加產出。
• 投資範圍已不只是廠房與設備，還包括工程、技術、認證自動化、製造能力、人才、系統與組織基礎設施。

逐字稿

Presentation

Ali Chen - UBS, Analyst (00:00:02) OK, good afternoon and good morning. Welcome to join BizLink's second quarter earnings call hosted by UBS. I am Ali Chen covering BizLink and the industrial sector in Taiwan. It is our honor to host BizLink management today. Now, let me hand over the call to Mike Wang, the senior IR manager. Mike, please.

Mike Wang - BizLink, Investor Relations (00:00:24) Thank you for the intro, and once again, thank you to UBS for hosting our results call. Good afternoon, everyone, and welcome to this week's second quarter 2026 earnings conference call. My name is Mike Wang, senior IR manager. Joining me today are Roger Liang, chairman; Felix Teng, CEO; and Charles Tai, CFO. Our earnings results were released earlier today and are available on our IR website, where you can download the latest earnings materials and access the call through MOPS.

Mike Wang (00:00:53) Today's call will begin with Felix, who will share strategic updates. Charles will then conclude with our financial highlights before we move to the Q&A session. You may submit your questions at any time through the public or private chat function, and we will address as many as time permits. Before we begin, please note that today's discussion may contain forward-looking statements based on our current expectations and are subject to risks and uncertainties.

Mike Wang (00:01:18) Actual results may differ materially. Please refer to the safe harbor notice in our earnings materials for further details. This call is being recorded and will be available on our IR website within 24 hours. With that, I would now like to turn the call over to Felix.

Felix Teng - BizLink, CEO (00:01:35) All right, thank you, Mike. Good afternoon, everyone, and thank you for joining us. Before Charles discusses our second quarter financial performance, I would like to spend some time on how we see our markets evolving, what is changing in our customers' requirements, and how BizLink is positioning itself for these changes. Over the past several years, BizLink has changed significantly.

Felix Teng (00:02:03) Connectors, cables and interconnect products remain important parts of our business, but they no longer fully describe where our business is heading. The more important question today is what engineering problems our customers need us to solve, and whether we have the capabilities to solve them. AI infrastructure is entering a new phase. The first phase of this investment cycle was heavily focused on securing compute. That remains important, but the challenge is increasingly shifting toward deploying that compute at scale. As compute density rises, the infrastructure surrounding it has to keep pace. More power has to be delivered safely and efficiently. More data have to move at higher speeds.

Felix Teng (00:02:59) Thermal and mechanical requirements are becoming more demanding, and increasingly these systems need to be designed, manufactured, and deployed together. This is why we continue to emphasize deployment first. The demand for compute is substantial, but compute only creates economic value after it is deployed and operating. The ability to deploy therefore depends not only on GPUs or accelerators, but on the availability and readiness of power, networking, cooling, mechanical infrastructures, facilities, and utilities. As AI factories become larger, synchronization across these infrastructure layers becomes increasingly important. This also means deployment will not always be linear.

Felix Teng (00:03:56) At different points in the cycle, the constraints can shift from compute to networking, from networking to power, or from equipment availability to facilities and utilities. This can create variability in procurement and deployment schedules, even when the underlying requirements for this infrastructure remain strong. The gap between underlying demand and deployment timing is increasingly important. For BizLink, it also expands the opportunities. As performance requirements increase and systems become more complex, customers need suppliers that can address a broader range of engineering and manufacturing requirements. We have been building BizLink around that direction for several years.

Felix Teng (00:04:54) On how we frame the business, we do not view our strategy as building a collection of individual products. We view it as expanding the capability we can bring to customers. What began with traditional interconnect solutions has evolved to include power delivery, high-speed data connectivity, optical connectivity, engineering, new product introductions, advanced manufacturing, and increasingly broader system-level integration. The value of these capabilities increases when they work together. Engineering allows us to engage earlier

Felix Teng (00:05:39) in a customer's deployment process. NPI converts engineering solutions into products that can be qualified and manufactured reliably. Manufacturing allows those products to scale, and system integration allows us to address a larger part of the customer's requirement. This is the direction in which we have been moving. We are already seeing this evolution across several businesses. In capital equipment, our scope has expanded from components toward higher-level assemblies and system integrations. Our engineering involvement is also increasing as customers engage us earlier in the development process. In HPC, we have expanded from traditional connectivity

