高塔半導體 (TSEM) 2025 Q4 法說會逐字稿

Good day, and thank you for standing by. Welcome to the Tower Semiconductor fourth-quarter 2025 earnings conference call and webcast. (Operator Instructions) Please note that today's conference is being recorded.

您好，感謝您的耐心等待。歡迎參加 Tower Semiconductor 2025 年第四季財報電話會議和網路直播。（操作員說明）請注意，今天的會議正在錄音。

I would now like to hand the conference over to your first speaker, Noit Levy, Investor Relations and Corporate Communications. Please go ahead.

現在我謹將會議交給第一位發言人，投資者關係和企業傳播部的諾伊特‧萊維。請繼續。

Thank you. Good day, and thank you, everyone, for joining us today. Welcome to Tower Semiconductor's fourth quarter and full-year 2025 financial results conference call. With us today are Mr. Russell Ellwanger, our Chief Executive Officer; and Mr. Oren Shirazi, our Chief Financial Officer.

謝謝。大家好，感謝各位今天蒞臨。歡迎參加 Tower Semiconductor 2025 年第四季及全年財務業績電話會議。今天與我們在一起的有我們的執行長 Russell Ellwanger 先生和我們的財務長 Oren Shirazi 先生。

Before we begin, please note that certain statements made during today's call may be forward-looking and subject to risks and uncertainties that could cause actual results to differ materially. These risks are detailed in our SEC filings, Form 20-F and 6-K, as well as filings with the Israeli Securities Authority, all available on our website. Tower assumes no obligation to update any such forward-looking statement.

在開始之前，請注意，今天電話會議中作出的某些陳述可能具有前瞻性，並受到風險和不確定性的影響，這些風險和不確定性可能導致實際結果與預期結果有重大差異。這些風險已在我們的美國證券交易委員會文件（20-F 表格和 6-K 表格）以及以色列證券管理局文件中詳細說明，所有這些文件都可以在我們的網站上找到。Tower公司不承擔更新任何此類前瞻性聲明的義務。

Our fourth quarter and full-year 2025 results are prepared in accordance with US GAAP. Some data presented may include non-GAAP financial measures as defined under SEC Regulation G. Reconciliations to GAAP figures and full explanations are provided in today's press release and financial tables. For your reference, the supporting slide deck is available on our website and integrated into this webcast.

我們 2025 年第四季及全年業績均依照美國通用會計準則編製。部分數據可能包含根據美國證券交易委員會 G 條例定義的非公認會計準則財務指標。與公認會計準則數據的調節表和完整解釋已在今天的新聞稿和財務報表中提供。供您參考，配套幻燈片已發佈在我們的網站上，並整合到本次網路直播中。

With that, I'd like to turn the call over to our CEO, Mr. Russell Ellwanger. Russell?

接下來，我想把電話交給我們的執行長羅素·埃爾旺格先生。拉塞爾？

Thank you, Noit. Hello, everybody. Thank you for joining our call today. Very pleased to share our results for the fourth quarter and full year of 2025. Additionally, we are extremely excited to present how these results have redefined our financial milestones and accelerated the timeline for achievement of the same. The updated financial model, which we will present, is the result of already strong partnerships with our lead customers having grown into deeply trust rooted supplier customer partnership technical alliances.

謝謝你，Noit。大家好。感謝您今天參加我們的電話會議。非常高興與大家分享我們2025年第四季和全年的業績。此外，我們非常高興地向大家展示這些成果如何重新定義了我們的財務里程碑，並加快了實現這些里程碑的時間表。我們將展示的更新後的財務模型，是我們與主要客戶之間已經建立的牢固夥伴關係發展成為根植於信任的供應商客戶夥伴關係技術聯盟的結果。

We entered our fourth quarter of 2025 with a company revenue of $440 million and 11% quarter-over-quarter growth, 14% year-over-year growth, fulfilling our beginning of the year target of quarterly sequential growth. In addition to the top line, we achieved bottom-line growth throughout the year. Fourth-quarter net profit was $80 million or 18% net margin, up from 11% in Q1 '25, 13% in Q2 '25, and 14% in Q3 indicative of a value-based growth being driven by technology mix enrichment. The revenue growth from Q1 to Q4 of 2025 was $82 million, of which there was a $40 million net profit drop down and almost 50%, to be exact 48.78%, and this was due to the high value of the incremental photonics revenue.

進入 2025 年第四季，公司營收達到 4.4 億美元，季增 11%，年增 14%，實現了年初設定的季度環比成長目標。除了營收成長外，我們全年也實現了利潤成長。第四季淨利潤為 8,000 萬美元，淨利潤率為 18%，高於 2025 年第一季的 11%、2025 年第二季的 13% 和第三季的 14%，這表明以技術組合豐富為驅動的價值型成長。2025 年第一季至第四季的營收成長為 8,200 萬美元，其中淨利潤下降了 4,000 萬美元，降幅接近 50%，確切地說是 48.78%，這是由於光子學增量收入價值較高所致。

Revenue for the full year was $1.566 billion, $130 million or 9% increase as compared to 2024 revenue. Now to review our 2025 revenue breakdown and discuss the key trends, please see slides 5 and 6 as reference. We achieved year-over-year growth across our key technology platforms namely Power Management, Image Sensors, and 300-millimeter RF SOI, on top of which, record achievements and unprecedented growth of our market leading optical transceiver offerings, Silicon Germanium and SiPho advanced platforms, has propelled us into a favored and unique position, both driving our growth for 2026 and additionally giving us the ability to redefine our financial model, which I will present at the end of my comments.

全年營收為 15.66 億美元，比 2024 年營收成長 1.3 億美元，增幅達 9%。現在，為了回顧我們 2025 年的收入組成並討論主要趨勢，請參閱投影片 5 和 6 作為參考。我們在關鍵技術平台（即電源管理、影像感測器和 300 毫米射頻 SOI）上實現了同比增長。此外，我們市場領先的光收發器產品（矽鍺和 SiPho 先進平台）也取得了創紀錄的成就和前所未有的增長，這使我們處於有利且獨特的地位，不僅推動了我們 2026 年的增長，而且還使我們能夠重新定義我們的財務模式，我將在發言的最後介紹這一點。

RF Infrastructure showed a 75% revenue increase, 2025 over 2024, being our fastest growing application in '25 driven by hyperscaler rapid adoption of Silicon Photonics in 800G and 1.6T pluggable transceivers. Silicon Germanium and Silicon Photonics revenues represented 27% of our corporate revenues or $421 million, up from $241 million or 17% in 2024. SiPho revenues alone were $228 million in 2025 up from $106 million in 2024.

射頻基礎設施在 2025 年的營收比 2024 年成長了 75%，成為我們在 2025 年成長最快的應用，這主要得益於超大規模資料中心快速採用 800G 和 1.6T 可插拔收發器中的矽光子技術。矽鍺和矽光子學收入占公司總收入的 27%，即 4.21 億美元，高於 2024 年的 2.41 億美元（佔 17%）。光是 SiPho 一家公司的營收就從 2024 年的 1.06 億美元成長到 2025 年的 2.28 億美元。

Specific to the fourth quarter, RF Infrastructure revenues were 32% of corporate revenue with SiPho having achieved $95 million or a $380 million annual run rate. Included in this number is some non-wafer NRE to enhance future development for Gen+1 and Gen+2.

具體到第四季度，射頻基礎設施收入占公司收入的 32%，其中 SiPho 實現了 9,500 萬美元的收入，年化收入達 3.8 億美元。該數字中包含一些非晶圓非重複性工程費用，以促進 Gen+1 和 Gen+2 的未來發展。

As highlighted in our recent announcement with Nvidia, the insatiable demand for compute bandwidth in both scale-up and scale-out architectures and Tower's exceptional ability to scale the capacity flawlessly in partnership with our customers has made 1.6 terabyte per second the fastest growing Silicon Photonics node in the industry to date with Tower being by far the majority supplier of 1.6T Silicon PICs. The partnership announced with Nvidia, as with all our direct module customers, underscores our commitment to deliver best-in-class technology and the manufacturing agility required to meet such an exceptional demand trajectory. In addition to Fab 3 Newport Beach, this past year, we successfully ramped Silicon Photonics production in Fab 9 San Antonio, Fab 7 Uozu, Japan, and are on track to ship the first production of very large SiPho ramp in 2026 from Fab 2 Migdal Haemek. Given an even stronger customer demand than was known at our last quarterly release, we have increased our CapEx plan for 2026 with multiple customer requests to enter into capacity reservation agreements through 2028, enabling our customers to in turn give firm commitments to their customers having ensured their supply.

