格芯 (GFS) 2026 Q3 法說會逐字稿

Thank you for standing by, and welcome to the GLOBALFOUNDRIES third-quarter fiscal year 2026 financial results conference call. (Operator Instructions) As a reminder, today's program is being recorded.

感謝各位稍候，歡迎參加 GLOBALFOUNDRIES 2026 會計年度第三季財務業績電話會議。(接線員指示) 提醒各位，今天的會議將被錄音。

And now I'd like to introduce your host for today's program, Eric Chow, Head of Investor Relations. Please go ahead, sir.

現在我想介紹今天會議的主持人，投資人關係主管 Eric Chow。先生，請開始。

Thank you, operator. Good morning, everyone, and welcome to GLOBALFOUNDRIES Third Quarter 2026 Earnings Call. On the call with me today are Tim Breen, CEO; and Sam Franklin, CFO. A short while ago, we released GS Third Quarter 2026 financial results which are available on our website at investors.gf.com, along with today's accompanying slide presentation. This call is being recorded, and a replay will be made available on our Investor Relations web page.

謝謝接線員。各位早安，歡迎參加 GLOBALFOUNDRIES 2026 年第三季法說會。今天與我一同出席的有執行長 Tim Breen，以及財務長 Sam Franklin。稍早我們已發布 GS 2026 年第三季財務結果，可於我們網站 investors.gf.com 查閱，並附有今天的簡報投影片。本次電話會議將被錄音，重播將於我們投資人關係網頁提供。

During this call, we will present both IFRS and non-IFRS financial measures. The most directly comparable IFRS measures and reconciliations for non-IFRS measures are made available on today's press release and accompanying slides. Please note that these financial results are unaudited and subject to change. Certain statements on today's call may be deemed to be forward-looking statements. Such statements can be identified by terms such as believe, expect, intend, anticipate, and may or by the use of the future tenants.

在本次電話會議中，我們將同時呈現 IFRS 與非 IFRS 財務衡量指標。與非 IFRS 指標最直接可比的 IFRS 指標，以及非 IFRS 指標的調節表，已於今日新聞稿與隨附投影片中提供。請注意，這些財務結果尚未經審計，且可能變更。今天電話會議中的某些陳述可能被視為前瞻性陳述。此類陳述可由 believe、expect、intend、anticipate、may 等用語，或使用未來時態來辨識。

You should not place undue reliance on forward-looking statements. Actual results may differ materially from these forward-looking statements, and we do not undertake any obligation to update any forward-looking statements we make today.

您不應過度依賴前瞻性陳述。實際結果可能與這些前瞻性陳述有重大差異，我們亦不承擔更新今日所作任何前瞻性陳述之義務。

For more information about factors that may cause actual results to differ materially from forward-looking statements, please refer to the press release we issued today as well as risks and uncertainties described in our SEC filings, including in sections under the caption Risk Factors in our annual report on Form 20-F and in any current reports on Form 6-K furnished with the SEC. In terms of upcoming events, we will be participating in a fireside chat at the UBS Global Technology and AI Conference in Scottsdale, Arizona on December 1. We will begin today's call with Tim providing a summary update on the business environment and technologies, followed by Sam, who will provide details on our third quarter results and fourth quarter guidance.

關於可能導致實際結果與前瞻性陳述有重大差異之因素的更多資訊，請參閱我們今日發布的新聞稿，以及我們向 SEC 提交文件中所描述的風險與不確定性，包括 20-F 年報中「風險因素」章節，以及向 SEC 提交的任何 6-K 當期報告。就近期活動而言，我們將於 12 月 1 日在亞利桑那州史考茲戴爾（Scottsdale）舉行的 UBS 全球科技與 AI 大會參與爐邊談話。今天的電話會議將先由 Tim 就商業環境與技術提供摘要更新，接著由 Sam 說明第三季業績與第四季展望指引。

We will then open the call for questions with Tim and Sam. We request that you please limit your questions to one with one follow-up. I'll now turn the call over to Tim.

之後我們將開放 Tim 與 Sam 回答提問。請各位將問題限制為一題，並可追加一個追問。接下來我把電話交給 Tim。

Thank you, Eric, and welcome, everyone, to our third quarter 2026 earnings call. GF delivered strong results in the third quarter with revenue and non-IFRS profitability metrics at the high end of our guidance ranges. We delivered record non-IFRS gross margin, which demonstrates the meaningful progress we are making to enhance the mix of our business by ramping our differentiated technologies in high-value, high-growth applications.

謝謝你，Eric，也歡迎各位參加我們 2026 年第三季法說會。GF 在第三季交出強勁成果，營收與非 IFRS 獲利能力指標皆落在我們指引區間的高端。我們創下非 IFRS 毛利率新高，顯示我們透過在高價值、高成長應用中加速導入差異化技術，正持續且顯著地改善業務組合。

Our communications infrastructure and data center end market saw record levels of growth, led by accelerating demand in optical networking across our silicon photonics and silicon germanium platforms. We also delivered strong growth in automotive and home and industrial IoT, reflecting broad-based secular momentum across our key markets. Thanks to the accelerated growth across several end markets, the proportion of revenue outside of Smart mobile devices was the highest level in our company's history and is driving a richer, more diversified, more resilient revenue base, thanks to the team's excellent efforts and execution we are ahead of plans to deliver the strategic and financial targets we set out at our Investor Day earlier in the year.

我們的通訊基礎建設與資料中心終端市場成長創下新高，主要由矽光子與矽鍺（SiGe）平台在光網路領域需求加速所帶動。我們在車用以及家用與工業物聯網（IoT）亦實現強勁成長，反映我們核心市場普遍的長期結構性動能。受惠於多個終端市場的加速成長，智慧型行動裝置以外的營收占比達到公司史上最高水準，並在團隊卓越努力與執行下，推動更高品質、更分散、也更具韌性的營收基礎；我們正超前進度，朝今年稍早投資人日所設定的策略與財務目標邁進。

We will look back at 2026 being a pivotal year for GF. In recent weeks, we have made several significant announcements that meaningfully advance our strategy, support our financial objectives and reinforce our foundation for durable long-term growth. We are proud to have announced a significant number of new customer and investment partnerships such as our agreements with Marvell to accelerate capacity for our industry-leading high-performance SiGe with monolithic power for critical power management solutions with TSMC to establish a US-based supply of silicon interposers, for the CoWos ecosystem and the Landmark US Japan fab expansion plan that will transform our multi-new York site. These milestones firmly position GF as a diversified holistic technology provider critical to governments and partners around the world.

回顧 2026 年，這將是 GF 的關鍵轉捩點。近幾週我們發布多項重大公告，實質推進我們的策略、支持財務目標，並強化我們可持續長期成長的基礎。我們很自豪宣布多項新的客戶與投資合作夥伴關係，例如與 Marvell 的協議，以加速我們業界領先、具單晶整合電源（monolithic power）的高效能 SiGe 產能，用於關鍵電源管理解決方案；與台積電（TSMC）合作建立美國在地的矽中介層（silicon interposers）供應，以支援 CoWoS 生態系；以及具里程碑意義的美日晶圓廠擴建計畫，將改造我們位於紐約州莫爾他（Malta, New York）的廠區。這些里程碑使 GF 穩固定位為多元且全方位的技術供應商，對全球各地政府與合作夥伴皆至關重要。

Let me cover these areas in more detail, beginning with our strong investment partnerships announced last month, the US Japan initiative provides a new and innovative model for GF to expand our critical US semiconductor manufacturing capacity at accretive economic returns. This initiative will unlock substantial capacity growth, support significant future customer demand and bring the scale of our US operations in Mother New York to mega fab level.

我將更詳細說明上述領域，先從我們上月宣布的強勁投資合作夥伴關係談起。美日倡議為 GF 提供一種全新且創新的模式，使我們能以具增益的經濟報酬擴充美國關鍵半導體製造產能。此倡議將釋放可觀的產能成長、支援未來顯著的客戶需求，並使我們在紐約州莫爾他（Malta, New York）的美國營運規模提升至超級晶圓廠（mega fab）等級。

The unique and innovative partnership structure allows GF to expand capacity faster without investing capital with an economic model that is accretive to our long-range plans. Through a robust structure of technology, construction and operations fees, cost reimbursements and EBITDA sharing, the initiative is designed to generate the profit per wafer comparable if we have built and own the capacity ourselves.

此獨特且創新的合作架構，使 GF 能在不投入資本的情況下更快擴充產能，且其經濟模型可為我們的長期計畫帶來增益。透過技術、建設與營運費用、成本補償以及 EBITDA 分潤等健全架構，該倡議的設計目標是產生與我們自行建置並擁有該產能時相當的每片晶圓利潤。

As another indicator of our deepening partnership with the US government, GF reached an agreement with the US Department of Commerce to reallocate USD955 million of our existing chips and Science Act award to accelerate investments at our Multi New York and Burlington, Vermont facilities. The amended award will accelerate CapEx and R&D investments in our highest value differentiated technologies that our customers demand such as silicon photonics, silicon germanium and advanced packaging, including the first US-based source of silicon interposers.

作為我們與美國政府夥伴關係深化的另一項指標，GF 已與美國商務部達成協議，將我們既有《晶片與科學法案》（CHIPS and Science Act）獎助中的 9.55 億美元重新分配，以加速對我們紐約州莫爾他（Malta, New York）與佛蒙特州伯靈頓（Burlington, Vermont）廠區的投資。修訂後的獎助將加速資本支出（CapEx）與研發（R&D）投入於客戶所需求、最高價值的差異化技術，例如矽光子、矽鍺以及先進封裝，其中包括美國首個在地矽中介層供應來源。

In recognition of the growing role that GF playing in the US semiconductor ecosystem a separate agreement will be entered into to issue 4% equity to the US government, which provides a vehicle for the American people to share GF's future success. Combined with previous announcements across the quantum, silicon photonics and advanced packaging investments, we anticipate the total equity ownership of the US government to be approximately 7% when all agreements are finalized.

為表彰 GF 在美國半導體生態系中日益重要的角色，將另行簽訂協議向美國政府發行 4% 股權，使美國民眾得以分享 GF 未來的成功。結合先前在量子、矽光子與先進封裝投資方面的公告，我們預期在所有協議完成後，美國政府的總股權持有比例約為 7%。

Having attained a meaningful position in GF, our Board of Directors and our US government partners do not plan on any further government equity position beyond these announced equity stakes. In summary, this landmark agreement represents a strong vote of confidence in and underscores our critical role in advancing US semiconductor leadership.

在美國政府已取得 GF 的重要持股後，我們董事會與美國政府合作夥伴不計畫在上述已宣布的股權持有之外，再進一步增加政府持股。總結而言，這項具里程碑意義的協議代表對我們的高度信心，也凸顯我們在推進美國半導體領導地位方面的關鍵角色。

Let me now turn to our significant and expanding opportunity in the AI data center. We see significant momentum with customers for our optical networking solutions, where customers value GF's unique combination of high-performance photonic devices, advanced fiber coupling, RF integration and advanced packaging, backed by more than a decade of investment and substantial clean room space within our modern, highly automated 300-millimeter footprint, GF can deliver differentiated performance and scale capacity rapidly as customer demand grows.

接下來我談談我們在 AI 資料中心領域重大且持續擴大的機會。我們看到客戶對光網路解決方案的動能顯著提升；客戶重視 GF 獨特的組合能力，包括高效能光子元件、先進光纖耦合、RF 整合與先進封裝。這些能力建立在十多年的投資基礎之上，並結合我們現代化、高度自動化的 300 毫米產線所具備的大量無塵室空間，使 GF 能提供差異化效能，並在客戶需求成長時快速擴充產能。

Our competitive advantages are translating into strong design win momentum and robust customer demand. We are actively engaged with more than 50 silicon photonics customers, including 7 of the top 10 optical transceiver suppliers with active designs slotted to ramp in 2027 and 2028 to scale across, scale out and scale up networks. In 2026, we expect to deliver $0.5 billion of revenue from silicon photonics and on the back of the strength of customer demand, order visibility and our capacity growth plans, we now expect to exit 2027 at an annualized silicon photonics revenue run rate of approximately USD1 billion a full year ahead of prior expectations.