Felix Teng (00:06:41) into power solutions. Across the company, the common denominator is that our customers are asking us to do more. This is important strategically. The more capabilities we have, the more problems we can potentially solve. Earlier participation gives us a better understanding of the customer's system requirements, and when we execute successfully, that creates opportunities to participate in more programs and more content over time. It also makes our business less dependent on the success of any single product or architecture. Individual products will change. Technology architectures will change. Customer requirements will change. Our objective is to build capabilities that

Felix Teng (00:07:41) support those changes. This is one reason we have continued to expand both the depth of our products and the breadth of our capabilities. Greater product depth allows us to participate in more parts of a customer's system. Broader capabilities allow us to respond as the technology changes. We do not need to predict every technology transition correctly. We need the engineering and manufacturing capabilities to remain relevant as those transitions occur. This is also why NPI has become increasingly important to BizLink. Engineering creates an opportunity, but engineering alone does not create revenue. The solutions still have to be validated, qualified,

Felix Teng (00:08:39) industrialized, and produced reliably at scale. NPI connects those stages. As our NPI capabilities strengthen, we can work with customers earlier, improve manufacturability, shorten the transition into volume production, and build the operating knowledge required to support a program over its life cycle. This makes the relationship deeper than supplying an individual component. And as the system becomes more complex, customers increasingly value suppliers that can take on a broader part of that responsibility. This does not mean one supplier needs to do everything. It means the value of having suppliers with broader engineering,

Felix Teng (00:09:38) manufacturing and integration capabilities is increasing. That is where BizLink wants to compete. So what is changing? AI infrastructure provides the clearest example of the evolution today. Power requirements are rising rapidly as compute density increases. This is changing how power needs to be generated, transmitted, and delivered all the way to the rack. At the rack level, the industry is evaluating multiple approaches, including higher amperage and higher voltage. We do not believe there will be one solution for every customer or every data center. Existing infrastructure matters, deployment schedules matter,

Felix Teng (00:10:35) reliability matters. The transition will therefore involve multiple power architectures operating at the same time. For BizLink, this is important because we are not dependent on one architecture winning. Our power capabilities span cables, busbars, and increasingly higher power solutions. As customer requirements evolve, our objective is to support those requirements across different architectures. The direction is clear: more compute requires more power, and higher power requirements create more engineering challenges throughout the infrastructure.

Felix Teng (00:11:31) The growth in AI-related electricity demand is accelerating investments across power generation, grid infrastructure, and electrical equipment. As those systems become increasingly electrified, capabilities developed around power delivery, power management, and higher density electrical systems can also become relevant to industrial automation, capital equipment, and other end markets over time. Data connectivity is evolving in parallel. As compute clusters become larger, more data need to move within and between those systems at higher speed. Copper continues to play an important role,

Felix Teng (00:12:24) particularly where distance and system architecture allow it. At the same time, optical connectivity becomes increasingly important as bandwidth requirements and transmission distances increase. We have capabilities in both. Our position in copper connectivity continues to evolve with higher speed architectures. While XFS has materially expanded our optical capabilities, we do not see this as a simple transition from copper to optics. Different system architectures will require different combinations of both. And those requirements will continue to evolve. For BizLink, the strategic point is straightforward.

Felix Teng (00:13:18) We can participate across power, copper, and optics. This gives us more than additional content opportunities. It gives us greater flexibility as customer architectures evolve. If content shifts between technologies, our objective is to participate in the transition rather than depend on one technology outcome. As these infrastructure layers become more closely connected, that breadth becomes increasingly valuable. And importantly, we do not expect the industry to change every part of the infrastructure at the same time. Customers need to deploy. New technologies therefore have to coexist with existing infrastructure, and architectural transitions

Felix Teng (00:14:18) depend on the customers and applications. This reinforces our view of deployment first. The winning solution is not necessarily the newest architecture in isolation. It is a solution that allows customers to bring reliable compute capacity online, at the scale and timing they require. Our execution strategy follows directly from these changes. If customers' requirements are becoming broader, our capabilities need to become broader as well. We continue to invest organically in engineering and NPI. We are expanding manufacturing capacity and capabilities where customer demand supports those investments.