正如我們最近與英偉達的公告中所強調的那樣，無論是縱向擴展還是橫向擴展架構，對計算頻寬的巨大需求，以及 Tower 與客戶合作完美擴展容量的卓越能力，使得每秒 1.6T 字節成為迄今為止業界增長最快的矽光子節點，而 Tower 則是迄今為止 1.6T 太矽光子 (PIC) 太矽光子 (PIC) 的主要供應商。與英偉達的合作，以及我們與所有直接模組客戶的合作，都凸顯了我們致力於提供一流技術和製造靈活性的承諾，以滿足如此巨大的需求成長。除了 Fab 3 Newport Beach 之外，去年我們還成功地在 Fab 9 San Antonio 和 Fab 7 Uozu, Japan 提高了矽光子裝置的產量，並且預計在 2026 年從 Fab 2 Migdal Haemek 開始交付第一批超大規模的矽光子裝置。鑑於客戶需求比我們上次季度報告中預測的還要強勁，我們增加了 2026 年的資本支出計劃，並且收到多位客戶的請求，希望簽訂到 2028 年的產能預訂協議，這使得我們的客戶能夠反過來向他們的客戶做出明確的承諾，從而確保他們的供應。

For next generation 400 gigabit per lane, we continue to make strong progress with heterogeneously integrated indium phosphide on silicon and other material systems. We are playing a key role partnering with our elite customers to define material systems that will be chosen, refining the flow and hence ensuring manufacturability readiness and immediate ramp capability upon 3.2T market introduction. We also see co-packaged optics as a substantially incremental opportunity for us in the coming years as optics gets adopted in scale-up interconnects as well as XPU, the high bandwidth memory interconnects, that are today largely copper.

對於下一代每通道 400 千兆位元的傳輸，我們在矽上異質整合磷化銦和其他材料系統方面繼續取得顯著進展。我們與我們的精英客戶合作，發揮關鍵作用，確定將要選擇的材料系統，改進流程，從而確保在 3.2 噸級產品上市時具備可製造性準備和立即提高產能的能力。我們也認為，隨著光元件在規模化互連以及目前主要採用銅的高頻寬記憶體互連 XPU 中得到應用，共封裝光元件在未來幾年將為我們帶來巨大的增量機會。

In Q4, we announced the expansion of our mature 300-millimeter wafer bonding technology to enable wafer-to-wafer integration of Silicon Photonics ICs and Silicon Germanium electrical ICs. In addition, we continue to work with several customers on dense wavelength division multiplexing laser sources which are a critical component of many CPO implementations and can significantly expand our served optical market by now including the laser source.

第四季度，我們宣布擴展我們成熟的 300 毫米晶圓鍵合技術，以實現矽光子積體電路和矽鍺電子積體電路的晶圓對晶圓整合。此外，我們繼續與幾位客戶合作開發密集波分複用雷射源，這是許多 CPO 實現的關鍵組件，現在將雷射源納入其中，可以顯著擴大我們服務的光學市場。

Beyond optical transceivers, our Silicon Photonics platform continues to be the technology of choice for physical AI applications, particularly frequency modulated continuous wave LiDAR. Ahead of CES, two of our FMCW LiDAR partners, Aeva and LightIC, publicly announced their collaboration with us and bringing to market disruptive products. The proven robustness of our Silicon Photonics platform supported by many tens of thousands of high-yielding, high-quality wafers shipped to date is enabling Silicon Photonics to capture growing share in the LiDAR market, unlocking new automotive and robotics opportunities.

除了光收發器之外，我們的矽光子平台仍然是物理人工智慧應用的首選技術，特別是調頻連續波雷射雷達。在 CES 展會之前，我們的兩家 FMCW LiDAR 合作夥伴 Aeva 和 LightIC 公開宣布與我們合作，並將顛覆性產品推向市場。我們矽光子平台的穩健性已得到充分驗證，迄今已交付數萬片高良率、高品質晶圓，這使得矽光子技術公司能夠在雷射雷達市場佔據越來越大的份額，從而開啟汽車和機器人領域的新機會。

Our Silicon Germanium platform delivered strong growth year over year in 2025 of 43%, remaining the optimal platform solution for low-power, low-latency, high-performance components such as drivers, transimpedance amplifiers for pluggables, LPOs, and active copper and active optical cables. Alongside our Silicon Photonics production, our Silicon Germanium platform is now running in high volume across Fab 3 Newport Beach, Fab 9 San Antonio, Fab 2 Migdal Haemek, and we have shipped 300-millimeter prototypes from Fab 7 Uozu.

我們的矽鍺平台在 2025 年實現了 43% 的強勁同比增長，仍然是低功耗、低延遲、高性能組件（例如驅動器、可插拔器件的跨阻放大器、LPO 以及有源銅纜和有源光纜）的最佳平台解決方案。除了我們的矽光子裝置生產之外，我們的矽鍺平台目前已在 Fab 3 Newport Beach、Fab 9 San Antonio、Fab 2 Migdal Haemek 大規模運行，並且我們已經從 Fab 7 Uozu 交付了 300 毫米原型。

RF Mobile represented 23% of our 2025 corporate revenue and 24% of our Q4 '25 revenues. 300-millimeter RF SOI was up 5.5% while RF Mobile as a whole was down 15% year over year. Those were primarily due to our proactively working with our customer partners to responsibly reduce exposure to lower margin controller offerings in favor of higher value optical and RF mix in the Fab and also influenced with the front-end module market shift from 200 millimeter to the higher digital content better served with more advanced nodes in 300 millimeter. Our latest technology which we presented last quarter with substantial improvement of our Ron-Coff relative to the competition and reduced layer count, therefore higher overall value per customer dollar, continues to see robust customer adoption. Lead customers have recognized it as best in class and are preparing to ramp to high volumes.

射頻行動業務占我們2025年公司總營收的23%，佔2025年第四季總營收的24%。 300毫米射頻SOI業務成長了5.5%，而射頻行動業務整體較去年同期下降了15%。這主要是因為我們積極與客戶合作夥伴合作，負責任地減少對低利潤控制器產品的投資，轉而支持晶圓廠中更高價值的光學和射頻產品組合；同時也受到前端模組市場從 200 毫米向更高數字內容的轉變的影響，而更先進的 300 毫米節點能更好地滿足這一需求。我們上個季度推出的最新技術，相對於競爭對手，Ron-Coff 有了顯著改進，層數減少，因此每位客戶的每一分錢都物有所值，繼續受到客戶的廣泛歡迎。主要客戶已將其視為同類最佳產品，並準備大幅提高銷售量。

Across the board, we continue to see strong design win momentum that positions our 300-millimeter RF SOI platform for sustained secular growth. In 2025, we achieved major wins, namely three of the top four Tier-1 RF front-end module providers. One has begun production with all planning for strong ramp in 2027 towards achieving appreciable revenue volumes in 2028.

總體而言，我們持續看到強勁的設計中標勢頭，這將使我們的 300 毫米射頻 SOI 平台實現持續的長期成長。2025年，我們取得了重大勝利，即贏得了四大一級射頻前端模組供應商中的三家。其中一家已開始生產，並計劃在 2027 年實現強勁增長，以期在 2028 年實現可觀的收入。

Power Management grew 20% year over year, demonstrating strong year-over-year revenue growth in both 200-millimeter and 300-millimeter offerings, representing 16% of our 2025 corporate revenues and 15% of our Q4 '25 revenues. In 300 millimeter, this includes the ramp of the Tier-1 handset envelope tracker previously announced, which is expected to continue to gain share in the years to come. Overall, our revenue growth in 300-millimeter power has significantly outpaced the rate of growth for both the power market as well as the mobile handset market, demonstrating the strength of our offering and share gains in this significant space.

電源管理業務年增 20%，200 毫米和 300 毫米產品均實現了強勁的同比增長，占我們 2025 年公司總收入的 16% 和 2025 年第四季度總收入的 15%。300 毫米，其中包括先前宣布的 Tier-1 手機信封追蹤器的斜坡，預計在未來幾年將繼續獲得市場份額。總體而言，我們在 300 毫米功率放大器領域的收入成長速度明顯超過了功率放大器市場和行動手機市場的成長速度，這表明我們產品的優勢以及我們在這一重要領域的市場份額成長。

Sensors and Displays grew 10% year over year, representing 16% of our 2025 corporate revenue and 15% of the fourth-quarter revenue. We have seen strength and continue to see strength in the machine vision market with new advanced products ramping to production alongside existing products that continue to gain share. We also expect our first ramp in the AR display segment with our silicon backplane for OLED on silicon, which has started production this past quarter. We are tracking this first adoption and its overall market carefully and with optimism as it may have significant value for Tower in the following years.

感測器和顯示器業務年增 10%，占我們 2025 年公司營收的 16%，佔第四季營收的 15%。我們已經看到機器視覺市場強勁成長，這種強勁成長動能也將持續下去，新的先進產品正在加速投產，而現有產品也不斷擴大市場份額。我們也預計將在 AR 顯示器領域實現首次量產，這得益於我們用於矽基 OLED 的矽背板，該產品已於上個季度開始生產。我們正在密切關注這項首次應用及其整體市場，並抱持樂觀態度，因為它在未來幾年可能對 Tower 產生重大價值。

Mixed Signal CMOS represented 7% and Discrete represented 11% of our 2025 corporate revenues. Year on year, we have seen decreases of 18% and 14% respectively, supporting our value-driven growth strategy allowing additional capacity for the higher margin and highest margin platforms to replace these two application sets.