我們的競爭優勢正轉化為強勁的設計導入（design win）動能與穩健的客戶需求。我們正積極與超過 50 家矽光子客戶合作，其中包括前十大光收發器供應商中的 7 家；其活躍設計已排定於 2027 與 2028 年放量，以支援 scale across、scale out 與 scale up 網路的擴展。在 2026 年，我們預期矽光子營收可達 5 億美元；基於強勁的客戶需求、訂單能見度與產能成長規劃，我們現在預期於 2027 年底的矽光子年化營收 run rate 約為 10 億美元，較先前預期提前整整一年達成。

To support the ramping demand for optical networking in 2028 and beyond, we plan to approximately triple our silicon photonics capacity over the course of next year, which Sam will cover in more detail. In addition to silicon photonics, silicon germanium is another important and complementary driver of GS growth in optical networking.

為了支援 2028 年及以後光學網路需求的加速成長，我們計劃在明年期間將矽光子產能約提升至三倍，Sam 將會更詳細說明。除了矽光子之外，矽鍺（SiGe）也是光學網路中推動 GS 成長的另一項重要且互補的驅動因素。

Our SiGe technology enables high-speed optical transceivers through transmit drivers and receive amplifiers, key interface elements between the electrical and optical domains. Customer demand for our SiGe technology continues to increase beyond current capacity and supported by customer prepayments, we are accelerating our capacity investments to ensure that we can meet this demand.

我們的 SiGe 技術透過發射驅動器與接收放大器實現高速光學收發器，這些是電域與光域之間的關鍵介面元件。客戶對我們 SiGe 技術的需求持續超出目前產能；在客戶預付款的支持下，我們正加速產能投資，以確保能滿足此需求。

Our recently announced multiyear strategic partnership with Marvell will help support significant SiGe capacity expansion in our Burlington fab to enable next-generation optical solutions. Together with the reallocated US government funding from our prior chip France awards, this gives us additional resources and strong customer alignment to scale a highly differentiated platform, meaningfully accretive to GF's long-term margin targets.

我們近期宣布與 Marvell 的多年期策略合作夥伴關係，將有助於支持我們在 Burlington 晶圓廠大幅擴充 SiGe 產能，以實現下一代光學解決方案。再加上我們先前法國晶片獎助案所重新分配的美國政府資金，這為我們提供了額外資源與強勁的客戶一致性，以擴大高度差異化的平台規模，並對 GF 的長期毛利率目標帶來顯著增益。

Lastly, advanced packaging is another key enabler of our AI data center strategy. Under our five-year USD2 billion manufacturing agreement with TSMC GF will establish a US-based supply of silicon interposers for the CoWoS ecosystem, with volume production expected to ramp at our multisite in the first half of 2028 and with a framework to expand the agreement over time. The economics of this arrangement are attractive for GF and the strategic benefits are clear. This agreement establishes a new multiyear revenue stream for GF, with gross margins above our long-term targets and with co-investment from TSMC that will help fund the necessary capital associated with this project.

最後，先進封裝是我們 AI 資料中心策略的另一項關鍵推動因素。根據我們與台積電（TSMC）的五年 20 億美元製造協議，GF 將為 CoWoS 生態系建立美國本土的矽中介層供應，預期在 2028 年上半年於我們的多據點產線開始放量爬坡，並具備隨時間擴大協議的框架。此安排對 GF 的經濟效益具吸引力，策略性效益也十分明確。該協議為 GF 建立新的多年期營收來源，其毛利率高於我們的長期目標，且台積電的共同投資將有助於支應此專案所需的資本支出。

The silicon interposers are yet another way GF can serve growing AI and high-performance computing infrastructure demand by combining GF trusted US manufacturing capabilities with a leading advanced packaging ecosystem globally, we are expanding into another essential layer of AI infrastructure and supporting the tighter integration of compute, memory and connectivity. We believe we are in the early innings of GF AI data center opportunity underpinned by the critical needs of our customers for higher bandwidth connectivity, tighter integration of compute and memory, greater power efficiency and secure scalable supply. The result for GF, more avenues to win greater content and value in each system and a clear path to growth at a structurally richer business mix.

矽中介層是 GF 服務日益成長的 AI 與高效能運算基礎設施需求的另一種方式；透過結合 GF 可信賴的美國製造能力與全球領先的先進封裝生態系，我們正擴展至 AI 基礎設施的另一個關鍵層級，並支持運算、記憶體與連結性的更緊密整合。我們相信，GF 在 AI 資料中心的機會仍處於早期階段，其基礎來自客戶對更高頻寬連結性、更緊密的運算與記憶體整合、更高電源效率，以及安全且可擴充供應的關鍵需求。對 GF 而言，結果是：在每個系統中有更多途徑贏得更高的內容占比與價值，並在結構性更優質的業務組合下具備清晰的成長路徑。

Let me turn to physical AI as well as the meaningful progress we are making across IP, software and custom silicon or IFC. GF's physical AI opportunity spans a wide variety of applications across connectivity, compute, power actuation and sensing and our IP, software and custom silicon capabilities unlock even more ways for GF to win. We believe the deployment of intelligence into vehicles, factories, robots, drones, satellites, medical devices and other edge systems represents a significant and long-duration demand vector for GF.

接下來我談談實體 AI，以及我們在 IP、軟體與客製化矽（或 IFC）方面所取得的顯著進展。GF 的實體 AI 機會涵蓋連結、運算、功率致動與感測等多元應用，而我們的 IP、軟體與客製化矽能力也為 GF 帶來更多勝出方式。我們相信，將智慧部署到車輛、工廠、機器人、無人機、衛星、醫療裝置及其他邊緣系統，代表著一個規模可觀且持續時間長的需求動能。

Together with the Synopsys ARC processor IP business we acquired earlier this year, MPS expands our role into that of a holistic technology partner, servicing a broader set of customers with our custom silicon platforms. MIPS recently announced three physical AI-native reference platforms, focusing on compute, AI inference and real-time control, and we are now working with several customers designing systems focused on physical AI, leveraging our expanded capabilities. The combination of Risk 5 compute and AI design capability and AI native platforms based on GF's broad manufacturing capabilities puts us in a unique position to accelerate share gain in this area.

結合我們今年稍早收購的 Synopsys ARC 處理器 IP 業務，MPS 使我們的角色擴展為全方位的技術合作夥伴，透過客製化矽平台服務更廣泛的客戶群。MIPS 近期宣布三個以實體 AI 原生為核心的參考平台，聚焦於運算、AI 推論與即時控制；我們目前正與多家客戶合作設計以實體 AI 為重點的系統，並運用我們擴增後的能力。結合 RISC-V 運算與 AI 設計能力，以及基於 GF 廣泛製造能力的 AI 原生平台，使我們處於獨特位置，可加速在此領域的市占提升。

Just over one year after we initially acquired MIPS, we are very encouraged by the strong customer traction, which is now driving several platform-level wins. During the third quarter alone, we secured custom silicon design wins that totaled nearly $1 billion of lifetime revenue. These included platform wins for our RADAR system on chip and a risk-based processor core designed for real-time control. In addition, we secured a win with a leading provider of high-performance data center computing solutions to design their custom Board management controller system-on-chip, integrating our processor technology with third-party IP.

在我們最初收購 MIPS 之後僅一年多，我們對強勁的客戶採用動能感到非常振奮，這正推動多項平台層級的勝出。僅在第三季，我們取得的客製化矽設計案合計帶來接近 10 億美元的全生命週期營收。其中包括我們 RADAR 系統單晶片（SoC）的平台勝出，以及一款為即時控制所設計、以 RISC 為基礎的處理器核心。此外，我們也與一家領先的高效能資料中心運算解決方案供應商取得合作勝出，為其設計客製化的板級管理控制器（BMC）系統單晶片，將我們的處理器技術與第三方 IP 進行整合。

We also secured a notable aerospace and defense win to support the engine control processor for a leading global aerospace and propulsion company. Ultimately, these wins demonstrate how GF can capture more opportunities and support our customers across the semiconductor value chain, deepen our relationships with industry partners and convert initial design engagements into concrete long-duration growth opportunities.

我們也取得一項重要的航太與國防領域勝出，為一家全球領先的航太與推進公司提供引擎控制處理器的支援。最終，這些勝出展現了 GF 如何在半導體價值鏈中捕捉更多機會並支援客戶、深化與產業夥伴的關係，並將初期的設計合作轉化為具體且長期的成長機會。

In summary, the multifaceted progress we demonstrated this quarter shows our strategy in action. We are scaling differentiated technologies into high-growth AI markets and driving meaningful customer value with our capabilities in software and custom silicon. We are transforming GF into a more diversified, more durable and more profitable business, propelled by the defining megatrends of our time.

總結來說，我們本季所展現的多面向進展，顯示我們的策略正在落地執行。我們正將差異化技術擴展至高成長的 AI 市場，並透過我們在軟體與客製化矽方面的能力為客戶創造顯著價值。在當代關鍵巨型趨勢的推動下，我們正把 GF 轉型為更具多元性、更具韌性且更高獲利能力的企業。

I'll now pass the call over to Sam for a deeper dive on third quarter 2026 financials.

接下來我把電話交給 Sam，請他更深入說明 2026 年第三季財務表現。

Thank you, Tim. For the remainder of the call, including guidance other than revenue, cash flow and net interest income, I'll reference non-IFRS metrics. The third quarter was a significant milestone for GF, not only for the continued progression in our profitability metrics, but also for the continued mix shift in our portfolio. For the first time in our history, more than two-thirds of our Q3 revenue came from outside Smart Mobile devices, a record for GF and further evidence of our increasingly diversified and margin-accretive mix.

謝謝你，Tim。在接下來的電話會議中，除營收、現金流與淨利息收入之外的指引，我將引用非 IFRS 指標。第三季對 GF 而言是一個重要里程碑，不僅因為我們獲利能力指標持續改善，也因為我們產品組合持續發生結構性轉移。在我們歷史上首次，第三季超過三分之二的營收來自智慧行動裝置以外的領域，這是 GF 的新紀錄，也進一步證明我們的組合正日益多元化，且對毛利率具有增益。

We delivered record non-IFRS gross margin while continuing to fund the strategic investments that will drive GF's next phase of growth. Our strategic decisions are already enhancing our earnings power, and we believe that we're only in the early innings of this evolution.

我們創下非 IFRS 毛利率新高，同時持續投入將推動 GF 下一階段成長的策略性投資。我們的策略決策已在提升獲利能力，我們相信這場演進仍處於早期階段。

Now on to the results. We delivered third quarter revenue of $ 1.899 billion, up 6% sequentially and 13% year-over-year. We shipped approximately 647,300-millimeter equivalent wafers in the quarter, up 4% sequentially and 7% from the prior year period. Revenue from manufacturing services accounted for approximately 88% of total revenue. Revenue from Technology Services, which includes revenue from IP licensing, software, reticles, non-recurring engineering, expedite fees and other items accounted for approximately 12% of total revenue for the third quarter.

接著看業績結果。我們第三季營收為 18.99 億美元，季增 6%，年增 13%。本季出貨約 647,300 片（以 300 毫米等效晶圓計），季增 4%，較去年同期增加 7%。製造服務營收約占總營收的 88%。技術服務營收（包含 IP 授權、軟體、光罩、一次性工程費用、加急費及其他項目）約占第三季總營收的 12%。

As we continue to invest in a broader range of capabilities for our customers, we expect Technology Services revenue will represent approximately 12% of total revenue in 2026, which is at the upper end of the range we indicated and with a gross margin profile that continues to be above our long-term target. Beyond 2026, we expect that the growth of revenue from technology services associated with IP licensing and software combined with strong momentum in customer design wins and tapeouts will be important drivers of GF's revenue growth and gross margin appreciation.