Felix Teng (00:15:17) We are increasing the level of system integration we can provide, and we continue to evaluate strategic investments where acquiring or establishing capability can accelerate our progress. Our global footprint is an important part of this strategy. Customers increasingly require suppliers that can support them across multiple regions while maintaining consistent engineering quality and manufacturing standards. We therefore continue to invest in our manufacturing network, where we see sustained customer requirements. But strong demand does not mean we need to pursue every available opportunity. We remain selective in how we allocate capacity and capital. We prioritize

Felix Teng (00:16:14) programs where our capabilities, customer relationships, and expected returns justify the resources required. This is particularly important in the current environment. We do not intend to build capacity simply because demand expectations are high. Capacity expansion needs to be supported by attractive long-term economics. We also actively manage how existing capacity is used. As our business mix evolves, we can reallocate manufacturing space, equipment, and organizational resources toward opportunities where we see stronger growth, greater strategic value, or better returns. In some cases, this may mean

Felix Teng (00:17:16) allowing a business to run down rather than adding capacity simply to preserve revenue. Our objective is therefore not to maximize revenue at any cost. It is to allocate our resources toward the opportunities where BizLink can create the greatest long-term value. This discipline is important because capacity decisions made during periods of very strong demand can affect returns for many years. We want to participate fully in the growth opportunities ahead of us without building the company around the assumption that every demand signal will continue indefinitely. At the same time, we are seeing increasing opportunities to participate earlier in customer projects.

Felix Teng (00:18:14) Earlier engagement allows us to contribute more engineering value and gives us a better understanding of how the complete system is evolving. The progression we are pursuing is clear: earlier engineering engagement, stronger NPI, scalable manufacturing, and broader system integration. This is also how we approach M&A. We evaluate acquisitions based on more than their immediate revenue contribution. We look for capabilities, technologies, customer relationships, geographic reach, and other strategic strengths that can make BizLink more competitive over time. Increasingly, we also think about M&A through the durability of the capabilities we acquire.

Felix Teng (00:19:13) Individual products can have relatively short technology cycles. Strong engineering, design or manufacturing capabilities can potentially remain relevant across multiple product generations, customers and end markets. That matters to how we allocate capital. We prefer strategic assets that strengthen what BizLink is capable of doing, rather than simply adding exposure to a product that is attractive at one point in the cycle. The contemplated acquisition of Interplex Datacom fits our broader capability expansion framework.

Felix Teng (00:20:02) It should not be viewed as a change in strategic direction, but rather as another step along a path we have been following for many years. Interplex Datacom's capabilities in mechanical engineering, precision manufacturing and design would complement BizLink's existing strengths across electrical, optical, power and manufacturing engineering. Following closing, we believe these complementary capabilities will broaden the engineering and manufacturing capabilities available across the business platform, particularly where mechanical and electrical requirements increasingly intersect.

Felix Teng (00:20:57) Because the transaction has not yet closed, Interplex Datacom remains a separate business, and we will therefore remain disciplined about what we discuss before closing. Our broader acquisition philosophy remains unchanged. We will continue to develop capabilities organically where that is the best approach. We will consider acquisitions when they can accelerate our strategy and we will remain selective. The objective is not simply to make BizLink larger. The objective is to make BizLink more capable. So what does this mean going forward? Looking forward, we see a larger opportunity set for BizLink than we had several years ago.

Felix Teng (00:21:51) AI infrastructure is creating new requirements across power, data connectivity, thermal and mechanical systems. Semiconductor production equipment is becoming more complex and creating opportunities for greater engineering and system integration content. Industrial automation, transportation, healthcare and other markets continue to evolve toward higher levels of electrification and automation. These markets will not grow at the same rate every quarter. We also do not expect the current AI investment cycle to move in a straight line. The scale and pace of investment across AI infrastructure today are significant.

Felix Teng (00:22:46) We remain constructive on the structural opportunity. However, an investment cycle of this magnitude will include periods of acceleration and periods of digestion. Our responsibility is not to predict every turn of that cycle. It is to ensure that the business is positioned to capture the upside while remaining resilient if conditions change. Our strategy is increasingly designed around that flexibility. We are building depth across products so that we can participate as customer requirements move between technologies. We are building broader capabilities that can remain relevant across multiple products and markets. We are selective about the demand and capacity we pursue.