2025 年，混合訊號 CMOS 占公司總收入的 7%，分立元件占公司總收入的 11%。與前一年相比，我們分別下降了 18% 和 14%，這支持了我們以價值為導向的成長策略，從而為利潤率更高和利潤率最高的平台提供了額外的產能，以取代這兩套應用。

Regarding capacity expansion, during our previous earnings release in November '25, we announced an increase of investment for Silicon Photonics and Silicon Germanium growth targeting a tripling of SiPho capacity against our targeted Q4 25 Silicon Photonics actual shipments, having stated a target that this would be online to begin Silicon starts in the second half of 2026. Due to continued growth in demand, we are announcing today additional CapEx investment of $270 million on top of the previously announced $650 million capacity expansion plan. This total capacity is targeted to yield capacity growth greater than 5x of the actual fourth-quarter monthly wafer shipments of Silicon Photonics wafer shipments to be compared to the 3x target that we gave during the Q3 public release, and over 70% of the total SiPho capacity is either presently reserved or in the process of being reserved through 2028 firmly backed with customer prepayment.

關於產能擴張，我們在 2025 年 11 月的上一次收益發布會上宣布，將增加對矽光子學和矽谷增長的投資，目標是使矽光子產能比我們 2025 年第四季度矽光子學實際出貨量目標提高三倍，並表示目標是在 2026 年下半年開始投產三倍學。由於需求持續成長，我們今天宣布，在先前宣布的 6.5 億美元產能擴張計畫的基礎上，再追加 2.7 億美元的資本支出投資。這個總產能的目標是實現比第四季實際月矽光子晶圓出貨量高出 5 倍以上的產能成長，而我們在第三季公開發佈時給出的目標為 3 倍。目前，超過 70% 的矽光子總產能已被預訂或正在預訂中，直至 2028 年，並有客戶預付款作為有力保障。

For the fourth quarter, utilization rates were Fab 2 operated at about 60% utilization as we are now in the final stages of Silicon Germanium and Silicon Photonics capacity qualification for a variety of flows. Fab 3 maintained our model full utilization of 85% and is still adding capacity for increasing Silicon Photonics capability. Fab 5 was at 75% utilization and Fab 7 was fully utilized, well above our 85% utilization model. Fab 9 was at 65% utilization presently in a Silicon Photonics and Silicon Germanium ramp.

第四季度，Fab 2 的利用率約為 60%，因為我們目前正處於矽鍺和矽光子產能認證的最後階段，該產能認證針對各種製程。Fab 3 保持了我們 85% 的滿載利用率，並且仍在增加產能以提高矽光子學能力。Fab 5 的利用率為 75%，Fab 7 的利用率為 100%，遠高於我們 85% 的利用率模型。目前 Fab 9 的利用率為 65%，用於矽光子學和矽鍺的產能爬坡。

As stated in our press release, Intel has expressed its intention not to perform under the September 23 Fab 11X agreement. We are presently in a mediation process. All flows which have been transferred or are in the process of being transferred to Fab 11X originally qualified in our Japanese 300-millimeter factory Fab 7. Customers are being redirected to be supported by this Fab in Japan.

正如我們在新聞稿中所述，英特爾已表示無意履行 9 月 23 日與 Fab 11X 達成的協議。我們目前正在進行調解。所有已轉移或正在轉移至 Fab 11X 的製程，最初均在我們日本 300 毫米工廠 Fab 7 中得到驗證。客戶將被轉接到日本的這家工廠以獲得支援。

For guidance, we guide our first quarter of 2026 mid-range revenue to be $412 million plus/minus 5%, representing a 15% increase as compared to the start of 2025. We target quarter-over-quarter revenue and profitability growth throughout 2026.

作為參考，我們預計 2026 年第一季的中期收入為 4.12 億美元，上下浮動 5%，比 2025 年初增長 15%。我們的目標是到 2026 年實現季度環比營收和獲利成長。

Based upon the thriving corporate ecosystem we have developed intertwined with deeply trusted customer partner alliances, we are pleased to provide a revised financial model. This new model demonstrates our value-driven growth strategy. Please refer to slide 7.

基於我們已建立的蓬勃發展的企業生態系統，以及與客戶夥伴之間建立的深厚信任的聯盟，我們很高興提供修訂後的財務模型。這一新模式體現了我們以價值為導向的成長策略。請參考投影片7。

First, the assumption. Beyond the $920 million CapEx plans that have been released, no additional CapEx, clean room space, or otherwise additional monies are required to achieve this model. This model is based on utilizing our owned capacity at an 85% utilization level. Intel Fab 11X is not included in this model. Revenue is $2.84 billion which will create 39.4% gross margin, 31.7% operating margin, a 7.7 point drop from gross operating margin, demonstrating a highly efficient business, and if not the very best, certainly among the best in our industry.

首先，我們來談談這個假設。除了已公佈的 9.2 億美元資本支出計劃外，無需額外的資本支出、無塵室空間或其他額外資金即可實現此模式。此模式基於我們自有產能利用率達到 85% 的原則。本機型不包含 Intel Fab 11X。營收為 28.4 億美元，毛利率為 39.4%，營業利潤率為 31.7%，比毛營業利潤率下降了 7.7 個百分點，這表明公司營運效率很高，即使不是業內最好的，也絕對是最好的之一。

Such efficiency is seen in more than just margin dollars. It is reflective to the speed of decision making and execution. Speed is a sustainable differentiator. Net profit is $750 million or 26.4% net profit margins. All tools and customer qualifications are planned to be fully completed within 2026, hence and most importantly, we target to achieve this model in the calendar year 2028.

這種效率不僅體現在利潤額上。這反映了決策和執行的速度。速度是一項可持續的差異化優勢。淨利為7.5億美元，淨利率為26.4%。所有工具和客戶資格都計劃在 2026 年內全面完成，因此，最重要的是，我們的目標是在 2028 年實現這一模式。

Now I would like to turn the call to our CFO Oren Shirazi. Oren, please.

現在我想把電話交給我們的財務長奧倫希拉齊。奧倫，拜託了。

Hello, everyone. Earlier today, we released our financial results for the fourth quarter of 2025 and for the full year and also released our balance sheet and cash flow report. Now I will review the results highlights as well as the highlights of our CapEx investment. And afterwards, I will present our updated target financial model resulting in higher revenue and profit margin than the prior model.

大家好。今天早些時候，我們發布了 2025 年第四季和全年的財務業績，並發布了資產負債表和現金流量報告。現在我將回顧業績亮點以及我們資本支出投資的亮點。之後，我將展示我們更新的目標財務模型，該模型將比之前的模型帶來更高的收入和利潤率。

Let's first look into the P&L. In 2025, we achieved quarter-over-quarter revenue increase during the year, which had accelerated in the second half of 2025, resulting in record revenue of $440 million in the fourth quarter of 2025, reflecting a year-over-year revenue increase of 14% and a quarter-over-quarter revenue increase of 11%. Gross profit for the fourth quarter of 2025 was $118 million an increase of $25 million or 26% compared to the previous quarter. And operating profit was $71 million, 40% higher as compared to the prior quarter.

我們先來看看損益表。2025 年，我們實現了季度環比營收成長，並在 2025 年下半年加速成長，最終在 2025 年第四季創下 4.4 億美元的營收紀錄，年成長 14%，季增 11%。2025 年第四季毛利為 1.18 億美元，較上一季增加 2,500 萬美元，增幅達 26%。營業利潤為7,100萬美元，較上一季成長40%。

Net profit for the fourth quarter of 2025 was $80 million, an increase of $26 million or 49% compared to net profit of $54 million in the prior quarter, and the earnings per share was $0.71 basic and $0.70 diluted compared to $0.48 basic and $0.47 diluted earnings per share reported for the prior quarter. Please note that income tax expenses line in the P&L includes a non-recurring tax benefit recorded in the fourth quarter of 2025, resulting in an all-in 2% effective tax rate. For 2026 and beyond as required by Pillar 2 regulation, we estimate all-in tax effective rates to be at least 15% in all our manufacturing sites.

2025 年第四季淨利潤為 8,000 萬美元，比上一季的淨利 5,400 萬美元增加了 2,600 萬美元，增幅達 49%；每股收益基本為 0.71 美元，稀釋後為 0.70 美元，而上一季報告的每股收益基本為 0.48 美元，稀釋後為 0.47 美元，稀釋後為 0.40.47 美元。請注意，損益表中的所得稅費用項目包含 2025 年第四季確認的一項非經常性稅務優惠，因此綜合實際稅率為 2%。根據第二支柱法規的要求，我們預計 2026 年及以後所有生產基地的綜合有效稅率至少為 15%。

For the full-year 2025, we reported revenue of $1.57 billion, 9% higher as compared to $1.44 billion in 2024. Gross profit and operating profit for '25 were $364 million and $194 million respectively compared to $339 million and $191 million in 2024 respectively. Net profit for 2025 was $220 million or $1.97 basic and $1.94 diluted earnings per share compared to $208 million net profit in 2024.