隨著我們持續為客戶投資更廣泛的能力，我們預期 2026 年技術服務營收將約占總營收的 12%，位於我們先前所指引區間的上緣，且其毛利率結構仍將高於我們的長期目標。在 2026 年之後，我們預期與 IP 授權與軟體相關的技術服務營收成長，加上客戶設計案與投片（tapeout）的強勁動能，將成為推動 GF 營收成長與毛利率提升的重要驅動因素。

Let me now provide an update on our revenue and outlook by end markets. Communications infrastructure and data center represented approximately 17% of third quarter total revenue. Revenue increased 14% sequentially and 81% year-over-year, driven by robust growth in optical networking which includes silicon photonics and silicon germanium. Not only did Q3 represent record year-over-year growth for this end market, it also marked the eighth consecutive quarter of double-digit percentage year-over-year growth for communications infrastructure and data center.

接下來我將依終端市場提供我們營收與展望的最新情況。通訊基礎設施與資料中心約占第三季總營收的 17%。在光學網路（包含矽光子與矽鍺）強勁成長帶動下，該領域營收季增 14%，年增 81%。第三季不僅創下此終端市場的年增率新高，也標誌著通訊基礎設施與資料中心連續第八季維持年增率達雙位數百分比。

With accelerating customer demand and improved manufacturing productivity, our expectation is that revenue from communications infrastructure and data center will grow in the range of 60% to 70% in 2026. Furthermore, our design win momentum across this end market has continued to strengthen. In the third quarter, we secured design wins for both silicon photonics and SiGe and a new optical transceiver customer for hyperscaler demand.

隨著客戶需求加速以及製造生產力提升，我們預期通訊基礎設施與資料中心的營收在 2026 年將成長 60% 至 70% 的區間。此外，我們在此終端市場的設計導入（design win）動能持續增強。在第三季，我們於矽光子（silicon photonics）與 SiGe 皆取得設計導入，並新增一位光學收發器客戶以滿足超大規模雲端業者（hyperscaler）的需求。

In data center power, we take out a smart power stage gate driver on our BCD platform and secured a design win on our 650-volt GaN platform. Automotive represented approximately 19% of third quarter total revenue. Automotive revenue increased 11% sequentially and 20% year-over-year. Consistent with our prior commentary, our expectation is that automotive revenue will grow low double-digit percentage in 2026, marking another strong year of growth for this end market, and GF continues to win with our customers to support the expansion of automotive semiconductor content.

在資料中心電源方面，我們在 BCD 平台推出智慧功率級閘極驅動器，並在 650 伏 GaN 平台上取得設計導入。汽車業務約占第三季總營收的 19%。汽車營收季增 11%，年增 20%。與我們先前的評論一致，我們預期 2026 年汽車營收將以低雙位數百分比成長，為此終端市場再創強勁成長的一年；同時，GF 持續與客戶贏得更多機會，以支持汽車半導體含量的擴張。

Smart Mobile devices represented approximately 32% of third quarter total revenue, the lowest proportion in our history and a step forward to our long-term strategic goal to enhance the diversification of mix. Revenue decreased 5% sequentially and 19% from the prior year period. Consistent with the commentary from our customers, we continue to expect smart mobile devices to decline by a low teens percentage year-over-year in 2026.

智慧行動裝置約占第三季總營收的 32%，為我們歷史上最低占比，亦是朝向我們長期策略目標—提升產品組合多元化—邁進的一步。營收季減 5%，較去年同期減少 19%。與客戶的評論一致，我們仍預期 2026 年智慧行動裝置將年減低十位數百分比（low teens）。

We continue to have a skew towards the premium end of the smart mobile market, and we expect to largely track the broader demand dynamics as the market adjusts to the challenges associated with memory-related supply constraints. The long-term demand indicators continue to be positive. In the third quarter, we secured new design wins across form factors spanning both smartphones and smart glasses. These included additional content wins in mobile power and a first-ever design win for field gauge ICs in premium tier handsets. Customer engagement continues to be strong for GF's differentiated and high-performance RF, power, audio, display and imaging capabilities to service the demand for smart and connected mobile devices.

我們在智慧行動市場仍偏向高階（premium）端，並預期在市場因記憶體相關供應限制所帶來的挑戰而調整之際，我們的表現將大致追隨整體需求動態。長期需求指標仍持續正向。在第三季，我們在多種外形規格（form factor）上取得新的設計導入，涵蓋智慧型手機與智慧眼鏡。其中包括在行動電源領域取得更多內容（content）導入，以及在高階機種中首次取得用於場域量測（field gauge）的 IC 設計導入。客戶對 GF 具差異化且高效能的 RF、電源、音訊、顯示與影像能力之合作互動仍然強勁，以服務智慧且互聯的行動裝置需求。

Finally, home and industrial IoT represented approximately 20% of third quarter total revenue, revenue increased 16% sequentially and 49% year-over-year on broad-based strength across connectivity, tracking and sensing applications. As workloads for edge AI applications continue to grow, alongside the requirement for Bluetooth wireless, cellular and power management components to perform at higher standards, we have seen strong traction with our customers in this end market for GF's highly differentiated platforms. As a result, we continue to expect Home & Industrial IoT to grow in the range of 10% to 15% in 2026.

最後，家用與工業 IoT 約占第三季總營收的 20%；在連線、追蹤與感測應用的廣泛強勁帶動下，營收季增 16%，年增 49%。隨著邊緣 AI 應用的工作負載持續成長，加上藍牙無線、蜂巢式通訊與電源管理元件需達到更高標準的要求，我們在此終端市場看到客戶對 GF 高度差異化平台的強勁採用動能。因此，我們仍預期 2026 年家用與工業 IoT 將成長 10% 至 15%。

Looking ahead to the future growth opportunities, the design win momentum has continued. In the third quarter, we taped out next-generation cellular IoT and Bluetooth products on our FDX platform for wireless connectivity. We're encouraged by the momentum we're seeing across the end markets we serve, and believe keep us on track with the strategic diversification objectives that we set out during our Investor Day this year.

展望未來的成長機會，設計導入動能持續延伸。在第三季，我們在 FDX 平台上完成下一代蜂巢式 IoT 與藍牙產品的流片（tape-out），用於無線連線。我們對於在所服務的各終端市場看到的動能感到鼓舞，並相信這將使我們持續朝今年投資人日所提出的策略性多元化目標前進。

Moving now to other key financial performance metrics in the quarter. In the third quarter, we delivered gross profit of $597 million, which translates into approximately 31.4% gross margin, up 540 basis points year-over-year. Q3 gross margin achieved a company record and was at the high end of our guidance range. R&D for the quarter was $164 million, and SG&A was $93 million. Total operating expenses of $257 million represented approximately 14% of total revenue and were up 9% quarter-over-quarter, reflecting recent acquisitions.

接著說明本季其他關鍵財務績效指標。第三季我們的毛利為 5.97 億美元，約當毛利率 31.4%，年增 540 個基點。第三季毛利率創公司紀錄，且位於我們財測區間的高端。本季研發費用為 1.64 億美元，銷售、一般及行政費用（SG&A）為 9,300 萬美元。總營業費用為 2.57 億美元，約占總營收 14%，季增 9%，反映近期併購影響。

We delivered operating profit of $340 million for the quarter at an operating margin of 17.9% at the high end of our guided range and up 250 basis points from the prior year period. Third quarter net interest income was $9 million and other expense was $6 million, and we incurred tax expense of $45 million in the quarter. We delivered third quarter net income of approximately $298 million, an increase of approximately $66 million from the prior year period. Diluted earnings of $0.53 per share was at the high end of the guidance range based on a fully diluted share count of approximately 558 million shares.

本季我們的營業利益為 3.40 億美元，營業利益率 17.9%，位於財測區間高端，並較去年同期提升 250 個基點。第三季淨利息收入為 900 萬美元，其他費用為 600 萬美元，且本季認列所得稅費用 4,500 萬美元。第三季淨利約為 2.98 億美元，較去年同期增加約 6,600 萬美元。以完全稀釋後股數約 5.58 億股計算，稀釋後每股盈餘 0.53 美元，位於財測區間高端。

Let me now provide some key cash flow and balance sheet metrics. Cash flow from operations in the third quarter was $341 million. Third quarter CapEx net of proceeds from government grants was $205 million or roughly 11% of revenue. Adjusted free cash flow for the quarter was $136 million. Free cash flow in the third quarter benefited from approximately $150 million of customer prepayments and government grants received.

接下來提供幾項關鍵現金流與資產負債表指標。第三季營運活動現金流為 3.41 億美元。第三季資本支出（CapEx）扣除政府補助款入帳後為 2.05 億美元，約占營收 11%。本季調整後自由現金流為 1.36 億美元。第三季自由現金流受惠於約 1.5 億美元的客戶預付款及收到的政府補助款。

At the end of the third quarter, our combined total of cash and cash equivalents and marketable securities stood approximately $3.2 billion. Our total debt was $1.1 billion and we also have a $1.5 billion revolving credit facility, which remains undrawn. On October 9, we paid a quarterly cash dividend of $0.12 per share. On October 27, the Board of Directors approved a dividend of $0.12 per share payable on January 14, 2027, to shareholders of record as of December 22, 2026. In the third quarter, we also repurchased $100 million of our shares.

截至第三季末，我們的現金及約當現金與有價證券合計約為 32 億美元。總負債為 11 億美元，另有 15 億美元的循環信用額度，目前仍未動用。10 月 9 日，我們支付每股 0.12 美元的季度現金股利。10 月 27 日，董事會核准每股 0.12 美元的股利，將於 2027 年 1 月 14 日支付給截至 2026 年 12 月 22 日登記在冊的股東。第三季我們亦回購 1 億美元的公司股票。

Let me now provide you with our outlook for the fourth quarter of 2026. We expect total GF revenue to be $2 billion, plus or minus $25 million. We expect gross margin to be approximately 31.5%, plus or minus 100 basis points, which at the midpoint reflects approximately 250 basis points of year-over-year expansion. Excluding share-based compensation, we expect total operating expenses to be $235 million, plus or minus $10 million. We expect operating margin to be in the range of 19.8%, plus or minus 170 basis points.

接下來提供我們對 2026 年第四季的展望。我們預期 GF 總營收為 20 億美元，上下浮動 2,500 萬美元。我們預期毛利率約為 31.5%，上下浮動 100 個基點；以中位數計算，代表年增約 250 個基點的擴張。扣除股份基礎給付（share-based compensation）後，我們預期總營業費用為 2.35 億美元，上下浮動 1,000 萬美元。我們預期營業利益率介於 19.8% 左右，上下浮動 170 個基點。

At the midpoint of our guidance, we expect share-based compensation to be approximately $68 million of which roughly $16 million is related to cost of goods sold. We expect net interest and other income for the quarter to be between negative $5 million and $1 million. and income tax expense to be between $44 million and $68 million.

以財測中位數計算，我們預期股份基礎給付約為 6,800 萬美元，其中約 1,600 萬美元與銷貨成本相關。我們預期本季淨利息及其他收入介於負 500 萬美元至正 100 萬美元之間。所得稅費用預期介於 4,400 萬美元至 6,800 萬美元之間。

Based on a fully diluted share count of approximately 575 million shares, which reflects the expected issuance of equity to the US government, we expect diluted earnings per share to be $0.59, plus or minus $0.05. Subject to the timing of the closures associated with the remaining equity issuances to the US government we expect that our fully diluted share count will be approximately 600 million shares commencing 2027.

以完全稀釋後股數約 5.75 億股計算（反映預期向美國政府發行股權），我們預期稀釋後每股盈餘為 0.59 美元，上下浮動 0.05 美元。視向美國政府進行其餘股權發行相關交割（closures）的時點而定，我們預期自 2027 年起完全稀釋後股數將約為 6.00 億股。

Now let me provide an update on some broader financial drivers, milestones and outlook. With respect to pricing, the third quarter saw meaningful blended ASP growth, both sequentially and year-over-year. This marked the first quarter of year-over-year blended pricing improvements since 2023 and is consistent with the commentary we provided on our last earnings call, especially as we execute the continued mix shift of our business.

接下來我將更新一些更廣泛的財務驅動因素、里程碑與展望。就定價而言，第三季混合平均售價（blended ASP）出現顯著成長，無論是季比或年比皆然。這是自 2023 年以來首次出現混合定價年比改善的一季，亦與我們在上次法說會所提供的評論一致，特別是在我們持續推動業務組合轉移（mix shift）的過程中。

With respect to tax, we continue to expect a non-IFRS effective tax rate in the mid-teens percentage range for the full year 2026. We Net CapEx for 2026 is expected to be approximately 15% to 20% of revenue, which remains below our long-term model, and we continue to expect free cash flow margin of approximately 10% in 2026.