Felix Teng (00:23:43) And we continue to diversify the earnings base of the company. We do not need perfect visibility into every technology transition or every point in the investment cycle. We need the ability to adapt as those conditions change. We are not building our strategy on the assumption that today's rate of investment will continue indefinitely. We are building the company so that we can capture the structural opportunities through different parts of the cycle. Across the board, our strategy remains consistent. We want to participate earlier. We want to solve more complex problems. We want to increase our content and value to customers. And we want to convert those capabilities

Felix Teng (00:24:39) into sustainable long-term growth. Our opportunity is no longer defined by any single product. It is increasingly defined by the range of problems BizLink is capable of solving. With that, I will turn the call over to Charles.

Charles Tai - BizLink, CFO (00:24:59) Thank you, Felix, and good afternoon, everyone. Felix has discussed how our opportunity set and capabilities are evolving. I will focus on what that evolution means financially: our second quarter performance, the economics of the business as we scale, the breadth of our growth drivers, and how we're managing our capital structure following the recent fundraising.

Charles Tai (00:25:28) So for the second quarter of 2026, consolidated revenue was 23.28 billion NT dollars [轉寫原文為 33.28 billion，見待校正紀錄], representing a 37 percent year-on-year growth and 12 percent sequential growth. Gross profit was 7.15 billion NT dollars with a gross margin of 30.73 percent compared with 28.77 percent in the first quarter. Operating profit was 4.01 billion NT dollars with an operating margin of 17 percent compared with 14.9 percent in the first quarter. Net income attributable to shareholders was 2.98 billion, and EPS was 15.28 NT dollars.

Charles Tai (00:26:22) Net income and EPS reached a new quarterly high. The second quarter showed a clear improvement from the first quarter. Revenue increased sequentially, utilization improved, and the operational issues that affected our first quarter profitability began to normalize. Last quarter, we explained that gross margin was affected by several factors occurring at the same time, including product transitions, customer deployment schedules, foreign exchange, and business mix. During the second quarter, we focused on the areas that we can control. The first is cost efficiency.

Charles Tai (00:27:09) As volume improved and utilization increased over time, we continued working on manufacturing efficiency, material cost, productivity, and execution across our operations. The second is operating expense efficiency. We continue to invest where necessary to support growth, but the organization also needs to become more efficient as the revenue base expands.

Charles Tai (00:27:40) Importantly, the improvement in the second quarter was not simply a function of higher volume; we also made progress in addressing the operational issues that affected the first quarter, while continuing to improve manufacturing efficiency, material cost, productivity, and operating efficiency across the organization. The results this quarter showed progress in both areas.

Charles Tai (00:28:06) We should also be clear about how we manage the business. Our objective is not to maximize gross margin in a particular quarter. Our business model is to deliver strong scale growth with healthy and sustainable margins. Margins will move from quarter to quarter because of product mix, customer mix, utilization, foreign exchange, and program timing. We expect that. What matters is whether we can maintain reasonable profitability while expanding the absolute earnings and cash generation capacity of the company. The second quarter is a strong demonstration of that model. So, BizLink today is materially larger than it was several years ago. For several years,

Charles Tai (00:29:02) we focused on improving the quality of the business. We improved gross margin. We improved operating efficiency. We maintained discipline around operating expenses. We improved cash generation. And we became more selective about where we deploy capital. Those priorities remain. But the next stage of BizLink's development is not about maximizing any one financial ratio. It's about maintaining healthy economics while continuing to scale. A healthy margin on a larger revenue base generates greater operating profit, greater net income, and greater cash flow. That increases earnings capacity.

Charles Tai (00:29:53) It gives us more resources to invest in customer programs, technology, capacity, people, and strategic opportunities. The important point is that scale and profitability have to work together. Growth without adequate return does not create sustainable value. But maximizing margin by walking away from attractive growth opportunities is also not how we intend to manage the company. Our objective is to capture growth where we have a competitive advantage while maintaining discipline, profitability, and returns.

Charles Tai (00:30:38) This is increasingly relevant because the composition of BizLink's business is changing. Our growth businesses are becoming larger. Our engineering content is increasing. Our system integration capabilities are expanding. And in several businesses, we are moving closer to the customer and participating in more of the value chain. This creates opportunity to increase both the size and the quality of our earnings base.

Charles Tai (00:31:08) We're already seeing the financial impact of that evolution. We can see this change in how we engage with customers. Across several of our higher growth businesses, our participation is moving beyond individual components toward broader assembly, subsystem, and engineering solutions. We're also becoming involved earlier in customer development cycles and supporting more complex programs. This increases the value that we can provide to customers and expands the opportunity available to BizLink with each program. So what is changing? The sources of growth across BizLink are becoming broader.