2025 年全年，我們的營收為 15.7 億美元，比 2024 年的 14.4 億美元成長了 9%。2025 年的毛利和營業利潤分別為 3.64 億美元和 1.94 億美元，而 2024 年的毛利和營業利潤分別為 3.39 億美元和 1.91 億美元。2025 年淨利為 2.2 億美元，基本每股收益為 1.97 美元，稀釋每股收益為 1.94 美元，而 2024 年淨利為 2.08 億美元。

Moving to our balance sheet. Our balance sheet is very strong evidenced by the following indicators and financial ratios. As of end of December 2025, our assets totaled over $3 billion primarily comprised of $1.5 billion in fixed assets predominantly comprised of FAB machinery and $1.7 billion of current assets. The recent increase in other long-term assets as compared to past periods is mostly attributed to the Newport Beach FAB lease extension prepayment as was announced in November 2025, and paid, which is presented as an asset as required by GAAP. Current assets ratio is very strong at about 6.5 while shareholders’ equity reached a record number of $2.9 billion at the end of December 2025.

接下來來看看我們的資產負債表。我們的資產負債表非常穩健，以下指標和財務比率可以證明這一點。截至 2025 年 12 月底，我們的資產總額超過 30 億美元，主要包括 15 億美元的固定資產（主要由 FAB 機械組成）和 17 億美元的流動資產。與以往期間相比，其他長期資產的近期增加主要歸因於 2025 年 11 月宣布並已支付的 Newport Beach FAB 租賃延期預付款，根據 GAAP 的要求，該款項已列為資產。流動資產比率非常強勁，約 6.5，而截至 2025 年 12 月底，股東權益達到創紀錄的 29 億美元。

Hedging, I would now like to describe our currency hedging activities. In relation to the Japanese yen since the majority of TPSCo revenue is denominated in yen and the vast majority of TPSCo costs are in JPY we have a natural hedge over most of our Japanese business and operations. To mitigate part of the remaining yen exposure, we are executing zero-cost cylinder transactions to hedge currency fluctuations. Hence, while the yen rates against the dollar may fluctuate, there is limited impact on our margin.

接下來，我想介紹一下我們的貨幣對沖活動。就日圓而言，由於 TPSCo 的大部分收入以日圓計價，而 TPSCo 的絕大部分成本也以日圓計價，因此我們對大部分日本業務和營運都擁有天然的對沖作用。為了減輕剩餘日圓風險敞口的一部分，我們正在執行零成本的鋼瓶交易來對沖貨幣波動。因此，儘管日圓兌美元匯率可能會波動，但對我們的利潤率影響有限。

Similar concept goes to the Israeli shekel. In relation to the Israeli shekel currency, while we have no revenue in this currency, since a portion of our costs in Israel is denominated in shekel, we also hedge a large portion of such currency risk by engaging in zero-cost cylinder transactions to mitigate this exposure. Hence, while the shekel rate against the dollar may fluctuate, the impact on our margins is limited.

以色列謝克爾也遵循類似的概念。關於以色列謝克爾貨幣，雖然我們沒有以該貨幣計價的收入，但由於我們在以色列的部分成本是以謝克爾計價的，因此我們也通過進行零成本鋼瓶交易來對沖很大一部分此類貨幣風險，以降低這種風險敞口。因此，儘管謝克爾對美元的匯率可能會波動，但對我們的利潤率的影響有限。

Now moving into our CapEx investment plan and its impact on our financial model. As we announced today, in order to support the increasing SiPho and SiGe demand, we are allocating an additional $270 million of cash to invest in capacity and capability SiPho equipment which would result in a total of $920 million cash investments in CapEx including the $650 million we already announced during 2025. These $920 million CapEx investments will expand our Fabs’ capacity in our 8-inch fabs in Israel, Newport Beach, Texas, and also in our 12-inch Uozu fab in Japan. This CapEx plan includes a large portion of capability CapEx for advanced development and high-end RF technology-related projects. Approximately 28% of the above-stated $920 million CapEx investments were already paid to date, while the remaining 72% of the $920 million are expected to be paid in 2026 and 2027.

接下來，我們將探討我們的資本支出投資計畫及其對我們財務模型的影響。正如我們今天宣布的那樣，為了支持不斷增長的矽磷化物 (SiPho) 和矽磷化物設備的產能和能力，這將使資本支出現金投資總額達到 9.2 億美元，其中包括我們已經宣布的 6.5 億美元。這9.2億美元的資本支出投資將擴大我們在以色列、德州紐波特海灘的8吋晶圓廠以及我們在日本Uozu的12吋晶圓廠的產能。此資本支出計劃包括很大一部分用於先進開發和高端射頻技術相關項目的能力資本支出。上述 9.2 億美元的資本支出投資中，約有 28% 已支付，其餘 72% 預計將在 2026 年和 2027 年支付。

Moving to the financial model. Following these investments, which are expected to drive greater revenue and incremental margins as compared to our prior model, which we released more than two years ago, we are providing an updated target financial model resulting in significantly higher revenue, profitability, and margin targets. Please note, the model is based on many forward-looking operational business and financial assumptions including the assumption that all our fabs will operate at 85% utilization post installation and qualification of the $920 million equipment tools we are investing in. Assumptions considering a modest average wafer selling price reduction of existing products and/or flows that we target will be offset by new products and/or flows introductions. Assumptions that our cost estimates will not differ significantly from our current assumptions.

接下來進入財務模型。經過這些投資，預計與兩年前發布的先前模型相比，將帶來更高的收入和利潤率，因此我們提供了一個更新的目標財務模型，該模型將帶來更高的收入、獲利能力和利潤率目標。請注意，該模型基於許多前瞻性的營運業務和財務假設，包括假設我們所有晶圓廠在安裝和驗證我們投資的 9.2 億美元設備工具後，利用率將達到 85%。假設現有產品和/或流程的平均晶圓售價略有下降，那麼新產品和/或流程的推出將抵消這一下降。假設我們的成本估算不會與我們目前的假設有顯著差異。

And lastly, please note that the model does not include Fab 11X capacity, revenue, and margin. Now any possible additional fabs and/or new capacity that has not yet been obtained, established, or announced to date. Under this model, which you may see in the slide for your reference, we are targeting $2.84 billion in annual revenue, which is $1.27 billion higher or 81% higher in revenue than our actual full-year 2025 revenue, $1.12 billion in gross profit which is more than tripling our 2025 gross profit. This level of gross profit reflects approximately 40% gross margin which reflects a 59% incremental gross profit that are derived from the incremental revenue when comparing the model to FY 2025 actual results.

最後，請注意，該模型不包括 Fab 11X 的產能、收入和利潤率。現在還有任何可能新增的晶圓廠和/或新增產能，但這些尚未獲得、建立或公佈。根據該模型（您可以在幻燈片中看到以供參考），我們的目標是年收入達到 28.4 億美元，比我們 2025 年全年的實際收入高出 12.7 億美元，增幅達 81%；毛利達到 11.2 億美元，是 2025 年毛利潤的三倍多。這一毛利水準反映了約 40% 的毛利率，與 2025 財年的實際結果相比，該模型帶來的增量收入產生了 59% 的增量毛利。

It also states $900 million in annual operating profit, which is 4.6x our actual FY25 operating profit, reflecting 32% operating margin. This reflects 55% incremental operating profit margins that are derived from the incremental revenue when comparing the model to FY 2025 actual results.

報告也指出，年度營業利潤為 9 億美元，是 2025 財年實際營業利潤的 4.6 倍，營業利益率為 32%。這反映了與 2025 財年實際結果相比，模型帶來的增量收入帶來了 55% 的增量營業利潤率。

And lastly, on net profit, $750 million, more than tripling the full-year 2025 net profit, reflecting 26% net margin, like Russell stated, which reflects 42% incremental net profit margins that are derived from the incremental revenue when comparing the model to FY 2025 actual results.

最後，淨利潤為 7.5 億美元，是 2025 年全年淨利潤的三倍多，淨利潤率為 26%，正如 Russell 所說，這反映了與 2025 財年實際結果相比，模型帶來的增量收入所對應的 42% 的增量淨利潤率。

To summarize, comparing this updated financial model to the prior financial model that we presented more than two years ago, gross profit, operating profit, and net profit are much higher, 50%, 60% each higher as compared to the prior model, mostly driven by the higher SiPho and SiGe mix and the additional value we bring to our customers.

總而言之，將這項更新後的財務模型與我們兩年多前提出的先前財務模型進行比較，毛利潤、營業利潤和淨利潤都高得多，分別比先前的模型高出 50% 和 60%，這主要是由於矽磷化物和矽鍺比例的提高以及我們為客戶帶來的額外價值。

That concludes my prepared remarks. Now I'd like to turn the call back to the operator, so we can take your questions.

我的發言稿到此結束。現在我想把電話轉回給接線員，以便我們回答您的問題。

(Operator Instructions)

（操作說明）

Mehdi Hosseini, Susquehanna Financial Group.

Mehdi Hosseini，Susquehanna Financial Group。

A couple of questions from me. Russell, I want to dive into the announcement that you had last Thursday, increased collaboration with Nvidia. The press release was making a reference to module. And I want to better understand what that implies. Does this mean that you will be manufacturing a transceiver for Nvidia or the module is more of a broader -- a reflection of a broader services that you would provide for this customer?