就稅務而言，我們仍預期 2026 全年非 IFRS 有效稅率將落在十幾個百分點的中段區間。我們預期 2026 年淨資本支出（Net CapEx）約為營收的 15% 至 20%，仍低於我們的長期模型，且我們仍預期 2026 年自由現金流利潤率約為 10%。

Touching briefly on the CapEx outlook for the year ahead, we expect 2027 to represent a year of accelerated investment for GF. Throughout the second half of this year, our customers have substantially increased their demand for our differentiated solutions across critical technologies such as silicon photonics, silicon germanium, FDX and advanced packaging.

簡要談一下明年的資本支出展望，我們預期 2027 年將是 GF 加速投資的一年。今年下半年，我們的客戶已大幅提高對我們差異化解決方案的需求，涵蓋矽光子、矽鍺、FDX 與先進封裝等關鍵技術。

Importantly, we have the ability to grow our capacity quickly and efficiently within our existing four-walls manufacturing footprint, resulting in a strong return on investment for new CapEx spending, robust demand and customer commitments form the foundation of our capacity investment principles.

重要的是，我們有能力在既有的四面牆（four-walls）製造廠房範圍內，快速且有效率地擴充產能，使新增資本支出具備強勁的投資報酬率；強勁需求與客戶承諾構成我們產能投資原則的基礎。

As these signals strengthen, we expect approximately one-third of our net OpEx in 2027 to be supported by customer prepayments for incremental capacity agreements. In parallel, we also expect to secure CapEx offsets from investment tax credits and government grant frameworks, including a proportion of the recently announced $955 million repurposing from our previously awarded chips grants.

隨著這些訊號愈發明確，我們預期 2027 年約三分之一的淨營運費用（net OpEx）將由客戶針對增量產能協議的預付款所支持。同時，我們也預期可透過投資稅額抵減與政府補助架構取得資本支出抵減（CapEx offsets），其中包括近期宣布將先前核定的晶片補助款中 9.55 億美元的一部分改作他用（repurposing）。

Against this backdrop of customer demand, prepayments and government support we now expect net CapEx investments of approximately $4 billion into our manufacturing footprint in 2027, while continuing to deliver our free cash generation targets. These critical investments will unlock capacity constraints accelerate future growth and continue to evolve our business into a more diversified, resilient and accretive portfolio.

在客戶需求、預付款與政府支持的背景下，我們目前預期 2027 年將在我們的製造版圖投入約 40 億美元的淨資本支出，同時持續達成我們的自由現金產生目標。這些關鍵投資將解除產能瓶頸、加速未來成長，並持續推動我們的業務演進為更為多元、具韌性且具增益性的產品組合。

Our third quarter results demonstrate that GF strategy is translating into stronger financial performance. Importantly, we believe these improvements are structural. Our revenue mix continues to shift towards differentiated high-value technologies and services, while our government and customer partnerships are enabling us to make capital-efficient investments to scale and accelerate key technologies with favorable economics.

我們第三季的結果顯示，GF 的策略正轉化為更強勁的財務表現。重要的是，我們相信這些改善具有結構性。我們的營收組合持續轉向差異化的高價值技術與服務；同時，我們與政府及客戶的合作夥伴關係，使我們能以資本效率更高的方式進行投資，擴大規模並加速具備有利經濟性的關鍵技術。

All of these drivers reinforce our confidence in the long-term financial targets we presented at our Investor Day. GF today is more diversified and better positioned to capture more value than it ever has been. In closing, I want to thank our employees around the world for their continued focus and execution. Their work made this quarter's performance possible and gives us confidence in what GF can achieve next.

所有這些驅動因素都強化了我們對投資人日所提出長期財務目標的信心。如今的 GF 更加多元化，且比以往任何時候都更有利於捕捉更多價值。最後，我要感謝全球員工持續專注並落實執行。他們的努力讓本季的表現成為可能，也讓我們對 GF 接下來能達成的成果充滿信心。

With that, let's open the call to Q&A. Operator?

接下來，我們開放問答。接線員？

CJ Muse, Cantor Fitzgerald.

CJ Muse，Cantor Fitzgerald。

Yeah, good morning. Thank you for taking the question. I think one of the key themes from this call is you guys move into a structurally richer business mix, particularly as I think about silicon photonics pulled in by a year. And so I'm curious, you laid out targets of exiting 27% to 35% gross margins, 40% exiting two you're raising pricing starting Jan 1. How, I guess, is that portfolio of changes impacting your thinking about your gross margin trajectory into '27 and '28.

是的，早安。謝謝讓我提問。我認為本次電話會議的一個關鍵主題是，你們正轉向結構上更優質的業務組合，特別是我想到矽光子提前一年拉進來。因此我想請教，你們提出的目標是到 2027 年退出時毛利率 27% 到 35%，到 2028 年第四季退出時達到 40%，而且你們將自 1 月 1 日起調漲價格。我想問，這一系列變化的組合，如何影響你們對 2027 與 2028 年毛利率走勢的看法？

Yeah. Hey, CJ, I'll take that one, and then Tim probably overlay some of the strategic dynamics as well. We're really pleased with the margin trajectory. And I think you've seen from us consistently this year, near on 500 basis points of year-over-year margin growth. So if you go back to what we said we were going to do at the Investor Day back in May, we feel like we're doing exactly that mix has been a key component of it. We've also seen some gradual pickup in utilization as well through this year. But fundamentally, it's been driven by those strategic decisions that we've taken over the last couple of years.

是的。嗨，CJ，我先回答這題，之後 Tim 可能也會補充一些策略面的動態。我們對毛利率的走勢非常滿意。而且你今年也持續看到我們年對年毛利率提升接近 500 個基點。所以如果回頭看我們在 5 月投資人日所說要做的事，我們覺得我們正確實照著做，其中組合（mix）一直是關鍵因素之一。今年以來我們也看到產能利用率逐步回升。但從根本上來說，這是由我們過去幾年所做的那些策略性決策所驅動。

And technology services as a percentage of revenue clearly falls through as a richer mix and margin driver as well. That's continued to grow during the course of this year. So on balance, we feel like we're actually hitting that target, and we're well on track to the target that we set out in May. Remember, at the start of this year, we said we wanted to exit the fourth quarter of 2026 with a 30% gross margin handle. We were bumping up on that in our 2Q results, our 3Q results.

此外，技術服務占營收的比重顯然也帶來更優質的組合，並成為毛利率的驅動因素。這部分在今年過程中持續成長。整體而言，我們覺得我們實際上正在達成該目標，並且正穩步朝 5 月所設定的目標前進。請記得，在今年年初我們說過，希望在 2026 年第四季退出時毛利率能站上 30% 的水準。我們在第二季與第三季的結果就已經接近那個水準。

Obviously, we've gone through that target we're quite pleased with even where the guidance for fourth quarter, it shows about 1.5 points of growth relative to the original target we have. So very much feel at this point that we are on track to those gross margin goals that we set out earlier on this year.

很明顯地，我們已經超過該目標；我們對此非常滿意，即便看第四季的指引，相較於我們原先的目標也顯示約 1.5 個百分點的成長。因此，就目前而言，我們非常確信我們正按計畫達成今年稍早所設定的毛利率目標。

Obviously, there's still some way to bridge from, call it, the 31.4% in 3Q up to the 40% that we want to exit the fourth quarter of 2028 with -- but that mix dynamic and where we're continuing to invest in those corridors where we're seeing very strong demand signals from customers is continuing to push us in the right direction as it relates to that margin accretion. So no change to the goals that we laid out just a few months ago, but maybe I'll pass it to Tim for some of the other structural and strategic dynamics.

當然，從第三季的 31.4% 到我們希望在 2028 年第四季退出時達到 40% 之間，仍有一段差距需要彌補——但這種組合動態，以及我們持續在那些走廊（corridors）加碼投資、而客戶在其中釋出非常強勁需求訊號的領域，正持續把我們推向正確方向，並帶來毛利率的增益。因此，我們不會改變幾個月前所提出的目標；不過我也許把時間交給 Tim，請他談談其他結構性與策略性的動態。

Yeah. Thank you, Sam. No, CJ, I think the right way to think about this is we're delivering what we said earlier this year, right? We're winning in those secular growth markets, and we're winning with leading customers who, in many cases, are new to GF. And so the announcement we made just this quarter customers like MPS, customers like TSMC, expanded business with Marvell.

是的，謝謝你，Sam。不，CJ，我認為正確的思考方式是：我們正在交付今年稍早所說的內容，對吧？我們在那些長期結構性成長（secular growth）的市場中持續取勝，也與領先客戶合作取勝，而在許多情況下，這些客戶是 GF 的新客戶。因此，我們就在本季宣布了像 MPS 這樣的客戶、像 TSMC 這樣的客戶，以及與 Marvell 擴大的業務合作。

By the way, there are many others that supported our several announcements over the quarter, just tell you kind of the breadth of engagement we have. and that's giving us a lot of confidence to bring those technologies to market, accelerate those brands to accelerate the capacity investments and obviously unlock even more growth through the partnership we struck with the US and Japan. So look, it's been a great period of accelerating our business model and we're very confident in the trajectory ahead.

順帶一提，還有許多其他客戶支持了我們本季的多項公告，這也反映出我們合作互動的廣度。而這讓我們對將這些技術推向市場、加速這些品牌、加速產能投資充滿信心，並且透過我們與美國及日本達成的夥伴關係，顯然能解鎖更多成長。所以你看，這段期間我們的商業模式加速推進得很不錯，我們對未來的走勢非常有信心。

Thanks, Tim. I guess as a follow-up question, could you speak to, I guess, your early look to Q1 seasonality. If I look at it, the implied SMD, you're talking down 22% year-on-year for Q4 and down 40% from your prior peak. And at the same time, CID is tracking 80% implied in Q4. So would love to hear how you're thinking about the March quarter.

謝謝，Tim。我想追問一下，你能否談談你們對第一季季節性的初步看法。如果我看這裡，隱含的 SMD，你們提到第四季年對年下滑 22%，且較先前高點下滑 40%。同時，CID 在第四季的隱含值追蹤到 80%。所以很希望聽聽你們如何看待 3 月季度。

Yeah. Happy to give you a little bit of color, CJ, Obviously, just mindful that we're going to guide 1 quarter at a time. But the way you've characterized it is the right way. And we think that relative to where we said earlier on this year, we're broadly on track with several of those end markets, right? We expected that mobile would be down in the low teens percentage, no change for that. For the full year, 26%. We expected that IoT would be in that 10% to 15% range full year.

好的，CJ，我很樂意提供一些補充說明。當然也要留意，我們會一次只提供一個季度的指引。但你這樣的描述方向是對的。我們認為，相較於今年稍早我們所說的情況，我們在其中幾個終端市場上大致符合進度，對吧？我們原先預期行動裝置（mobile）會下滑到低十幾個百分點的幅度，這點沒有改變。就全年而言，是 26%。我們原先預期 IoT 全年會落在 10% 到 15% 的區間。

Again, no change in automotive, low double-digit percentage growth. That's been a consistent theme for us. So really, the step change has been the accelerated growth that we're seeing in comms infrastructure and data center and the way you thought about it for the fourth quarter is the right way.

同樣地，汽車業務沒有變化，維持低雙位數百分比成長。這一直是我們一貫的主題。因此，真正的躍升在於我們在通訊基礎設施與資料中心所看到的加速成長，而你對第四季的看法是正確的。

Looking out to 2027, obviously, we'll put more color on that when we get to the next call. But just going back to what we said as part of our Investor Day in May, look, I think the general trends are very constructive as it relates to automotive, IoT, remember in that space, we said low double-digit growth is our long-term goal for automotive and sort of mid-teens for IoT. Comms infra and data center is where we think that there's very strong momentum to continue growing.