Charles Tai (00:31:58) AI infrastructure remains one of our strongest structural growth drivers. The underlying demand for compute remains strong, and as Felix discussed, the infrastructure required to deploy that compute is expanding across power, connectivity and other areas. Another important characteristic of this business is that we are participating across multiple technology generations at the same time. While today's generation is already contributing to our revenue and profit growth, our engineering teams are working with customers on future platforms. In some areas, that work already extends beyond the next generation into N plus two solutions.

Charles Tai (00:32:55) This matters because each generation is becoming more demanding: power density is increasing, data speeds are rising, and system requirements are becoming more integrated. Early participation allows us to accumulate engineering knowledge, qualification experience, and customer trust that can carry forward into subsequent generations.

Charles Tai (00:33:24) As a result, the capabilities supporting today's earnings also help position us for future opportunities. At the same time, we do not assume that the current pace of AI infrastructure investment will continue in a straight line. Large investment cycles inevitably include changes in deployment timing, capacity digestion, and shifts in where capital is being directed. We do not need to predict exactly

Charles Tai (00:33:53) when those changes will occur. We need to manage BizLink so that we can continue creating value through them. That starts with the quality of the demand we choose to support. We do not need to address every unit of incoming demand. When demand is strong, it can be tempting to add capacity simply to capture additional revenue. We do not believe that is always the right decision.

Charles Tai (00:34:21) We evaluate the capital required, the expected return, the strategic importance of the program, and whether existing resources can be redeployed before committing incremental capacity. This allows us to participate in attractive growth without automatically converting every increase in demand into higher capital intensity over time. We intend to keep capital expenditure discipline relative to the scale of the business and broadly consistent with the way we have historically managed capital intensity. There may be periods where strategic investments justify higher spending, but the principle is clear. We will not build capacity simply because demand and expectations are high.

Charles Tai (00:35:15) Capital has to be supported by customer requirements and expected returns. Product and capability depth provide another form of resilience, as Felix discussed. Broader participation across technologies gives us greater flexibility as architectures evolve. From a financial perspective, that matters because our growth does not have to depend on one product or one technology outcome. The same principle applies to strategic investment. When we invest in capability that can support multiple products, customers, or end markets, we create more ways to generate return from that investment over time.

Charles Tai (00:36:04) We believe this is a more durable approach than concentrating capital around a narrow product opportunity simply because the demand is strong today. Capital equipment has become another important structural growth engine. AI is increasing manufacturing complexity across advanced logic, memory, and packaging, supporting continued investment in increasingly sophisticated semiconductor equipment. Our role in this market has expanded from components toward high-level assembly and system integration, with deeper engineering engagement and participation in more complex, higher value programs.

Charles Tai (00:36:51) This is already translating into accelerating growth with quarterly sales reaching a new level. Today, we have reasonable order visibility extending into the second half of 2027, while discussions with customers are already beginning to extend into 2028. We are preparing for the next stage of growth by deepening our engineering engagement and expanding our subsystem capabilities. We also see a gradual recovery across our broader industrial businesses. The recovery is not uniform, and different end markets are moving at different speeds, but the direction has improved.

Charles Tai (00:37:41) This gives us multiple sources of growth rather than requiring one business to carry the entire company. Our diversification is therefore not simply defensive. It allows us to participate in multiple structural growth opportunities while giving us greater flexibility in where we allocate capital and capacity. So how are we executing? This brings me to capital allocation and capital structure. Following our recent fundraising, I want to be clear about how we think about both. We manage our balance sheet against a clear internal financial framework. This framework is our north star for how much leverage and liquidity we believe are appropriate for BizLink.

Charles Tai (00:38:35) We do not need the balance sheet to remain at exactly the same position at every point in time. Working capital changes, we invest in capacity, we pursue strategic opportunities, and there may therefore be periods where the balance sheet moves away from our preferred position. But we operate within defined guardrails. If the balance sheet moves outside of those guardrails, we'll take the necessary steps to bring it back to where our preferred position is. That discipline does not change with market sentiment. This is how we intend to manage the company through different parts of the cycle. Our recent financing reflects this philosophy.

Charles Tai (00:39:25) We had alternatives, including relying more heavily on debt financing. We chose not to do that. We believe maintaining a stable balance sheet is the right approach for BizLink, particularly as the size of our customer programs and investment opportunities continue to increase. This is also why we chose to support our growth through a combination of equity and equity-linked capital.