我有幾個問題。Russell，我想深入探討你上週四宣布的與英偉達加強合作的消息。新聞稿中提到了模組。我想更好地理解這意味著什麼。這是否意味著您將為英偉達生產收發器，還是該模組代表更廣泛的服務——反映了您將為該客戶提供的更廣泛的服務？

And I have a follow-up.

我還有一個後續問題。

No, the part of our role in the module is the output parameters of our photonics or of the TIA or of the drivers or of the -- for the pluggable or as well for the copper or optical cable. But the partnership is referring to the fact of alignments and needs directly and through our module customers and understandings of supply needs and commitments on supply shipments.

不，我們在模組中扮演的角色是光子裝置、TIA、驅動器或可插拔元件的輸出參數，以及銅纜或光纜的輸出參數。但這種合作關係指的是我們與模組客戶之間直接或間接的協調一致和需求，以及對供應需求和供應出貨承諾的理解。

Okay. And your 5x capacity increase for silicon photonics, silicon germanium, is that -- does that include incremental demand from Nvidia and partners?

好的。你們矽光子學、矽鍺的產能增加5倍，是否包括英偉達及其合作夥伴的增量需求？

Yeah, that's referring to total demand. Well, I wouldn't say it's necessarily referring to total demand. It's an answer to demand, but it's the actual capacity that we're building. So if you look at -- we had referred to as the $380 million run rate that we had in Q4, take off of that some small amount of NRE, which we don't specify, the silicon wafers that we shipped for the fourth quarter, that exact amount of silicon wafers by capacity, we plan to have 5 times more of that in the fourth quarter of 2026.

是的，那指的是總需求。嗯，我倒不認為它一定是指總需求。這是對需求的回應，但這是我們正在建造的實際產能。所以，如果你看一下——我們之前提到的第四季度 3.8 億美元的運行率，減去一些我們沒有具體說明的非經常性工程費用，再減去我們第四季度出貨的矽晶圓數量（按產能計算），我們計劃在 2026 年第四季度達到這個數字的 5 倍。

Got it. Okay. And then on your power business line, does -- would you be able to also help prospective customers on the high voltage, especially as the next generation of AI server rack will require 800-volt DC?

知道了。好的。那麼，在您的電力業務方面，您是否也能幫助潛在客戶解決高壓問題，特別是下一代 AI 伺服器機架將需要 800 伏特直流電？

We have a variety of roadmap activities. We don't, at this moment, have an 800-volt platform on an IC. We do have 800-volt capabilities in fabs but not in an IC. But we do have higher voltage IC capabilities with and without SOI.

我們有多種路線圖活動。目前我們還沒有基於積體電路的 800 伏特平台。我們在晶圓廠中具備 800 伏特的製造能力，但在積體電路中則不具備。但是，無論是否採用SOI封裝，我們都有更高電壓的積體電路能力。

Tavy Rosner, Barclays.

Tavy Rosner，巴克萊銀行。

Just following up on the Nvidia question. So just to clarify, you're not actually shipping directly to Nvidia, you're shipping through resellers that will just send your technology on to them.

關於英偉達的問題，我再補充一下。所以澄清一下，你其實不是直接把產品寄給英偉達，而是透過經銷商寄送，經銷商再把你的科技產品轉寄給他們。

That is correct. We -- as far as the photonics itself, we do not ship that directly to Nvidia. And as far as specifics of projects or activities that we're doing with Nvidia that anything that was not specifically stated in the PR, I would not be at liberty to talk about. But as far as the present photonics -- silicon-based photonics ICs, they are all being designed by and shipped through other module makers or integrators.

沒錯。就光子學本身而言，我們並沒有直接向英偉達供貨。至於我們與英偉達合作的具體項目或活動，凡是公關稿中沒有明確說明的內容，我都不能妄加評論。但就目前的矽基光子積體電路而言，它們都是由其他模組製造商或整合商設計並透過他們進行銷售的。

Okay. Understood. And then around CPU, I mean, you spoke about the opportunity. I think I recall last quarter, maybe it was a different conversation. You guys spoke about the ability to add value to the ecosystem through lasers, power connectors, and you guys also doing any R&D on the actual CPU as well, maybe through like third-party packaging in order to have your kind of own end-to-end offering at some point?

好的。明白了。然後，關於 CPU，我的意思是，你談到了這個機會。我想我記得上個季度，也許是另一次談話。你們談到可以透過雷射、電源連接器等為生態系統增加價值，你們是否也在對 CPU 本身進行研發，也許會透過第三方封裝等方式，以便在某個時候推出自己的端到端產品？

Direct packaging of the CPU, no. We're certainly working on multiple architectures of CPO. And certainly, the XPU would be -- or could be incorporated into the CPO. But the specific activity right now of our engagement, well, that's not even 100% true. Yeah, I mean, we're certainly working with XPU makers on CPO strategies.

CPU直接封裝，否。我們當然正在研究多種 CPO 架構。當然，XPU 將會——或者說可以——被納入 CPO 中。但就我們目前的參與活動而言，嗯，這甚至不完全屬實。是的，我的意思是，我們當然正在與 XPU 製造商合作制定 CPO 策略。

Okay. Understood. And then very last one for me. The rollout of additional CapEx, I think, I recall you saying it's going to be all live by end of 2026. Is there any chance that it can come in sooner depending on several factors, maybe some of them beyond your control, but like is there any chance or you have the certainty that that's not going to be online before the end of the year?

好的。明白了。然後，對我來說，這是最後一個。關於額外資本支出項目的推出，我記得你說過，所有項目將在 2026 年底前全部上線。是否有可能提前上線，這取決於多種因素，其中一些因素可能是你無法控制的，但是，有沒有可能或你能確定它不會在年底前上線？

The capacity qualification ramp will be happening throughout the year. So it will -- the biggest portion of this $920 million should easily be online, I mean, fully qualified within the third quarter, on or before the third quarter, with growth happening in the first and the second quarter as well. The most recent orders that we've done also have tools that are coming in in the second quarter. But what we've stated is that and what I just stated is that expect and target that by December, everything will be fully qualified in order to be able to do customer starts. In order to have everything fully qualified by December, the tools really have to arrive before the end of the third quarter and nominally by mid-third quarter.

產能資格認證的逐步提升將貫穿全年。所以肯定會實現——這 9.2 億美元中的大部分應該很容易就能在線實現，我的意思是，在第三季度內，甚至在第三季度之前，就能完全獲得批准，而且第一季度和第二季度也會有增長。我們最近接到的訂單中也包含一些將在第二季到貨的工具。但我們已經說過，而且我剛才也說過，預計到 12 月份，所有環節都將完全合格，以便能夠開始為客戶服務。為了確保所有工作都能在 12 月完全完成，工具必須在第三季結束前到貨，名義上最遲也要在第三季中期到貨。

So that's where you could be thinking of is that linear or not, there will be a distribution of tools. Some have already arrived and the bulk of this $920 million will be arriving between now and mid-third quarter.

所以，你可以思考的是，無論這種分佈是線性的還是非線性的，工具都會有分佈。部分款項已經到位，這9.2億美元中的大部分款項將從現在到第三季中期陸續到位。

Cody Acree, Benchmark StoneX.

Cody Acree，Benchmark StoneX。

Congrats on the steady and impressive progress. Russell, maybe could you just give us a little more color on your expectations for your silicon photonics contribution in '26 and '27, specifically with the 70% commitment already talking about prepaid? It looks like your visibility should be pretty solid for the next couple of years.

恭喜你們取得穩定且令人矚目的進展。Russell，您能否更詳細地介紹一下您對 2026 年和 2027 年矽光子學貢獻的預期，特別是關於已經承諾的 70% 的預付款？看來未來幾年你的曝光度應該會相當不錯。

Yeah, definitely. The demand is there. Certainly, we're very aware of the demand. Right now, the -- if your question is really on the ramp profile, the ramp profile is pure operational execution at this moment. I mean, there's some technical execution still.

是的，當然。需求是存在的。當然，我們非常清楚市場需求。目前，——如果你的問題確實與爬坡曲線有關，那麼爬坡曲線目前純粹是運行執行的問題。我的意思是，還有一些技術執行的問題需要解決。

There's some flows that still would need to be qualified, be it San Antonio or be it Migdal Haemek that are not yet qualified that are in the first order -- not the first order, but solely qualified in Newport Beach as that was the fab that most all of this development was done at. So you have -- some more technical work has to be done. But that's, for the most part, behind us on the technical work.

還有一些流程仍需進行資格審查，無論是聖安東尼奧還是米格達爾·哈埃梅克，這些流程尚未獲得資格審查，它們屬於第一類——不是第一類，而是僅在紐波特海灘獲得資格審查，因為那裡是大部分開發工作完成的地方。所以，還需要做一些技術性的工作。但就技術工作而言，這大部分已經過去了。

From the time that everything is ready to be qualified, you still have several months for live testing in order to have customers qualify the flow themselves, if you know what I mean. So in some cases, it goes through HTOL and whatever other live tests the customer requires in its own commitments to their end customers, but the bulk of this is just operational execution.