展望到 2027 年，顯然我們會在下一次電話會議時提供更多細節。但回到我們在 5 月投資人日所說的，我認為就汽車、IoT 而言，整體趨勢非常正向；記得在那個領域，我們說汽車長期目標是低雙位數成長，而 IoT 大約是中十幾%的成長。通訊基礎設施與資料中心則是我們認為仍有非常強勁動能可持續成長的領域。

The target that we put out previously in our long-term model was 30-plus percentage points you can defer from where we reported in Q3 and the direction of travel for Q4 that we're seeing that continued momentum in that space. So I'd say that, that is probably a tailwind going into 2027. And then look, mobile, we've been consistent on the commentary there and largely tracking where we see the rest of the market. So I don't think any surprises on that front. But hopefully, that gives you a little bit of color as we're thinking about it for the full year, and obviously, happy to give you more color on Q1 when we get to the next quarter.

我們先前在長期模型中提出的目標是 30 個百分點以上；你可以從我們在第三季所報告的結果，以及我們看到第四季的走勢方向推導出，在該領域的動能仍在延續。所以我會說，這很可能是邁向 2027 年的一股順風。另外，行動裝置方面，我們在相關評論上一向一致，且大致上與我們對整體市場的觀察同步。因此我不認為那方面會有任何意外。但希望這能讓你對我們如何看待全年有一些補充說明；當然，等到下一季時，我們也很樂意就第一季提供更多細節。

Very helpful. Thank you.

非常有幫助。謝謝。

Sahej Singh, Stifel.

Sahej Singh，Stifel。

Hey, thank you for taking my question. And congrats, by the way, on the partnerships you've lined up this quarter. I think they say a lot about where Global sits in the AI supply chain. Tim, I guess, thanks for the color that customer prepayments will support roughly a third of 2027 CapEx, I think I heard. And since May, you've sort of talked about customers discussing advanced payments to secure FDX, IFO and silicon germanium capacity.

嗨，謝謝讓我提問。順帶一提，也恭喜你們本季敲定的合作夥伴關係。我認為這充分說明了 Global 在 AI 供應鏈中的位置。Tim，我想謝謝你補充說明客戶預付款將可支應約三分之一的 2027 年資本支出（CapEx），我想我聽到的是這個。而且自 5 月以來，你們一直提到客戶在討論以預先付款的方式來確保 FDX、IFO 以及矽鍺（silicon germanium）產能。

And you've had CPO tape-out in the first quarter and a scale tape-out in the second -- so I guess, even directionally, could you give some color into how much of your 2027, 2028 site-on-site capacity is now under signed commitments? And are those agreements even reaching into 2029 as NPL and CPO potentially come in?

而你們在第一季完成了 CPO tape-out、第二季完成了規模化（scale）tape-out——所以我想問，即使只是方向性地，你能否補充一下：你們 2027、2028 年各廠區（site-on-site）的產能，目前有多少已在已簽署的承諾之下？另外，隨著 NPL 與 CPO 可能導入，這些協議是否甚至延伸到 2029 年？

Yeah. No, thank you, Sahej. So I think the picture is as follows. The acceleration of optical communication is very clear. And I think we're still at the very early ins of that, as you see just optical links replacing copper in all aspects of the data center, just demand is truly skyrocketing.

是的。不，謝謝你，Sahej。我認為情況如下。光通訊加速的趨勢非常明確。而且我認為我們仍處於非常早期的階段；你可以看到在資料中心的各個面向，光連結正取代銅線，需求確實正在急遽攀升。

And so we're having those discussions with our customers and lining up the supply they need. We are absolutely booked out through 27, increasingly booked out through '28 and the question for us is how fast we can add more capacity to meet their requirements in '28. And many of them are seeking capacity assurance through '29 and even in some cases, beyond. And I think in every single case, they're willing to put their balance sheet to work to support us in those investments as well.

因此我們正與客戶進行這些討論，並為他們所需的供應做規劃。我們的產能在 27 年之前已完全被預訂，28 年也愈來愈多被預訂；對我們而言的問題是，我們能多快增加更多產能，以滿足他們在 28 年的需求。而且許多客戶正在尋求到 29 年、甚至在某些情況下更久的產能保障。我認為在每一個案例中，他們也都願意動用其資產負債表來支持我們進行這些投資。

So it's a very constructive environment in terms of visibility, in terms of demand, and it's really now down to us in terms of execution, right, installing tools, driving productivity. And that's really what you're seeing accelerate our own trajectory even within '26 we're doing better than we expected based on our acceleration of investments and the ability to ramp those tools to production quickly. And that's true across all those corridors you mentioned. And I'd say increasing that will also be true in advanced packaging with some of the interposer work we've also announced.

所以就可見度與需求而言，這是一個非常具建設性的環境；現在真正取決於我們的就是執行力，對吧——安裝設備、提升生產力。而這也正是你所看到的：我們自身的軌跡正在加速，甚至在 26 年內，我們的表現也優於原先預期，因為我們加速投資，並且能夠快速把這些設備爬坡到量產。而你提到的那些走廊（corridors）都同樣如此。我也會說，隨著我們也宣布了一些中介層（interposer）相關工作，先進封裝方面也會愈來愈是如此。

Yeah. Thank you for that. I guess interposer work is topical as well. And Sam, thank you for the color on the on the blended ASPs. I guess I'd like to build on CJ's question on the margin drivers. Particularly would help us price -- frame pricing specifically. At the Investor Day, you were pretty clear that the bridge to the 40% exiting '28 doesn't embed significant price increases from supply and demand. So now with ASPs up year-over-year for the first time since 2023, should we think of pricing as additive to that to the 40%, I mean?

好的。謝謝你的說明。我想中介層（interposer）的工作也很切題。Sam，也謝謝你對混合 ASP 的補充。我想在 CJ 關於毛利率驅動因素的問題上再延伸一下。特別是會有助於我們定價——具體來說是框架（frame）定價。在投資人日，你們相當清楚地表示，通往 28 年底達到 40% 的路徑，並未納入因供需而帶來的顯著漲價。所以現在 ASP 年增、且是自 2023 年以來首次年增，我們是否應該把定價視為對 40% 目標的額外加分呢？

And then on the TSMC interposer agreement. Thanks for the color that margin set above your long-term targets? And at roughly $400 million a year or about 4% of revenue against 28 consensus. Is it fair to think of that as adding on the magnitude of 50 basis points to corporate gross margins at run rate alongside the 100 to 200 basis points from the US Japan program?

另外關於與台積電（TSMC）的中介層協議。謝謝你補充說明其毛利率水準高於你們的長期目標。以每年約 4 億美元、或相對於 28 年市場共識營收約 4% 來看。是否可以合理地認為，在穩態（run rate）下，這將為公司整體毛利率帶來約 50 個基點的提升，並且與美國、日本計畫帶來的 100 到 200 個基點一起發揮作用？

So just a little bit to unpack that. So maybe I'll start with the pricing point and then we can talk more broadly around margins -- and look, it's encouraging what we've seen in the third quarter, but also not surprising. We gave the color on this in prior quarterly earnings updates where we said we expected the second half of 2026 to be the point where we lapped some of those pricing headwinds, if you like, that we've seen over the course of the last few years, and that was certainly the case in the third quarter, quarter-over-quarter blended ASPs were up about 2.5%. And for the year-over-year period, they were up almost 5%.

這裡有幾點需要拆解一下。我先從定價談起，然後我們再更廣泛地談毛利率——你看，我們在第三季看到的情況令人鼓舞，但也不意外。我們在先前的季度財報更新中已提供過相關說明：我們預期 2026 年下半年會是我們跨越（lapped）過去幾年所面臨的一些定價逆風（如果你願意這麼說）的時間點，而第三季確實如此；季對季來看，混合 ASP 上升約 2.5%。而年對年來看，幾乎上升了 5%。

So we feel that across all of the end markets that we operate in now that pricing is a much more constructive demat. But obviously, mix is a key component of that as well. We are servicing a richer mix of demand within the end market portfolio that we have. And so looking out to 2027, like I said, this combination of richer mix and blended ASP mix improvements, points us towards probably about a 10% uplift in blended ASP mix 2027, which I think is consistent more broadly with the market. And it's consistent with what we said we were going to do as it relates to pricing actions as well that we took towards the end of the first half of 2026.

因此我們認為，在我們所服務的所有終端市場中，定價環境現在更具建設性。但顯然，產品組合（mix）也是其中的關鍵因素。我們在終端市場組合中，正在服務更高價值的需求組合。因此展望 2027 年，如我所說，較高價值的組合與混合 ASP 組合改善的結合，指向 2027 年混合 ASP 組合大約有 10% 的提升；我認為這與更廣泛的市場趨勢一致。也與我們所說、在 2026 年上半年末所採取的定價行動相一致。

So naturally, that serves as a tailwind. But I would say, Sahej, in the same way that we -- when we see pricing going the other way in prior year, pricing is only one component from the margin perspective. And it's actually a far less impactful component than even the broader mix dynamics are concerned, the utilization, the depreciation, the productivity within our manufacturing footprint. So like I said to CJ, we're not looking to change the targets that we put out there from the perspective of our gross margin accretion over the course of the next couple of years, but you can infer that the decision, these announcements that we've made over the course of the last couple of months, really give us more clarity, more certainty on our ability to hit those targets.

所以自然地，這會形成一股順風。但我會說，Sahej，就像我們在前一年看到定價走向相反時一樣，從毛利率角度來看，定價只是其中一個構成要素。而且相較之下，它的影響其實遠不如更廣泛的組合動態、產能利用率、折舊，以及我們製造版圖內的生產力來得重要。所以如同我對 CJ 所說的，我們不打算改變先前所提出、未來幾年毛利率提升（accretion）的目標；但你可以推論，我們在過去幾個月所做的這些決策與公告，確實讓我們在達成這些目標的能力上有了更高的清晰度與確定性。

So I'd say no major change on that front, but we're encouraged by the momentum. And naturally, as you would expect, the announcement for the sort of in process with the SMC, when that starts to ramp in the first half of 2028, we would expect that to be margin accretive as well to our overall objective. So the change to the targets, greater visibility, and we think a high degree of momentum to get to those targets.

所以我會說在那方面沒有重大變化，但我們對這股動能感到鼓舞。而且如你所預期的，關於與 SMC 進行中的那類公告，當它在 2028 年上半年開始放量時，我們也預期這將對我們整體目標帶來毛利率的增益。因此，目標的調整帶來了更高的能見度，我們也認為達成這些目標的動能相當強。

Thank you. Very helpful.

謝謝。非常有幫助。

Jim Schneider, Goldman Sachs.

高盛，Jim Schneider。

Good morning. Thanks for taking my question. Maybe just sort of building on the theme that's been asked for the last few questions. wanted to maybe clarify more tactically where you see the various moving parts of utilization pricing and mix going into the first half of '27. It seems like, if anything, mix is improving pricing, as you said, is getting better. And then I guess the question is about utilization.

早安。謝謝讓我提問。也許延續前幾個問題的主題，我想更戰術性地釐清，你們如何看待到 27 年上半年為止，產能利用率、定價與產品組合等各項變動因素。看起來如果有什麼變化的話，正如你所說，產品組合正在改善、定價也在變好。那我想問的是產能利用率。

And so do you think you're going to be able to maintain sort of at least flattish utilization? Because it seems like given everything that's going on with the smartphone market and it's relatively dilutive margins that mix or gross margins could actually at least stay flat in the first half, if not go up?

所以你們認為能否至少維持大致持平的產能利用率？因為看起來，考量智慧型手機市場正在發生的一切，以及其相對稀釋的毛利率，產品組合或毛利率在上半年其實至少可以維持持平，甚至可能上升？

Yeah. Happy to take that one, Jim. And look, I think I like what I said around some of the structural drivers that we see across those end markets. And naturally, from a mobile point of view, we think we largely track that market. But the offset that we're seeing in terms of mix benefits from the other end markets or a tailwind from our point of view going into next year. I will say that naturally, there's a degree of seasonality from the fourth quarter into the first quarter, you can see that in prior years.