Charles Tai (00:39:52) We believe that when capital is deployed well, it should create long-term value for our customers, our business leaders, and the communities where we operate, as well as our shareholders. Over the years, as we execute our strategy, the value of BizLink has grown substantially, and the market's recognition of that strategy has grown with it.

Charles Tai (00:40:21) Our responsibility is to continue earning that recognition by converting the capital entrusted to us into greater capability, stronger earnings, and sustainable long-term value. Greater financial capacity expands our choices. It does not change our financial discipline or lower our return requirement. Our capital allocation priority remains consistent. We first invest in

Charles Tai (00:40:51) the existing business where customer demand and expected return justify the investment. That includes working capital, engineering, technology, capacity, automation, and NPI. But before adding capacity, we also evaluate whether existing resources can be used more effectively. As our business mix changes, we can shift resources away from more mature programs toward opportunities where we see stronger growth, greater strategic importance, or better economics. This is how we manage capital intensity. We do not measure success by how much capacity we have, but by what the capacity allows the company to earn. Not every incremental dollar of revenue deserves the same amount of capital.

Charles Tai (00:41:50) We will remain selective and prioritize opportunities where customer importance, strategic fit, and expected return justify the resources required. We can see the result of this discipline in our returns as BizLink has grown. Both our return on equity and return on invested capital have improved meaningfully over the past several years. We view this as an important measure of the quality of our growth. Our objective is not simply to make the company larger, but to deploy capital in ways that generate attractive returns over time.

Charles Tai (00:42:32) Following our recent financing, that responsibility becomes even more important. We now have greater financial capacity and our job is to convert that capital into capability and earnings while maintaining disciplined returns. Capital allocation also extends beyond physical capacity. As the business grows, we need more than factories and equipment. We need engineers who can solve increasingly complex customer problems, NPI teams that can convert those solutions into scalable production, leaders capable of managing larger and more geographically diverse operations, and digital systems and processes that allow our teams to operate efficiently across a global company.

Charles Tai (00:43:30) These investments may not produce immediate revenue, but they are necessary to support sustainable earnings growth. We also evaluate strategic opportunities through M&A. Our approach is increasingly capability-led. We look for engineering, design, manufacturing, technology, customer, or geographic capabilities that can strengthen the broader company. We prefer capabilities that can remain valuable across multiple products, customers, and technology cycles, rather than building our strategy around any single product opportunity.

Charles Tai (00:44:15) The contemplated Interplex Datacom acquisition is one example of how we may deploy capital when we identify a strategic opportunity that meets our long-term objectives and return criteria. As the transaction has not yet closed, Interplex Datacom remains a separate business and has not contributed to our reported results. Strategic opportunities do not necessarily appear when internally generated cash is most abundant. This is another reason we value maintaining financial flexibility. As the business itself grows larger, our customer programs are expanding, and the investment required to support those opportunities is also increasing.

Charles Tai (00:45:08) A strong balance sheet gives us the ability to act when an opportunity meets our strategic and return requirements without compromising the operating needs of the rest of the company. That is the purpose of financial capacity — not to spend more, but to ensure that capital availability does not become a constraint when the right opportunity appears. So, looking forward, we believe BizLink is entering the next stage of its development from a much stronger position. The earnings base is larger, our structural growth drivers are broader,

Charles Tai (00:45:57) our product and capability depth are greater, our engineering and manufacturing capabilities are deeper, our customer relationships increasingly span more products, programs and geographies, and our financial capacity has increased. Our responsibility is to convert those advantages into sustainable value. The process is straightforward. We need to convert customer opportunity into earnings, we need to convert earnings into cash, we need to allocate that capital with discipline, and we need to reinvest where we see attractive long-term returns. As we do that, we will maintain several disciplines; we will remain selective

Charles Tai (00:46:46) about the demand we pursue. We will manage capital intensity rather than automatically adding capacity. We will continue reallocating resources as our business evolves. We will favor capabilities that can remain relevant as products and technologies change. And we will manage the balance sheet against the financial guardrails that we have established.

Charles Tai (00:47:15) This discipline becomes more important when markets are strong. The scale of investment occurring across AI infrastructure today is significant, and we remain constructive on the long-term opportunities. But we're not managing BizLink on the assumption that today's demand environment continues indefinitely. There will be periods of acceleration. There will be periods of digestion. Technology will change.