從一切準備就緒可以進行驗證之時起，你還有幾個月的時間進行實地測試，以便讓客戶自己驗證流程，你明白我的意思嗎？因此，在某些情況下，它會經過 HTOL 以及客戶為履行對最終客戶的承諾而要求的任何其他現場測試，但大部分只是營運執行。

I think that's one reason that we're so bullish and confident on where we're at and where we're going on this model that we just gave of the 2.8 -- what was it $2.84 billion and the $750 million net profit. When your target and your plans are to have everything online for wafer starts in December, okay, let's say, worst case, you miss it by one month, two months, three months, okay, maybe, but it's there.

我認為，這就是我們對目前的發展狀況以及未來發展方向如此樂觀和自信的原因之一，我們剛剛公佈的模型顯示，營收為 28.4 億美元，淨利潤為 7.5 億美元。如果你的目標是讓所有晶圓生產在 12 月上線，那麼，好吧，假設最糟糕的情況是，你晚了一個月、兩個月、三個月，好吧，也許吧，但它就在那裡。

So if it's -- will you have the full start capability in December? We target to have that, and I believe that we will. Can it push out that one or two tools isn't fully qualified for whatever reason? Obviously, this is a lot of equipment coming from suppliers, and although we're very good and the suppliers are very good at doing a final test at the supplier site, the tools are all disassembled and shipped. During the disassembly and shipment, there can be something that's broken or goes wrong that isn't identified immediately during the, what's called Tier 1, Tier 2 start-up at our site.

那麼如果真是這樣的話──你們12月份能完全啟動嗎？我們的目標是實現這個目標，我相信我們能夠做到。它能否排除一兩個工具因某些原因而不完全合格的情況？顯然，這其中有許多設備來自供應商，雖然我們和供應商都非常擅長在供應商現場進行最終測試，但所有工具都是拆卸後運輸的。在拆卸和運輸過程中，可能會出現一些損壞或出錯的情況，而這些情況在我們現場所謂的 Tier 1、Tier 2 啟動過程中無法立即發現。

So it's possible that that could take a little bit longer and one or two, three tools can be delayed beyond the plan. It's also possible that for whatever reason, the supplier themselves misses their initial target, and that can happen. The same as not that Tower would ever do it, but sometimes wafer manufacturers miss their commitments. No, just a joke there.

所以，這可能需要更長時間，一到兩個、三個工具的交付可能會比計劃推遲。供應商本身也可能因為某些原因而未能達到最初的目標，這種情況是可能發生的。雖然 Tower 公司不太可能這樣做，但晶圓製造商有時會違背承諾。不，我只是開個玩笑。

But the point being, whether it's December or January, maybe February, I don't know. It could also be -- I mean, we've released that we tend to have everything up and running in December. People that want to make sure that future commitments will always give targets where they believe they have some leeway. So the internal target should probably be more aggressive than the express target to the street, right?

但關鍵是，到底是十二月、一月，還是二月，我不知道。也可能是——我的意思是，我們已經宣布，我們通常會在 12 月讓所有東西都正常運作。那些希望確保未來承諾始終設定有一定迴旋餘地的目標的人。所以內部目標可能應該比面向市場的快速目標更具侵略性，對吧？

But the big point I'm trying to make in maybe too many words is that whether we hit the full qualification of start capability in December or whether it's November or whether it's January or February, it will be hit. And the demand is there, it's committed, and it will be used. So the model will be hit. Now from the time that you start all the starts, it's some period of time to get everything ramped and qualified. And then it's some amount of months before you can ship and get the revenue.

但我想表達的重點（可能說得有點囉嗦）是，無論我們是在 12 月、11 月、1 月還是 2 月達到完全的啟動能力認證，我們最終都會達到這個目標。需求就在那裡，需求已經確定，而且一定會被使用。所以這個模型會受到影響。從你開始進行所有起跑訓練到一切準備就緒並獲得資格，需要一段時間。然後還需要幾個月的時間才能發貨並獲得收入。

So we feel very comfortable in talking about the 2028 to hit our model because the demand is there. Customers have committed to that demand to the extent we did not ask customers for reservation fees, they wanted it. They know how precious, especially from Tower, SiPho demand is as we are truly by far, the leader in silicon PICs. So they want the wafers, and it's just really in our hands as far as operational execution.

因此，我們有信心在 2028 年實現我們的目標，因為市場需求已經存在。客戶們已經接受了這項要求，以至於我們沒有向客戶收取預訂費，是他們主動要求的。他們知道矽光電晶片的需求有多寶貴，尤其是來自 Tower 的需求，因為我們是矽光子積體電路領域的絕對領導者。所以他們想要晶圓，至於營運執行方面，就完全取決於我們了。

Now I'd like to say it's 100% in our hands. It also is in the hands of our suppliers. But they're good suppliers. And we have a good relationship with our suppliers.

現在我想說，一切都完全掌握在我們自己手中。這也可能掌握在我們的供應商手中。但他們是很好的供應商。我們與供應商保持著良好的關係。

We really focus on having strong relationships with our customers. To have a good relationship with the customer, you must also have that same model with your suppliers, right? I mean, what goes around comes around. So you can't easily be someone that has a mentality to not treat a supplier well and expect a customer to treat you well and vice versa.

我們非常注重與客戶建立牢固的關係。要與客戶建立良好的關係，你也必須與供應商建立同樣的關係，對嗎？我的意思是，善惡終有報。所以，你不能抱著不善待供應商卻期望顧客善待你的心態，反之亦然。

So I think, Cody, maybe any -- too many words. But our plans are very firm, very strong. Can something be impacted by a month or two, one way or the other? Of course. But the plans are there, and if it is impacted by a month or two, it's not impacted by the bulk of the capacity growth. There will be one or two tools or, I don't know, a handful of tools that sometimes are called a lemon tool. It's not necessarily a lemon tool.

所以我覺得，科迪，也許——話太多了。但是我們的計劃非常堅定，而且非常有力。一兩個月的時間會對某件事產生影響嗎？當然。但計劃已經制定，即使受到影響一兩個月，也不會受到產能成長的大部分影響。會有一兩件工具，或者，我不知道，可能有一小撮工具，它們有時被稱為“檸檬工具”。它未必是個劣質工具。

It means that there's a problem that wasn't found immediately. That can delay something. But there will always, throughout this year, we will definitely have incremental capacity growth.

這意味著存在一個未能立即發現的問題。這可能會延誤某些事情。但今年全年，我們的產能肯定會持續成長。

Maybe can I just continue on with your mobile business. Any concerns about the ongoing memory shortages or the increased prices that have been called out by some of your peers in the industry and the impact to potential unit volumes in the handset market?

或許我可以繼續做你的手機業務？您是否擔心業內一些同行指出的持續記憶體短缺或價格上漲問題，以及這些問題對手機市場潛在銷售的影響？

Cody, I mean, there's always a concern when you have something in the market that you yourself have no say in or control of. So yes, there's definitely a concern there. We work with our customers closely to understand what their inventory levels are. They try to understand what their customer inventory levels are.

科迪，我的意思是，當市場上出現一些你既沒有發言權也沒有控制權的東西時，總是會讓人感到擔憂。所以，是的，這確實令人擔憂。我們與客戶密切合作，以了解他們的庫存水準。他們試圖了解客戶的庫存水準。

And to be as convinced as possible that the plan that we have first that our start plan for the year can be hit. But are we -- I'd love to be able to say that there's no concern we're impervious to it. We're not. There are factors in the market that always play that you never want to be a victim, so you try to do as good a planning as you can and, in the best case, to have alternatives should a certain capacity not be used in the fab that it can be replaced with something else where we talked about the fact of intentionally working out some lower-margin products to allow room for higher-margin products.

而且要盡可能確信，我們所製定的計劃能夠實現，即我們今年的啟動計劃能夠成功。但我們真的——我很想說我們完全不用擔心，我們對此免疫。我們不是這樣。市場中總有一些因素在起作用，你永遠不想成為受害者，所以你要盡可能做好規劃，最好是有備選方案，以防工廠的某個產能沒有被利用，可以用其他東西來替代。我們之前討論過，要有意生產一些利潤較低的產品，以便為利潤較高的產品騰出空間。

The lower-margin products are still in demand and there's always the possibility if there is a gap in the fab because a demand of what you thought would be there is not there, we have the opportunity to backfill it with something else.

低利潤產品仍有需求，如果因為沒有達到預期需求而導致生產線出現缺口，我們總有機會用其他產品來填補空缺。

And if that something else is maybe not preferred because it's not the same margin profile, but it can be done. So at least you're absorbing your fixed cost.

如果其他方案可能不是首選，因為它的利潤率不同，但也是可以實現的。所以至少你承擔了固定成本。

Richard Shannon, Craig-Hallum Capital Group.

理查德·香農，克雷格-哈勒姆資本集團。

This is Tyler on for Richard. Sorry to disappoint. I have a question. I had a question on this model that you gave and the 2028 timeline. Is this a run rate in 2028, or is this the full year?

