Credo Technology Group Holding Ltd (CRDO) 2026 Q4 法說會逐字稿

Ladies and gentlemen, thank you for standing by. (Operator Instructions) I would now like to turn the conference over to Dan O'Neil. Please go ahead, sir.

各位女士、先生，感謝您耐心等候。(接線員指示) 我現在把電話會議交給 Dan O'Neil。先生，請開始。

Good afternoon. Thank you all for joining our fourth quarter fiscal 2026 earnings call. Today, I am joined by Bill Brennan, Credo's Chief Executive Officer; and Dan Fleming, Credo's Chief Financial Officer. During this call, we will make certain forward-looking statements. These forward-looking statements are subject to risks and uncertainties discussed in detail in our documents filed with the SEC.

下午好。感謝各位參加我們 2026 會計年度第四季財報電話會議。今天與我一同出席的有 Credo 執行長 Bill Brennan，以及 Credo 財務長 Dan Fleming。在本次電話會議中，我們將作出若干前瞻性陳述。這些前瞻性陳述受我們向美國證券交易委員會（SEC）提交文件中所詳述之風險與不確定性影響。

These documents can be found in the Investor Relations portion of the company's website. It is not possible for the company's management to predict all risks nor can the company assess the impact of all factors on its business or the extent to which any factor or combination of factors may cause actual results to differ materially from those contained in any forward-looking statement. Given these risks, uncertainties and assumptions, the forward-looking events discussed during this call may not occur, and actual results could differ materially and adversely from those anticipated, implied or inferred.

這些文件可在公司網站的投資人關係（Investor Relations）專區查閱。公司管理層無法預測所有風險，公司亦無法評估所有因素對其業務的影響，或任何單一因素或多項因素組合可能導致實際結果與任何前瞻性陳述所載內容出現重大差異的程度。基於上述風險、不確定性與假設，本次電話會議中討論的前瞻性事件可能不會發生，且實際結果可能與所預期、暗示或推論者存在重大且不利的差異。

The company undertakes no obligation to publicly update forward-looking statements for any reason after the date of this call to conform these statements to changes in the company's expectations or to actual results, except as required by law.

除法律要求外，公司在本次電話會議日期之後，無義務因任何原因公開更新前瞻性陳述，以使該等陳述符合公司預期之變化或實際結果。

Also, during this call, we will refer to certain non-GAAP financial measures, which we consider to be important measures of the company's performance. These non-GAAP financial measures are provided in addition to and not as a substitute for or superior to financial performance prepared in accordance with US GAAP. A discussion of why we use non-GAAP financial measures and reconciliations between our GAAP and non-GAAP financial measures is available in the earnings release we issued today, which can be accessed using the Investor Relations portion of our website.

此外，在本次電話會議中，我們將提及若干非 GAAP 財務衡量指標，我們認為這些是衡量公司表現的重要指標。這些非 GAAP 財務衡量指標係作為依照美國 GAAP 編製之財務表現的補充，而非替代，亦不應被視為更優。關於我們為何使用非 GAAP 財務衡量指標，以及 GAAP 與非 GAAP 財務衡量指標之間的調節表，已載於我們今日發布的財報新聞稿中，可透過本公司網站的投資人關係專區取得。

I will now turn the call over to our CEO. Bill?

我現在把電話會議交給我們的執行長。Bill？

Thanks, Dan, and thank you all for joining our fourth quarter and full fiscal year '26 earnings call. I'll begin with a review of our fiscal '26 performance, discuss the major developments across our business and share our perspective on the opportunities ahead. Dan Fleming, our Chief Financial Officer, will then provide additional detail on our Q4 and fiscal year '26 results, along with guidance for the first quarter of our fiscal '27. We'll then open the call for questions. Fiscal '26 marked another defining year for Credo.