Charles Tai (00:47:44) And there will be changes in where the bottlenecks sit. We do not need to predict each one. We need a business model that can adapt through them. This is why product and capability depth matter. This is why diversification matters. This is why capital discipline matters. And this is why we maintain financial governance. The second quarter gives us greater confidence in this model. We have returned to sequential growth. Profitability has improved from the first quarter, and our absolute earnings continue to demonstrate the benefit of greater scale. Looking further ahead, we remain constructive on AI infrastructure. We remain constructive on capital equipment.

Charles Tai (00:48:35) We're seeing improving conditions across nearly all parts of our industrial portfolio, and we continue to see opportunities to expand our content and capability across the customers and markets we serve. We do not need every business to grow at the same rate every quarter. We need each business to create value over time. And we need to allocate capital toward the opportunities where BizLink has the strongest competitive position and the most attractive returns.

Charles Tai (00:49:09) Our objective is clear: grow the earnings base, maintain healthy profitability, allocate capital selectively, and preserve the strategic and financial flexibility required to keep investing through the cycle. This is how we intend to create sustainable long-term value for our shareholders. Thank you. So now let me turn the call over to Mike.

Q&A

Mike Wang (00:49:36) Thank you, Felix and Charles. This concludes our prepared statements section. Now let us begin the Q&A section. Please type in your questions, and then we will answer as many of them as possible in the remaining time. I want to remind everyone that there will be no forward-looking comments. Looking at some of the questions, I think the one that we want to address first is the one on our competitive position.

Mike Wang (00:50:04) To give you a summary of what's being asked: as competitors broaden their AI infrastructure portfolio, how do you think about BizLink's competitive position and ability to maintain or increase content? So for this one, I'd like to hand it over to Roger.

Roger Liang - BizLink, Chairman (00:50:29) Thank you, Mike. We are seeing more competitors expand across data connectivity, power, and architecture. Our strategy is not to pursue price for its own sake, but to build deeper positions where performance requirements are high, qualification is difficult, and customers value reliability and time to market. High-speed copper is a great example.

Roger Liang (00:51:00) Multiple generations of development and qualification have built engineering know-how, manufacturing experience, and customer relationships. That position has helped us accelerate the capability and system-level knowledge we build in power. We therefore see copper, power and optics increasingly reinforcing one another. Being early matters, because these markets do not reset with each and every generation. An incumbent supplier enters the next architecture with qualification history, customer relationships, and manufacturing experience.

Roger Liang (00:51:54) Later entrants need to catch up while the technology is continuously moving forward. The increasing complexity of AI infrastructure makes these advantages more valuable. Higher data rates, higher power density, and liquid cooling increasingly require electrical, mechanical, thermal, and manufacturing capabilities to work together. Customers need suppliers that can solve engineering problems, qualify solutions, and scale them reliably. We can also use targeted M&A to add complementary capability, when that is faster or more effective than building everything organically. Combined with organic development, this can shorten our strategic timeline.

Roger Liang (00:52:49) So we believe that our advantage can compound over time. Being ahead today can help us enter the next generation earlier, move faster, and potentially widen that lead as AI infrastructure becomes more complex.

Mike Wang (00:53:08) Thank you, Roger. We've seen a lot of interest on optics. So this is the next question: can you update us on the progress of the optics business and how you see the opportunity developing over the next several years? For this one, I'd like to hand it over to Felix.

Felix Teng (00:53:36) OK, thank you, Mike. We see optics as a structural extension of our data connectivity business. As AI clusters become larger and more complex, bandwidth requirements continue to increase. But the opportunity is about more than successive transitions from 800G to 1.6T at naturally higher speed. Density is becoming increasingly important. The opportunity is also broadening as optics move

Felix Teng (00:54:21) [closer to the compute] — which many people have heard about — and these technologies may develop on different timelines but collectively they point toward greater optical complexity. Importantly, we see both greenfield and installed base opportunities. New AI infrastructure will be designed with greater optical content from the beginning. However, the current existing hyperscale infrastructure can be upgraded or evolved with network architecture change. So in this environment we see more and more demand and also inquiries about our solutions.

Felix Teng (00:55:13) We can establish position in fiber assemblies, harnesses, and also related optical connectivity solutions. We are therefore not trying to predict exactly which optical architecture wins or when each transition occurs, but almost every path toward larger and denser AI infrastructure requires more sophisticated optical connectivity, and indeed XFS gave us a meaningful platform from which to participate in that evolution.