這裡是泰勒替理查德發言。很抱歉讓您失望了。我有個問題。我對您提供的這個模型以及 2028 年時間表有一個疑問。這是 2028 年的運行率，還是全年的運行率？

Yes. No, certainly, we will achieve it by run rate, and we target to get a full year. But what we stated is that it would be achieved within the year. So we're -- our target and what I've stated is that is our target. One could definitely believe that we will hit it by run rate.

是的。不，當然，我們會透過提高運行率來實現目標，我們的目標是完成整個賽季。但我們之前說過，這個目標將在一年內實現。所以，我們的目標，也是我剛才所說的，就是我們的目標。人們完全可以相信，憑藉跑動率，我們一定能打進決賽。

And nominally, we'd love to hit it for full year, and it's possible.

名義上，我們希望全年都能達到這個目標，而且這是有可能的。

Okay. Great. And then the silicon photonics, I know you just mentioned you could backfill other things. But at this point, with all of the CapEx investments that you make, is this going to put the fabs at fully utilization for that model?

好的。偉大的。至於矽光子學，我知道你剛才提到你可以用其他東西來補充說明。但是，就目前而言，投入了這麼多資本支出，這能使晶圓廠在該模式下達到滿載運作嗎？

No. Silicon Photonics would not bring any of the factories to the full photo utilization. But it's not the Silicon Photonics that I was talking about as far as backfilling. That question was the specific question with regard to the RF Mobile because of the fear of the high-bandwidth memory manufacturers focusing on that for data center rather than supplying it elsewhere. And without the memory that it might not -- that there could be a decline in the overall mobile integrator by not having the memory they need for their phones.

不。矽光子技術並不能使任何一家工廠實現光能的充分利用。但我所說的回填技術並不是矽光子學。這個問題是針對射頻行動通訊技術提出的，因為人們擔心高頻寬記憶體製造商會將重點放在資料中心，而不是將其供應到其他領域。如果沒有足夠的內存，行動整合商的整體業務可能會下降，因為他們的手機沒有足夠的內存。

That was what the question was. So I was saying if that was the case, that capacity is fungible.

這就是問題所在。所以我的意思是，如果真是那樣的話，這種能力就是可以被取代的。

Got it. But with this, I think what I'm really getting at is with this CapEx spend that you're adding today, does that bring us to the 85% utilization?

知道了。但是，我想說的是，今天你們增加的這筆資本支出，是否能讓我們達到 85% 的利用率？

It does, providing that the other flows are used to the prescribed capacity that we allotted to them. So no, it's not -- if it was only silicon photonics, it would not be 85% utilization. But must understand as well, and this is an important point. We're focusing on the silicon photonics, our commitments around the silicon photonics, where I say that the RF SOI, if you will, for the most part, that's pretty fungible to power. I mean, there are some layers that are different, but relatively fungible for power, relatively fungible for imaging.

確實如此，前提是其他流量的使用量達到我們分配給它們的規定容量。所以，不，並非如此──如果只是矽光子學，利用率就不會達到 85%。但也要明白這一點，這一點很重要。我們專注於矽光子學，我們致力於矽光子學，我所說的射頻SOI，如果你願意這麼稱呼的話，在很大程度上是可以互換的電源。我的意思是，有些層雖然不同，但在功率方面相對可互換，在成像方面相對可互換。

The silicon photonics is under different ratios, but it's very fungible to silicon germanium.

矽光子學材料的比例各不相同，但它與矽鍺具有很強的可替代性。

Lisa Thompson, Zacks Investment Research.

Lisa Thompson，Zacks投資研究公司。

I have a few accounting questions for Oren. First off, could you tell us exactly what the dollar amount was for the one-time tax benefit in Q4?

我有一些會計方面的問題想請教奧倫。首先，您能否告知我們第四季一次性稅收優惠的具體金額是多少？

It's approximately the difference between if we had 15% tax or 16% or 17% by the model, which is about from the $81 million pretax income, we should have like a tax expense of about 15% to 17% of that, so about like $12 million, $13 million. Instead of that, we have $1.5 million. So the gap is about $10 million.

這大致相當於，如果我們按模型計算稅率為 15%、16% 或 17%，那麼在 8,100 萬美元的稅前收入中，我們應該有大約 15% 到 17% 的稅收支出，也就是大約 1,200 萬美元到 1,300 萬美元。但我們只有150萬美元。所以差距大約是1000萬美元。

Okay. And can you explain exactly what did you get for the $105 million for the lease extension?

好的。能具體解釋一下，你用1.05億美元續租得到了什麼嗎？

We got additional 3.5 years of lease of Newport Beach facility. We announced on November 13, 2025 press release. Instead of that that it was supposed to be ending in the beginning of '27, it is now until the end of 2030.

我們獲得了紐波特海灘設施額外3.5年的租賃權。我們在2025年11月13日的新聞稿中宣布了這項消息。原本應該在 2027 年初結束，現在卻要持續到 2030 年底。

Okay. And you paid the $105 million upfront cash?

好的。你們一次性支付了1.05億美元現金？

Yes, yes. Yes. And it's included in the cash flow operations of Q4, which is the reason why it is a one-time lower by $105 million than any model. But we announced it in November. So it's not new, no.

是的，是的。是的。這筆款項已計入第四季度的現金流營運中，因此比任何模型都少了 1.05 億美元。但我們在11月就宣布了。所以這並不是什麼新鮮事。

Right, right, right. And then I'm just curious as the change in the US depreciation rules of what you can write off, has that changed your model at all or changed your depreciation expectations going forward?

對，對，對。我很好奇，美國折舊規則的變化，即可以抵扣的項目的變化，是否改變了你們的模型或改變了你們對未來折舊的預期？

No. No impact on us.

不。對我們沒有任何影響。

No, not at all. Okay. Great.

不，完全不是。好的。偉大的。

This concludes the question-and-answer session. So I will now turn back to Russell for closing remarks. Thank you.

問答環節到此結束。現在我將把發言權交還給羅素，請他作總結陳詞。謝謝。

Thank you very much. 2025 marked the completion of my 20th year at Tower. So I thought I would give a little bigger picture view of what Tower is about, where we're going, what we're doing. For the year 2025, we had a corporate theme, and the theme was bold growth, limitless impact, infinite reach. I love that theme and put a lot of thought into it and truly would be my great honor if my life's journey would be worthy to have those words in my epitaph included, obviously, to loving, honorable, loyal, husband, father, grandfather, and friend. But if that was written on my epitaph, wow, what a value-add life I would have led.

非常感謝。 2025年是我在Tower工作的第20年。所以我想從更宏觀的角度介紹一下 Tower 的宗旨、我們的發展方向以及我們正在做的事情。2025 年，我們設定了一個企業主題，那就是大膽成長、無限影響、無限覆蓋。我非常喜歡這個主題，並為此深思熟慮。如果我的人生旅程能夠配得上將這些話語寫入我的墓誌銘，那將是我的莫大榮幸。當然，這些話語是獻給一位充滿愛、正直、忠誠的丈夫、父親、祖父和朋友的。但如果我墓誌銘上寫的是這樣的話，哇，我的一生該是多麼有意義啊！

If you look at bold growth, at least to me, it means being undaunted and creating a legacy much accretive to one's birth situation. In the case directly of corporate leadership, it would mean expanding the enterprise much, much beyond the situation from when one arrived. If you talk about limitless impact, that would mean that the individual or the corporate leader has been successful in importing knowledge and creating opportunities for employees, colleagues, community for one's family to have an advancing growth trajectory much beyond what they otherwise would have had, what otherwise would have been.

在我看來，大膽成長意味著無所畏懼，並創造出遠超自身出身的成就。就企業領導階層而言，這意味著要將企業規模擴大到遠超過你剛上任時的規模。如果影響力是無限的，那就意味著個人或企業領導者成功地引進知識，並為員工、同事、社區甚至家庭創造機會，使他們的發展軌跡遠遠超出他們原本所能達到的水平。

Infinite reach is a very interesting concept. I first encountered the term in David Deutsch's book, The Beginning of Infinity, and the meaning, as he put it forward, meant that truth discovered in any sphere is indeed a truth holds in all spheres. And it is very interesting.

無限延伸是一個非常有趣的概念。我第一次接觸到這個術語是在大衛·多伊奇的著作《無限的開端》中，他提出的意思是，在任何領域發現的真理確實在所有領域都成立。這很有意思。

If you look at learning, there are many things which truly cannot be taught, but rather must be learned and where the learning comes only through doing. And I thought about that quite a bit. It really -- many, many things can be taught, but those things that can be told are tools. Things that can be learned are principles and values, and it really only is learned through the doing.

如果你仔細觀察學習，​​你會發現有很多東西是無法真正教導的，而必須學習，而這種學習只能透過實踐來實現。我仔細考慮過這個問題。確實，很多很多東西都可以教，但那些可以教的東西只是工具。可以學習的是原則和價值觀，而這些只能透過實踐來真正學習。

I have a very, very fervent belief that work is the laboratory where one can and should learn and develop themselves in all capabilities, principles, and values needed to become the person that they aspire to be. Anybody worth their salt spends the bulk of their awakened hours at work. What a meaningless activity if that isn't the place where one develops as a person. And I thought very much that a good company must allow for financial and professional growth, but a great company allows for the same with the addition of personal growth.