好的，Jim，我很樂意回答。你看，我想我先前提到我們在這些終端市場看到的一些結構性驅動因素。而且自然地，從行動裝置的角度來看，我們認為我們大致會跟著該市場走。但我們從其他終端市場看到的產品組合效益所帶來的抵銷，或從我們角度看是進入明年的一股順風。我也要說，自然地，從第四季到第一季存在一定程度的季節性，你可以在過去幾年看到。

But given some of the structural shifts that we're seeing within those end markets, we do think that there should be a slightly less pronounced impact from a seasonality point of view fourth quarter into the first quarter. And obviously, I'll give more color around that when we get to the next call.

但鑑於我們在這些終端市場內部看到的一些結構性轉變，我們確實認為，從第四季到第一季的季節性影響應該會稍微不那麼明顯。當然，等到下一次電話會議時，我會提供更多細節。

But hopefully, that helps give you a little bit of insight from a tactical point of view. And then as it relates to more broadly and into 2028, like Tim said, and like I said, we think we're increasing our visibility on getting to those objectives that we set out from a margin point of view and really have our eyes set on getting to that 40% gross margin, we think we're making good progress towards that.

但希望這能從戰術角度給你一些洞見。另外，談到更廣泛的面向並延伸到 2028 年，如 Tim 所說、也如我所說，我們認為我們在達成先前設定的毛利率目標方面能見度正在提高，而且我們確實把目標鎖定在達到 40% 的毛利率；我們認為我們正朝此取得良好進展。

Maybe, Jim, one comment I'll add is about just thinking about utilization. It's also about flexibility. We spent a lot of time over the last three years really investing in our manufacturing footprint to make it as flexible as possible. so that we can respond when one market is weak and one market is stronger, we can respond and use that capacity across where those grow. Obviously, there are always limitations to that, but that's definitely a factor that's helping us overall improve utilization this year and picking up into next year as well.

Jim，我再補充一點關於產能利用率的看法。這也關乎彈性。過去三年我們花了很多時間投資我們的製造版圖，讓它盡可能具備彈性。如此一來，當某個市場疲弱、另一個市場較強時，我們就能因應，並把產能用在成長的地方。當然，這總是有其限制，但這確實是幫助我們今年整體提升產能利用率、並在明年進一步改善的一項因素。

That's fine. Thank you. And then maybe just to clarify, on the smart mobile device market, clearly, there's some correction there for all the reasons that we understand. But I just wanted to clarify, your outlook is unchanged on the full year. So can you maybe characterize where you think the downstream inventory levels stand with respect to component inventories and smart mobile devices and just kind of making sure that you're sort of aligned with inventory levels you can go in terms of unit demand tracking from there?

好的。謝謝。另外也許再釐清一下，在智慧行動裝置市場方面，顯然基於我們理解的各種原因，確實有一些修正。但我想確認的是，你們對全年展望沒有改變。所以你能否描述一下，你認為下游庫存水位在零組件庫存與智慧行動裝置方面目前處於什麼位置，並確認你們在單位需求追蹤上與庫存水位是對齊的？

Yeah. Thanks, Jim. I think the broader topic on SMB is just how the market is quite bifurcated. I think the high end is not that bad, right, all things considered. Obviously, can weather the memory impacts much more. The low end is much, much worse. Our exposure is much more to the high end of the market. And I'd even say in the current environment, some of the capacity in the industry more broadly is being taken up by other end markets, which actually, in turn, leads mobile customers to come back to us and look for more supply to make sure they have enough to do what they need to build.

好的，謝謝，Jim。我認為關於 SMB 的更大議題在於市場其實相當兩極化。我認為高階市場其實沒有那麼糟，綜合各項因素來看是如此。顯然，它更能承受記憶體的影響。低階市場則糟得多、糟很多。我們的曝險更多是在高階市場。我甚至會說，在目前環境下，整個產業的一部分產能更廣泛地被其他終端市場吸收，這反過來其實會讓行動客戶回到我們這裡尋求更多供應，以確保他們有足夠的量來完成他們需要的組裝。

So I'd say, overall, we're not seeing any kind of concerns from an inventory point of view. Like I said, overall, we're going to track the market going into 2027. But the situation is, that's not as bad as people might think.

所以我會說，整體而言，從庫存角度我們沒有看到任何疑慮。如我所說，整體而言，我們在進入 2027 年時會跟著市場走。但情況並沒有大家可能想的那麼糟。

Thank you.

謝謝。

Mehdi Hosseini, Sequanna International.

Sequanna International，Mehdi Hosseini。

Yes, sir. Thanks for taking my question. I want to dive into communication infrastructure. And Tim, can you please update us on how you see the mix between Sifang changing into 2027 and 2028. And I asked the question in this context because last time you highlighted that in '26, SiGe mix is higher than cycle. And again, I wanted to see how that mix is changing over the next couple of years. And I do have a follow-up.

是的，先生。謝謝讓我提問。我想深入談談通訊基礎設施。Tim，你能否更新一下你如何看待 Sifang 之間的產品組合在 2027 與 2028 年的變化？我在這個脈絡下提問，是因為上次你提到在 26 年，SiGe 的組合占比高於景氣循環水準。同樣地，我想了解未來幾年這個組合會如何變化。我還有一個追問。

No, thanks, Mehdi. I think the story I'd say is very simple. The optical transition is pulling both of them very strongly, and I'd say SiGe is getting in a particular pickup because of the transition to 200 gig per lane speeds, and I think that's going to continue going forward. We've sold out effectively of both. We're adding capacity for both.

好的，謝謝，Mehdi。我想我會說這個故事很簡單。光學轉型正非常強勁地拉動兩者，而我會說 SiGe 因為轉向每通道 200G 的速度而出現特別明顯的拉升，我認為這個趨勢未來會持續。我們基本上兩者都已經賣到滿載。我們也正在為兩者增加產能。

We have an excellent road map for both, right? And I think our strategy will continue to operate on that basis. I think both are going to see very strong growth going into '27 and beyond. And like I say, customers are now willing to put their balance sheet to work to help us invest and add capacity, again, both those end markets, so very strong. And you didn't ask it, but I'll answer anyway, the transition to technologies like NPO are going to accelerate that story even further because that requires even higher-performance photonics, but also extra SiGe content as well. So those are tailwinds for the business going forward.

我們對兩者都有非常出色的產品路線圖，對吧？我認為我們的策略會持續以此為基礎運作。我認為兩者在 27 年及之後都會看到非常強勁的成長。而且如我所說，客戶現在也願意動用他們的資產負債表來協助我們投資並擴增產能——同樣地，這兩個終端市場都非常強勁。你雖然沒有問，但我也一併回答：像 NPO 這類技術的轉換會讓這個故事進一步加速，因為那需要更高效能的光子技術，同時也需要更多的 SiGe 含量。所以這些都是未來推動業務的順風因素。

I was actually going to follow up on NPL, which is kind of a black box, in my opinion, is more about integration, heterogeneous integration, you have PIC and EIC, which is quite different than dispute components associated with the pluggable transceiver. And my follow-up question for you is how do you see GLOBALFOUNDRIES content changing from pluggable to NPO transceivers?

我其實正要追問 NPL；在我看來它有點像黑盒子，更偏向整合、異質整合，你有 PIC 和 EIC，這與可插拔收發器相關的分離式元件相當不同。我對你的追問是：你如何看待 GLOBALFOUNDRIES 的含量，從可插拔收發器轉向 NPO 收發器時會如何變化？

Yeah. No, it's a great question. I'd say, overall, a transition from pluggable to NPL and to CPO, by the way, which we still believe will continue to happen. All of that creates a growing content opportunity for us. Because obviously, as you go forward in high performance, the PIC has to do more right in different applications. You have more SiGe content in each of those applications. And then as you get closer to the XPU, you have some of the packaging components be that fiber attach guide away for bonding and so on also play an important role.

是的。不，這是個很好的問題。我會說，整體而言，從可插拔（pluggable）轉向 NPL，再到 CPO（順帶一提，我們仍然相信這個趨勢會持續發生）。這一切都為我們帶來了日益增加的內容（content）機會。因為很明顯，隨著你在高效能領域持續往前推進，PIC 必須在不同應用中把更多事情做對。而在每一個這些應用中，你會有更多的 SiGe 內容。接著當你更接近 XPU 時，一些封裝元件——無論是用於鍵合的光纖貼附導引（fiber attach guide）或其他——也都扮演重要角色。

So I'd say the opportunity increases as you get closer to the ultimate chip. And therefore, our investments in R&D, building the scale platform and others have really begun to pay off very well in terms of customer interest and traction for long-term road map in those areas.

所以我會說，當你越接近最終的晶片時，機會就越大。因此，我們在研發上的投資、建立規模化平台以及其他投入，已經開始在客戶興趣與在這些領域的長期路線圖推進（traction）方面，帶來非常好的回報。

Great. Thank you.

很好。謝謝你。

Harlan Sur, JPMorgan.

Harlan Sur，摩根大通（JPMorgan）。

Yeah. Good morning. Thank you for taking my questions, and great job on the quarterly execution. June quarter utilization were in the high 80% range. Did that increase in the September quarter, maybe towards the 90% range. That would imply -- the team still has another 8 to 10 percentage points of utilization increases ahead if that's the case, I mean, by my calculations, you still have roughly about 400 basis points of gross margin improvement in front of you, just from utilization right how much of that potentially gets unlocked in 2027 as manufacturing activity and utilizations continue to increase?

是的。早安。謝謝讓我提問，也恭喜本季執行表現很出色。6 月季度的產能利用率在 80% 高段。9 月季度是否有提高，可能接近 90%？如果是這樣，這表示——以我的計算——團隊在利用率上仍有另外 8 到 10 個百分點的提升空間；僅就利用率而言，你們前面大約還有 400 個基點的毛利率改善空間。隨著製造活動與利用率持續提升，其中有多少可能會在 2027 年被釋放出來？

Yeah. Hey, Harlan. Good morning. Maybe just to pick up on that, you're right, utilization in the high 80s in the second quarter. Utilization in the third quarter was also in the high 8s. One of the reasons for that is that we've been investing CapEx through this year. We've been adding to the capacity, which in effect adds to that denominator for the purpose of calculating utilization. So you're right, we do still see utilization upside as an opportunity from a margin point of view.

是的。嗨，Harlan。早安。先接續你這點，你說得對，第二季的利用率在 80% 高段。第三季的利用率也同樣在 80% 高段。其中一個原因是我們今年一直在投入資本支出（CapEx）。我們在增加產能，而這實際上會增加用來計算利用率的分母。所以你說得對，從毛利率角度來看，我們確實仍把利用率上行視為一個機會。

But as you heard on the call, we're going to be continuing to invest in that capacity and really sort of grow the overall corridors where we see strong demand to be able to expand margins from the mix point of view, but also from a utilization point of view into 2027. So at this point, headroom to certainly grow into from a utilization point of view, and we think we're going to keep growing that denominator as we invest CapEx dollars to grow our capacity.

但如同你在電話會議上聽到的，我們會持續投資這些產能，並真正去擴大我們看到需求強勁的整體走廊（corridors），以便從產品組合（mix）角度、也從利用率角度，在 2027 年之前擴大毛利率。所以就利用率而言，確實還有可成長的空間，而我們也認為，隨著我們投入資本支出資金來擴充產能，我們會持續把那個分母做大。

No, I appreciate that. And then for Tim, on the Photonics front, lasers are also in pipe supply, right, in high-quality lasers, very, very tight supply situation. I mean Marvell is paying a healthy premium to acquire Lumentum might partly for the indium phosphide laser portfolio. The GF team has done a very good job in compound semiconductors, right, silicon germanium, you talk about quite a bit. targeted for high-quality TIAs and drivers is a good example. Indium phosphide lasers would also add to your slide capabilities and portfolio as the team -- is the team considering adding indium phosphide manufacturing to your technology offerings?

了解，謝謝說明。接著給 Tim 一題：在光子（Photonics）方面，雷射在供應鏈中也屬於瓶頸供應，對吧？高品質雷射的供應非常、非常吃緊。我的意思是，Marvell 以不小的溢價收購 Lumentum，部分原因可能就是為了磷化銦（indium phosphide）雷射產品組合。GF 團隊在化合物半導體方面做得非常好，對吧？像你們談到很多的矽鍺（silicon germanium）。以用於高品質 TIA 與驅動器（drivers）的目標應用來說就是個好例子。若能加入磷化銦雷射，也會擴充你們的晶片能力與產品組合——團隊是否正在考慮把磷化銦製造納入你們的技術供應項目？

No, it's a great question, Harlan. And you're definitely seeing a strong interest in that part of the value chain expressed by the transaction you mentioned, but also other investments that we're hearing about in the industry. We're focusing on integration of indium phosphide into our technologies, right?