Mike Wang (00:55:57) Thank you, Felix. Now, let me turn to some financial questions. Everybody is asking about margin trajectory — the second quarter number that we just reported and of course what we see going forward. So: how should we think about margin trajectory into the second half, particularly given the change in mix and new platform ramps? A perfect question for our CFO Charles.

Charles Tai (00:56:35) OK, thank you Mike and thank you everyone. I think the first point I would like to make is that the second quarter showed the improvement that was expected from the first quarter: revenue increased sequentially, operating conditions improved and some of the factors that affected first quarter profitability began to normalize. We also continue working on the areas we can control, particularly in manufacturing efficiency, material cost, productivity, and operating expense efficiency.

Charles Tai (00:57:10) Looking into the second half, we've been constructive. But I would not suggest modeling gross margin as a straight line upward every quarter, as we discussed before. Product transitions, customer deployment cycles, foreign exchange and business mix can all create quarter to quarter fluctuation. So we're managing the business for stable and reasonable margins while growing the earnings base. I think the more important element is what is happening underneath the margin: the higher growth businesses are becoming larger, so our engineering content is increasing,

Charles Tai (00:57:58) and in several areas, we're moving closer to the customer and participating in more of the value chain. At the same time, some of our more mature businesses are also beginning to recover. So we're increasing both scale and mix working for us. That does not mean that margin will increase indefinitely. What it means is that if we can maintain a healthy margin while growing the revenue base and improving operating efficiency, the absolute earnings and cash generating capacity of the company can continue to increase. That is the financial model that we are trying to build.

Mike Wang (00:58:45) Thank you Charles. We can do one more question. We will address some of the other questions offline. Continuing on spending — capex. With demand strong across HPC and other growth businesses, how are we thinking about capacity expansion and capital intensity over the next one to two years?

Charles Tai (00:59:58) OK, thank you, Mike. The first thing is that we will definitely continue to invest because the opportunity set in front of us is expanding. But our approach to capacity has not changed. We do not believe that every dollar of incoming demand should automatically result in another dollar of capacity. We look at priority, strategic fit, return requirement, the durability of the opportunity, and whether existing resources can first be used more effectively. As our business mix changes, we can

Charles Tai (01:00:37) reallocate resources: capacity supporting more mature programs can sometimes be redirected to faster growing or higher value opportunities. Automation and productivity improvement can also generate additional output without requiring the same level of incremental physical capacity. So capacity is only one part of how we support growth.

Charles Tai (01:01:03) So where do we invest? The requirement is increasingly broader than factory space and equipment. We're investing in engineering, in technology, in qualification automation, in manufacturing capability, and also in people, systems and the organizational infrastructure required to support a larger company.

Charles Tai (01:01:29) We therefore intend to manage capital intensity with discipline and remain mindful of historical experience rather than building capacity on the assumption that demand continues indefinitely. [...] We don't need the balance sheet to remain exactly the same. So if we move away from our preferred position, we would take the necessary steps to bring it back.

Mike Wang (01:02:25) Greater financial capacity also gives us new choices. But it does not lower our hurdle rate. I hope that addresses your question. Thank you, Charles. And finally, thank you, Roger, Felix and Charles. This concludes our Q&A section. A replay of today's call will be available on our IR website within 24 hours from now. If you have any further questions, please feel free to reach out to BizLink Investor Relations. We thank you very much for joining today's call. You may now disconnect.

待校正紀錄

待確認事項

• XFS ＝ 新富生光電（已核對）：CEO 兩度提到「XFS 大幅擴充我們的光學能力」，對照庫內 3665_貿聯-KY（市）「光互連 / CPO / ALC」段落，XFS 即貿聯 2026 年 1 月併表的新富生光電，主營光纖跳線／Shuffle Box 相關，與康寧為主要代工合作關係。
• 2Q26 合併營收精確值：法說口徑 23.28bn vs 7 月自結 23,253mn，差約 27mn，需以正式財報／簡報（https://mopsov.twse.com.tw/nas/STR/366520260821M001.pdf）為準。
• 本場未提供任何 3Q26 或全年財測數字（IR 明確聲明 Q&A 不含前瞻性評論），僅有「半導體設備訂單能見度到 2027 下半年、討論延伸到 2028」這類定性說法。
• 募資金額、發行條件與 Interplex Datacom 交割時程本場均未揭露。