我堅信，工作是一個實驗室，人們可以而且應該在這裡學習和發展自己的所有能力、原則和價值觀，從而成為自己渴望成為的人。任何一個稱職的人，醒著的大部分時間都會花在工作上。如果那不是一個人成長的地方，那麼這項活動就毫無意義。我一直認為，一家好的公司必須允許員工獲得財務和職業發展，而一家偉大的公司不僅如此，還允許員工獲得個人發展。

Years back, earlier in my career, I had the great pleasure to reflect and thank Dr. Dan Maydan, at the time the President of Applied Materials, and this is directly what I wrote in. When I came to Applied, I believe I was a good person. Thank you for creating an environment that has allowed me to become a better person. Tower aspires to be such a high -- such a company. We focus on hiring most capable and passionate people and of equal importance to develop and nurture an environment where passionate and capable people can further grow in capability, in passion, and as well in virtue.

多年前，在我職業生涯的早期，我非常榮幸地回顧並感謝了當時的應用材料公司總裁丹·梅丹博士，以下是我當時寫下的內容。當我來到應用科技大學時，我相信我是一個好人。感謝您創造這樣一個環境，​​讓我成為一個更好的人。Tower 渴望成為如此高遠、如此卓越的公司。我們專注於招募最有能力、最有熱情的人才，同樣重要的是，我們致力於發展和培養一個環境，​​讓有熱情、有能力的人才能夠在能力、熱情和品德方面進一步成長。

We acknowledge and drive an understanding that the strongest catalyst for increased capability and enhanced passion or close collaborations with our customers and the excitement enjoy that is earned and truly earned from sharing in each other's successes. There's a quote of uncertain origin. If two people agree on everything, one of them is unnecessary. We treasure diversity. We treasure diversity of opinions, but only if it's directed to singleness in purpose and actions.

我們認識到，提高能力和增強熱情的最強催化劑，是與客戶緊密合作，以及由此獲得的興奮感，而這種興奮感是透過分享彼此的成功而真正獲得的。有一句出處不明的引言。如果兩個人所有事情都意見一致，那麼其中一個就顯得多餘了。我們珍惜多元化。我們珍惜意見的多樣性，但前提是這些意見必須指向同一個目標和行動。

Organizational anarchists do not do very well at Tower. But no matter what and how diverse the opinion is, if it's directed towards making things better, it's highly appreciated. I don't think I'm much different than anyone else. I don't like it when people disagree with me, but I truly value it. And it's a very strong thing, and that's the culture that we have. So we have worked hard to be a company that really does allow people to grow.

組織無政府主義者在塔樓裡表現不佳。但無論意見如何，無論意見多麼多樣化，只要是為了讓事情變得更好，我們都非常感激。我覺得自己和其他人沒什麼不同。我不喜歡別人跟我意見不一致，但我很珍惜這種意見。這是一種非常強大的力量，也是我們所擁有的文化。因此，我們一直努力成為一家真正能讓員工成長的公司。

And if you allow people to grow, you have an environment and a spirit in the company where the company has become truly a masterpiece. Now I don't know the attraction of any single piece of art or music to those on the call, but I can say that I cannot walk by a da Vinci without being drawn to it. That's the impact of a masterpiece. It's the same thing with the company. If a company has extremely passionate people and they're of the highest character and they are capable, knowledgeable people, they are a magnet for the customer, and the customer wants to be with them.

如果你允許員工成長，公司就會有良好的環境和精神，讓公司真正成為傑作。我不知道通話中的每個人對任何藝術品或音樂的吸引力在哪裡，但我可以說，我每次經過達文西的作品都會被它吸引。這就是傑作的影響力。公司的情況也一樣。如果一家公司擁有充滿熱情、品德高尚、能力出眾、知識淵博的員工，那麼他們就會對顧客產生巨大的吸引力，顧客也樂於與他們合作。

And that's what allows for corporate growth. That is one of the things that allows for bold growth that allows a company to have limitless impact, and it's based on the infinite reach of people as soon as you take on big responsibilities, and you take full ownership on those responsibilities, you learn so many truths.

而這正是企業發展壯大的動力。這是實現大膽成長、使公司擁有無限影響力的因素之一，它基於人們的無限影響力。一旦你承擔起重大的責任，並對這些責任完全負責，你就會學到很多真理。

And what is true in one sphere is true in everything. The principles that allow you to be a successful business leader allows you to be a successful father, a successful husband, a successful mother, successful wife, successful son or daughter, successful and value-added friend. So those are the things that Tower has truly worked on, how we continue to work on.

在一個領域成立的真理，在所有領域也成立。讓你成為成功商業領袖的原則，也能讓你成為成功的父親、成功的丈夫、成功的母親、成功的妻子、成功的兒子或女兒，以及成功且有價值的朋友。所以，這些就是 Tower 真正著手做的事情，也是我們繼續努力的方向。

Ex-US President, Bill Clinton, had a quote that, on first hearing, sounds very nice, and it says old age is when your memories outweigh your dreams. I'm not a spring chicken, so those are the type of things that I think about. And at first, again, it sounds very good, but certainly, having dreams in no way define vibrant youth. So the statement maybe is correct as far as if your memories outweigh your dreams, it shows that you're old. But having dreams does not show that you're youthful. Many, many people, even at a young age, only have dreams. They do nothing to try to make the dreams real.

美國前總統比爾柯林頓曾說過一句名言，乍聽之下似乎很不錯，這句話說，老年就是你的記憶超過了你的夢想。我年紀不小了，所以這些都是我會考慮的事情。起初，這聽起來很不錯，但當然，擁有夢想並不能定義充滿活力的青春。所以，這種說法或許是正確的，如果你的記憶比夢想更重要，那就表示你老了。但擁有夢想並不代表你年輕。很多人，即使在很小的年紀，也只有夢想。他們沒有採取任何措施來實現這些夢想。

So I added to this quote and took the liberty. Old age is when your memories outweigh your dreams and consequent actions to achieve them. Tower is in no way an aged company. We work off of the experience and knowledge that comes only through age, but with the full vibrance and excitement of youth, and that's a combination that's unbeatable and is truly a catalyst for customers to want to engage with your company. We showed this new financial model, which as having stated multiple times and being questioned about as well, it's our target to achieve it, be it by run rate or be it in the full year, but to achieve this model in 2028, relatively short term.

所以我對這句話進行了補充，並擅自做了修改。老年時，回憶會超越你的夢想以及為實現夢想而採取的行動。Tower公司絕非老牌公司。我們憑藉著歲月累積的經驗和知識，以及年輕人的活力和熱情開展工作，這種組合是無與倫比的，也是促使客戶想要與貴公司互動的真正催化劑。我們展示了這個新的財務模型，正如我們多次聲明並被問及的那樣，我們的目標是實現這個模型，無論是按運行率還是按全年計算，但我們的目標是在 2028 年實現這個模型，這是一個相對較短的時期。

The model, Oren went through all of the incremental margins, but what it is, it's a revenue CAGR of 22%. It's a very nice CAGR in our industry, a very nice foundry CAGR. So a three-year CAGR of 22%. But it's a net profit CAGR of 50.5%, and that's really incredible to have a 2.5% off of the 22% CAGR, to have a 2.5% increase on the CAGR of the net profit, which isn't something that's talked about often because most people don't have CAGRs on net profit. But from the present state to achieving this model, it's 50.5%.

Oren 的模型考慮了所有增量利潤，但結果是，收入複合年增長率為 22%。在我們產業中，這是一個非常好的複合年增長率，也是鑄造業非常好的複合年增長率。因此，三年複合年增長率為 22%。但淨利潤複合年增長率達到了 50.5%，這真的令人難以置信，比 22% 的複合年增長率提高了 2.5%，淨利潤的複合年增長率提高了 2.5%，這很少被提及，因為大多數人都沒有淨利潤的複合年增長率。但從目前的狀態到實現這一模式，需要 50.5%。

That is not an aged company. That is a company that is full, full of youthful exuberance based upon the capability of age, based upon experience, based upon having developed multiple years of strong, extremely strong relationships with customers.

那不是一家老牌公司。這是一家充滿活力、朝氣蓬勃的公司，這種活力源於其年齡、經驗以及多年來與客戶建立的牢固、極其牢固的關係。

So to close, we entered 2026 with very strong momentum towards bold growth, limitless impact, and infinite reach. Thank you for being with us. Thank you for continuing to be with us as we track towards the achievement of the $750 million net profit model. Thank you very much.

最後，我們帶著強勁的成長勢頭進入了 2026 年，朝著大膽成長、無限影響和無限覆蓋的方向邁進。謝謝你們的陪伴。感謝您一直以來與我們攜手並進，共同朝著實現 7.5 億美元淨利潤目標邁進。非常感謝。

This concludes today's conference call. Thank you all for participating. You may now disconnect your lines. Thank you.

今天的電話會議到此結束。感謝大家的參與。現在您可以斷開線路了。謝謝。