不，這是個很好的問題，Harlan。你確實看到價值鏈那一段的強烈興趣，除了你提到的那筆交易之外，我們也聽到產業內其他投資動向。我們的重點是把磷化銦整合進我們的技術中，對吧？

And I think part of the road map is about having integrated lasers into greater SOAs for future optical applications. By the way, just like we're also investing in novel materials for future enhanced modulation to go to 400 gig per lane and beyond. And part of our roadmap, what we can invest at the silicon level. Discrete laser manufacturing, not on our current roadmap, but obviously we'll continue to be open to opportunities going forward.

我認為路線圖的一部分，是在未來的光學應用中，把整合式雷射納入更大型的 SOA（半導體光放大器）。順帶一提，就像我們也在投資新型材料，用於未來的增強調變，以邁向每通道 400G（400 gig per lane）及更高。而我們路線圖的一部分，是我們能在矽層級投入的項目。離散式雷射製造目前不在我們的路線圖上，但顯然我們會持續對未來的機會保持開放。

Thank you.

謝謝你。

Joseph Moore, Morgan Stanley.

Joseph Moore，摩根士丹利（Morgan Stanley）。

Great. Thank you. In the last cycle, it seemed like a lot of the focus on wafer securing was coming from automotive. And you had big automotive OEM deals that we're working with GLOBALFOUNDRIES. And obviously, we had shortage of cars that was driven by all of this. And as I look at things now, I mean, your automotive business is doing fine, but I worry about looking at all the capacity going to interposers and silicon photonics and AI. These fabs are getting full.

很好。謝謝。在上一個景氣循環中，似乎很多對晶圓確保（wafer securing）的關注是來自汽車領域。而你們也有一些大型汽車 OEM 的合作案正在與 GLOBALFOUNDRIES 推進。很明顯地，我們當時也出現了由此導致的缺車。而我現在看，雖然你們的汽車業務表現不錯，但我擔心看到大量產能被分配到中介層（interposers）、矽光子與 AI。這些晶圓廠正在被填滿。

Like do you think we could be looking at another capacity crunch in the automotive space? And I guess, how are you thinking about that? And how are your customers thinking about that? Are you seeing people start to focus on those supply chains in the way that they did a couple of years ago?

你認為我們是否可能在汽車領域再度面臨一次產能吃緊？我想問，你們怎麼看這件事？你們的客戶又怎麼看？你是否看到大家開始像幾年前那樣，再度把焦點放在這些供應鏈上？

Yeah. Thank you, Joe. It doesn't get as much of a limelight, but I think the sector continues to be advancing forward in terms of automotive that is in terms of its thinking on supply security. We have a lot of conversations with automotive players, not just for like the next 12 months but the next kind of 3 to 5 years.

是的，謝謝你，Joe。這件事雖然沒有那麼受矚目，但我認為汽車產業在供應安全（supply security）的思維上仍持續向前推進。我們和許多汽車業者有很多對話，不只是針對接下來 12 個月，而是接下來大約 3 到 5 年。

And for a lot of them, they're very keen also to see that geographic supply diversification as well. And that plays particularly well for us given our US footprint and our European footprint. So I would expect continuous investment, more engagement, more direct business opportunities in auto as well. Many of them are doing that, both OEMs and Tier 1s.

而對其中許多人來說，他們也非常希望看到供應在地理上的多元化。考量到我們在美國與歐洲的布局，這對我們特別有利。所以我預期汽車領域也會持續投資、更多互動、以及更多直接的商機。其中許多公司都在這麼做，包含 OEM 與一階供應商（Tier 1）。

And that also plays, by the way, into our custom business where a part of what we've announced in our prepared remarks was progress in things like custom radar, which is another area we're developing further within the platform. So I'd say automotive continues to be in line with what we talked about at our Investor Day, strong consistent growth opportunity in where GS is really very well positioned to play across multiple technologies and with multiple end customers.

而且順帶一提，這也與我們的客製化業務相呼應；我們在事先準備的發言中提到的進展之一，就是像客製化雷達這類項目，這也是我們在平台內進一步發展的另一個領域。所以我會說，汽車領域的發展仍符合我們在投資人日所談到的方向：這是一個強勁且穩健的成長機會，而 GS 在多項技術、以及面向多個終端客戶方面，都處於非常有利的位置。

And Joe, maybe just more broadly, as you think about some of the CapEx that we've guided to today for 2027, as you'd expect, a lot of that is going into both the silicon photonics, SiGe data center applications. But we're not leaving our other end markets behind and automotive is a key component of that. So as we think about driving towards that low double-digit percentage growth, investing into some of our automotive capacity corridors across the footprint will also feature within that overall $4 billion net CapEx that we indicated for next year.

另外，Joe，更廣泛地說，當你思考我們今天對 2027 年所給出的資本支出指引時，如你所預期，其中很大一部分會投入在矽光子、SiGe 資料中心應用。但我們並沒有忽略其他終端市場，而汽車就是其中的關鍵組成。因此，當我們思考如何推動達到低雙位數百分比的成長時，對我們整體版圖中部分汽車產能走廊的投資，也會納入我們所提到明年 40 億美元淨資本支出（net CapEx）的整體規劃之中。

Yeah, that's helpful. Thank you. And in terms of the OEMs, you mentioned, that those relationships are direct. I mean, I feel like that's what's been unique about the last few years is that you're getting automotive OEMs showing awareness of foundry supply chains and things. Is it still -- I guess, are they emphasizing it enough, do you think? And is it still very much a focus on the geographic sourcing? Is that coming from the Tier 1s? Or is that coming from the OEMs?

是的，這很有幫助。謝謝。另外就你提到的 OEM 而言，你說那些關係是直接的。我的感覺是，過去幾年比較特別的是，你看到汽車 OEM 開始意識到晶圓代工供應鏈等議題。我想問的是——你認為他們現在是否足夠重視？而且是否仍然非常聚焦在地理來源（geographic sourcing）？這個需求是來自 Tier 1？還是來自 OEM？

I think it's both. And I think part of it also is increased differentiation. I mean if you look at today's car, the content growth is just tremendously increased over the last years. And all of them have road maps to do more, right, as you move to different architectures, some of that's linked to electrification, but some of it is just linked to performance and feature sets of the cars. I think it's also being led by the more kind of silicon native OEMs.

我認為兩者都有。而且我認為其中一部分也來自差異化程度的提升。我的意思是，如果你看看今天的汽車，內容（功能/電子含量）的成長在過去幾年大幅增加。而且當你轉向不同的架構時，它們都有更多的路線圖要做，對吧；其中一部分與電動化相關，但另一部分則純粹與汽車的性能與功能組合有關。我認為這也正由更偏向「矽原生」的OEM所帶動。

By the way, I would say, obviously, Tesla, but I would also cite the Chinese players that are increasingly setting the term globally for what the automotive industry used to do in terms of performance. So I just think you've got a much more aware end market, and that creates a lot of opportunities for players like us.

順帶一提，我會說，顯然是特斯拉，但我也會提到中國的業者，他們正日益在全球範圍內為汽車產業過去在性能方面所做的事情設定標準。所以我只是認為，你面對的是一個更有認知的終端市場，而這為像我們這樣的業者創造了很多機會。

Great. Thank you.

很好。謝謝。

Krish Sankar, TD Cowen.

Krish Sankar，TD Cowen。

Yeah. Hi. Thanks for taking my question. I had two of the first one, Tim or Sam, can you give us an update on the silicon photonics competitive landscape? Obviously, there is TSMC, Tower, STMicro or UMC. And also, along the same path, I think the Marvell-Lumentum merger highlights one of the best aspects which people worry that you could be into optical transceiver overbuilt in 2028. I'm just kind of curious if you have any thoughts around that, and I have a quick follow-up.

是的，嗨。謝謝讓我提問。我有兩個問題，第一個是，Tim或Sam，你們能否更新一下矽光子（silicon photonics）的競爭版圖？顯然有台積電、Tower、意法半導體或聯電。另外沿著同樣的脈絡，我認為Marvell與Lumentum的合併凸顯了其中一個大家擔心的面向：你們可能會在2028年面臨光收發器（optical transceiver）產能過剩。我想請教你們對此是否有任何看法，之後我還有一個簡短追問。

Yeah. No, great question, Chris. Look, overall, it's great to see other players believing the same thing, we believe that the optical transition is just getting started, and that's why you're obviously seeing people enter the space who previously weren't very engaged in photonics. We've been doing this for 10 years. We've got a very strong technology road map and positioning.

是的。Chris，這是個很好的問題。整體來看，很高興看到其他玩家也相信同樣的事情；我們認為光學轉型才剛開始，這也是為什麼你會看到一些過去並未深度投入光子領域的人開始進入這個市場。我們已經做了10年。我們有非常強的技術路線圖與市場定位。

Today, we've also got a lot of investments beyond the PIK level into the packaging side, which I think will play more and more into those new applications we talked about earlier, like NPO and CPO. But I think perhaps even more important right now is we've got the scalable fab footprint to invest and grow with our customers without adding fab fates for it, particularly for photonics and SiGe, where they need that capacity rapidly. And that's one thing they're betting on in terms of coming to GF for more of their capacity.

今天，我們在封裝端也有大量投資，超越PIK層級；我認為這將越來越多地支撐我們先前談到的新應用，例如NPO與CPO。除此之外，我認為目前可能更重要的是，我們擁有可擴展的晶圓廠佈局，能在不增加額外晶圓廠負擔的情況下，隨著客戶一起投資與成長，特別是在光子與SiGe方面，客戶需要快速取得那樣的產能。而這也是他們選擇把更多產能需求交給GF的一個押注點。

So I think we feel very confident in our competitive position within that landscape. Look, I view the recent announcement, the Marvell momentum announcement, again, it's a pretty strong endorsement of the optical transition, both betting a bit, if you like, their future on optical, which I think is reassuring. Both, by the way, our close partners with GF would be public about that as well. So we think this is a great opportunity for all of us to do more innovate faster and to grow the business. So I think it's a tailwind.

所以我認為，我們對於在這個競爭版圖中的競爭地位非常有信心。你看，我把最近的公告——Marvell與Lumentum的公告——視為對光學轉型相當強的背書；某種程度上等於把他們的未來押在光學上，我認為這令人安心。順帶一提，他們兩家也都是GF的緊密合作夥伴，這點我們也可以公開說明。因此我們認為，這對我們所有人而言都是一個很好的機會，可以做更多創新、更快創新，並擴大業務。所以我認為這是一股順風。

Great. Thanks for the time. And then a quick follow-up on the TSMC Cobalt inter-person agreement, it's like a $2 billion manufacturing agreement is any of this flowing through CapEx next year, and will the wafer supply there be dedicated to TSM for the first five years? Or do you have flexibility over there?

很好。謝謝你撥時間。接著一個簡短追問，關於台積電的Cobalt中介層（interposer）協議，這像是一份20億美元的製造協議；其中是否有任何部分會反映在明年的資本支出（CapEx）中？而且那邊的晶圓供應在前五年會不會是專供TSM？還是你們在那邊有彈性？

Yeah. So it's part of the CapEx that Sam has talked about. So definitely that's included within the $4 billion number that he shared. Look, our strategy in interpose -- it is a critical technology for all forms of high-performance compute, where you're integrating multiple die with HBM and so on. So we've had a road map for some time in that.

是的。這是Sam談到的資本支出的一部分。所以它確實包含在他分享的40億美元數字之內。你看，我們在中介層的策略——它是各種高效能運算的關鍵技術，因為你要把多顆晶粒與HBM等整合在一起。所以我們在這方面已經有一段時間的路線圖。

The TSMC agreement stands on its own, and that's accretive in its own right, but we're also engaging with a number of other customers for future interposer requirements. And these are getting increasingly sophisticated with integration of deep trench capacities and so on. So I think it's actually a very interesting area that we're spending R&D money and effort to actually build more for the future because it's still in our sweet spot in terms of technology mix.

與台積電的協議本身是獨立成立的，而且本身就具增益（accretive）。但同時，我們也正與多家其他客戶洽談未來的中介層需求。而且這些需求正變得越來越複雜，例如整合深溝槽（deep trench）能力等等。所以我認為這其實是一個非常有意思的領域；我們正在投入研發資金與努力，為未來建立更多能力，因為就技術組合而言，這仍然在我們的甜蜜點（sweet spot）之內。

Thank you, Tim.

謝謝你，Tim。

Atif Malik, Citi.

Atif Malik，花旗（Citi）。

Hi. Thank you for taking my questions. You guys have been busy with the partnership announcements with TSMC, Marvell, Monolithic Power. To what extent are these partnerships incremental to the long-term model I shared at the Investor Day?

嗨。謝謝讓我提問。你們最近很忙，宣布了與台積電、Marvell、Monolithic Power的合作夥伴關係。這些合作在多大程度上，是對我在投資人日分享的長期模型的增量貢獻？

I think yeah, Sam commented on that a bit earlier. The way I think about it, Atif is that we set out only 6 months ago, a set of strategic goals, which I think are both appropriate for the business and very much strategically align with where we have a right to win. And these are evidence points that we are winning in those directions with those customers. So too early to update the numbers of our strategic goals, but as Sam said, all of these give us increased confidence that we're on the right track to deliver and exceed those goals over time.

我想是的，Sam稍早有稍微評論過。Atif，我的看法是，我們在6個月前才訂出一套策略目標；我認為這些目標既適合公司業務，也在策略上與我們具備勝出權利（right to win）的方向高度一致。而這些合作就是我們正沿著那些方向、與那些客戶取得勝利的證據點。所以現在更新策略目標的數字還太早，但如Sam所說，這些都讓我們更有信心，我們正走在正確的軌道上，能隨時間推進達成並超越那些目標。

Great. And then, Sam, what is the satellite communications exposure within CID, and what's the gross margin profile?

很好。接著，Sam，CID之中衛星通訊（satellite communications）的曝險/比重是多少？毛利率輪廓如何？

Look, we don't break out the individual components. I will say that in 2025, we gave a little bit of color around SATCOM. We delivered about $100 million of revenue within that component of the end market from relative standing start in the prior year. You've seen publicly, we've made some announcements that associate ourselves with some of those key players within the satellite communications market more broadly.

你看，我們不會把各個組成項目單獨拆分揭露。我會說的是，在2025年，我們對SATCOM提供了一些色彩：相較於前一年幾乎從零起步，我們在該終端市場的那個組成部分交付了約1億美元的營收。你也在公開資訊中看到，我們發布了一些公告，讓外界知道我們與更廣泛衛星通訊市場中的一些關鍵玩家有所連結。

You've only got to look at lower orbit satellite launches and the continued momentum within that end market to see that it's an important growth vector within comms interim data center for us and look more broadly, this is a technology that we're servicing that demand which is accretive to the long-term goals that we've set out.

你只要看看低軌衛星的發射，以及該終端市場持續的動能，就會明白這對我們在「通訊與中介資料中心」領域而言是一個重要的成長向量；更廣泛地說，我們所服務的這項需求所對應的技術，對我們所設定的長期目標具有增益（accretive）。

Let me just add on my side. I mean, I think one of the reasons we're seeing strong long-term momentum in technologies like SiGe beyond the data center is actually things like satellite communications, right? When you want to launch satellite 300 miles in space and have a high performance high-bandwidth solution, bringing that bandwidth back to us, you need high-performance RF technologies. We've been doing that for a very long time, and SiGe is getting also traction from those areas as well. So it's very encouraging.

我這邊再補充一下。我的意思是，我認為我們之所以在資料中心以外的SiGe等技術上看到強勁的長期動能，其中一個原因其實就是像衛星通訊這類應用，對吧？當你要把衛星送到距離地面300英里的太空，並且要有高性能、高頻寬的解決方案，把那個頻寬回傳到地面，你就需要高性能的RF技術。我們做這件事已經很久了，而SiGe也正在從這些領域獲得更多牽引力。所以這非常令人鼓舞。

Thank you.

謝謝。

Chris Caso, Wolfe Research.

Chris Caso，Wolfe Research。

Yes, thank you. Good morning. My first question is with regard to the CapEx and the $4 billion number you mentioned, it's a big number relative to what you guys have spent in the past. So a few questions on that. One is, if you could discuss given some of the prepayments and credits in that, how much of that is cash out the door from GLOBALFOUNDRIES that would affect your cash flow for next year, just kind of separating out what's supported by others and what you're putting out yourself.

是的，謝謝。早安。我的第一個問題是關於資本支出（CapEx）以及你提到的40億美元數字；相較於你們過去的支出，這是一個很大的數字。所以我有幾個相關問題。第一個是，考量其中包含一些預付款與抵減（credits），你能否談談其中有多少是GLOBALFOUNDRIES實際付出去的現金、會影響你們明年的現金流？也就是把由他方支持的部分，與你們自己要支出的部分區分開來。

And then in terms of when that turns into revenue and kind of the timing of that? And I guess what I'd really like you to answer, which you probably won't, how much revenue gets generated by $4 billion CapEx, but I'll settle for the win.

接著，就這些投入何時會轉化為營收，以及其時點安排而言？我想我真正希望你回答的（你大概不會回答）是：40億美元資本支出能帶來多少營收，但我就先接受這個成果。

Yeah, Chris, I think that we've been relatively under investing from a CapEx point of view over the last, call it, three years. If you look at the aggregate over the last three years, it's only been a little north of revenue versus what we said previously was targeting a 20% net CapEx to revenue and what we reiterated earlier this year.

是的，Chris，我認為從資本支出（CapEx）的角度來看，過去大約三年我們的投資相對偏保守。如果看過去三年的合計，CapEx 只略高於營收的水準；而我們先前所說的目標是淨CapEx對營收比率20%，且我們在今年稍早也再次重申了這一點。

So the principal reason for that is the foundation of our investment decisions into capacity is demand. we've seen a step change function in that demand outlook as we go through this year. It's the technologies across Photonics, FDX, Advanced Packaging, SiGe, more recently, what we announced with TSMC on enterprises. So this has been the single biggest change in terms of how we think about that investment into the footprint.

造成這種情況的主要原因是：我們在產能上的投資決策，其基礎是需求。今年以來，我們看到需求展望出現了階躍式的變化。這涵蓋了光子（Photonics）、FDX、先進封裝（Advanced Packaging）、SiGe，以及更近期我們與台積電（TSMC）在企業領域所宣布的事項。因此，這是我們在思考對產能版圖投資方式上，最大的單一變化。

Now our customers recognize that as well. And actually, to Tim's earlier comments, they've been very, very constructive conversations in terms of the prepayment capacity reservation for incremental capacity discussions that we've been having. We expect those largely to close in, call it, the fourth quarter and into early 2027 as well.

我們的客戶也同樣認知到這一點。而且，呼應Tim先前的評論，我們就增量產能的預付款產能保留（prepayment capacity reservation）所進行的討論，對話非常、非常具建設性。我們預期這些大多會在第四季以及2027年初完成落地。

And so that $4 billion net number that I gave you, obviously, we don't net customer prepayments from that in terms of how we calculate net CapEx. So think of that as 4 billion net CapEx to GF, but within that $4 billion, we expect that we'll get about, call it, a third of that supported by customer prepayments. You won't see that flow through as a net to CapEx.

因此，我給你的40億美元淨額，顯然在我們計算淨CapEx時，不會把客戶預付款從中扣除。所以可把它視為GF的40億美元淨CapEx；但在這40億美元之中，我們預期約有三分之一會由客戶預付款支持。你不會看到它以CapEx淨額的形式反映出來。

But you will effectively see a flow-through towards our free cash flow. And so that's why we feel good about still remaining the targets around our free cash flow margin for 2027. And then look, the other dynamic, as you said, is around government grants and water support framework. If I take the ITC, as an example, that's now 35%. Now we don't get all of that 35% in the year.

但你會實際看到它流入並反映在我們的自由現金流上。這也是為什麼我們對於在2027年仍能維持自由現金流利潤率目標感到有信心。另外，如你所說，另一個變數是政府補助與相關支持架構。以投資稅額抵免（ITC）為例，現在是35%。當然，我們不會在同一年拿到全部35%。

In fact, some of what will get us a net to CapEx next year will be from ITC that was accrued in, call it, the 2025 time frame. We will, however, get as an offset to CapEx next year. The government grants that we've announced during the course of this year. Some of that is in relation to quantum computing.

事實上，明年能作為CapEx淨額抵減的一部分，將來自於大約在2025年期間累積（accrued）的ITC。不過，我們確實會在明年把它作為CapEx的抵減項。此外，我們在今年期間已宣布的政府補助，也會在明年作為CapEx的抵減。其中一部分與量子運算（quantum computing）相關。

We've also got the $955 million repurposing from the original chip offices. About 80% of that -- we expect we'll allocate towards CapEx. Not all of that will be allocated in 2027 as an offset, but we'll probably get the majority of that coming through. in 2027 as well. So all points towards good indicators from a free cash point of view. And then look, the way we think about the ROI is really a function of payback.

我們也有來自原始晶片辦公室（chip offices）的9.55億美元重新用途（repurposing）資金。其中約80%——我們預期會配置到CapEx。並非全部都會在2027年作為抵減項入帳，但我們可能會看到其中大部分也在2027年反映出來。因此，從自由現金流的角度來看，各項指標都指向良好。另外，我們看ROI的方式，主要取決於回收期（payback）。

And when you take that combination of the government grant, the ITC, the offset to our overall CapEx burden, we see strong paybacks inside a sort of two-year time horizon for most of those technologies that we outlined. So it forms an important part of our investment decision-making and given the margin profile and the mix associated with these corridors, we see a strong ROI and a relatively quick ramp to demand and revenue.

當你把政府補助、ITC，以及對我們整體CapEx負擔的抵減效果合併來看，對於我們所概述的大多數技術，我們看到在大約兩年的時間範圍內就能有很強的回收。因此，這是我們投資決策的重要組成部分；再加上這些走廊（corridors）所對應的毛利結構與產品組合，我們看到強勁的ROI，以及需求與營收相對快速的爬坡。

Obviously, it's really a function of how quickly you can get these tools installed, qualified and ramped that the beauty of our expansion footprint is that we can do it within our existing four walls. So we're not relying like some others are around building incremental product capacity to it, we could do within our reducting four walls. So that accelerates the time to revenue as well.

顯然，這很大程度取決於你能多快把這些設備（tools）安裝、認證（qualified）並完成量產爬坡；而我們擴產版圖的優勢在於，我們可以在既有廠房的四面牆內完成。因此我們不像其他一些公司那樣，還得依賴新建額外的產品產能；我們可以在既有廠房內完成。這也加速了營收實現的時間。

And maybe I'll just add. I mean, the way we've been thinking about this is -- it's a unique moment for us to accelerate our business with the participation of our customers, with the participation of government, maintaining a very strong balance sheet, maintaining our free cash flow generation objectives. It's the right way to navigate the current situation that we're in.

我再補充一下。我們的思考方式是——這對我們而言是一個獨特的時刻：在客戶參與、政府參與之下加速業務，同時維持非常強健的資產負債表，並維持我們自由現金流產生的目標。這是我們在當前所處情勢下，最合適的應對方式。

Thank you. This does conclude the question-and-answer session of today's program. I'd like to hand the program back to Eric Chow for any further remarks.

謝謝。以上為今天節目的問答環節結束。我想把節目交回給Eric Chow，請他做進一步的結語。

Thank you very much for joining today. We look forward to seeing you at the UBS conference on December 1. Thanks again.

非常感謝各位今天的參與。我們期待在12月1日的UBS會議上與各位見面。再次感謝。

Thank you, ladies and gentlemen, for your participation in today's conference. This does conclude the program. You may now disconnect. Good day.

感謝各位女士、先生參與今天的電話會議。本節目到此結束。您現在可以掛線。祝各位有美好的一天。