謝謝你，Dan，也感謝各位參加我們 2026 會計年度第四季與全年財報電話會議。我將先回顧我們 2026 會計年度的表現，說明公司各項業務的重大進展，並分享我們對未來機會的看法。接著，我們的財務長 Dan Fleming 將就 2026 會計年度第四季與全年業績提供更多細節，並提供 2027 會計年度第一季的展望指引。之後我們將開放提問。2026 會計年度是 Credo 另一個具指標性的年度。

Revenue exceeded $1.3 billion, more than tripling year-over-year, while non-GAAP net income increased more than 5 times to $662 million. Very few semiconductor companies have scaled at this pace while sustaining product leadership, strong margins and operational execution. In the fourth quarter of fiscal '26, revenue reached a record $437 million. Notably, our revenue in the quarter exceeded our entire fiscal '25 revenue. Q4 non-GAAP gross margin was 68.3%.

營收超過 13 億美元，年增逾三倍；同時，非 GAAP 淨利增加逾 5 倍至 6.62 億美元。能以如此速度擴張，同時維持產品領先地位、強勁毛利率與營運執行力的半導體公司屈指可數。在 2026 會計年度第四季，營收達到創紀錄的 4.37 億美元。值得注意的是，本季營收已超過我們整個 2025 會計年度的營收。第四季非 GAAP 毛利率為 68.3%。

Non-GAAP net income grew to $227 million and was more than 30% greater than our revenue in the year ago quarter. Producing these results required incredible effort and expertise, and I want to sincerely thank team Credo for their continued stellar performance. These results reflect Credo's ability to capitalize on a fundamental shift occurring across AI infrastructure. As AI clusters scale from tens of thousands to hundreds of thousands of GPUs, connectivity is no longer just about bandwidth. Reliability, power efficiency, signal integrity and telemetry have become critical architectural requirements.

非 GAAP 淨利成長至 2.27 億美元，且比去年同期季度的營收高出逾 30%。達成這些成果需要非凡的努力與專業能力，我要由衷感謝 Credo 團隊持續卓越的表現。這些成果反映 Credo 能夠把握 AI 基礎設施正在發生的根本性轉變。隨著 AI 叢集從數萬顆 GPU 擴展到數十萬顆 GPU，連線不再只是頻寬的問題。可靠性、能效、訊號完整性與遙測已成為關鍵的架構需求。

Today's AI infrastructure is increasingly constrained not by compute, but by the reliability and efficiency of the connectivity fabric tying these systems together. Over the past several years, AI network reliability has become Credo's North Star. Our road map, our product investments, our software architecture and our system-level approach have all been built around helping customers accelerate cluster bring up, maximize GPU utilization and maintain stable operation at unprecedented scale. Credo is purpose-built for this transition. Our strategy is centered on delivering connectivity solutions across the full spectrum of AI infrastructure from die-to-die and chip-to-chip connectivity to multi-rack scale copper and to road scale and facility-wide optical interconnect.

當今的 AI 基礎設施日益受限的不是運算能力，而是將這些系統連結在一起的連線網路（fabric）的可靠性與效率。過去幾年，AI 網路可靠性已成為 Credo 的北極星指標。我們的產品藍圖、產品投資、軟體架構與系統層級方法，皆圍繞著協助客戶加速叢集上線（bring up）、最大化 GPU 使用率，並在前所未有的規模下維持穩定運作。Credo 正是為這項轉型而打造。我們的策略聚焦於提供涵蓋 AI 基礎設施全光譜的連線解決方案，從 die-to-die 與 chip-to-chip 連線，到多機架規模的銅纜，再到機架規模與全機房範圍的光互連。

By extending both inward towards the silicon and outward across the data center, we've positioned Credo to become a foundational network architecture partner for our customers. Importantly, hyperscale and Neo cloud operators increasingly want partners capable of delivering multiple generations of connectivity solutions with deep system-level integration. This is where Credo differentiates itself through our vertically integrated approach, spanning core SerDes technology, silicon and system-level solutions, firmware and telemetry software and operational execution. I'll now discuss our businesses in more detail. First, regarding active electrical cables.

透過同時向內延伸至矽晶層、向外延伸至整個資料中心，我們已將 Credo 定位為客戶的基礎網路架構合作夥伴。重要的是，超大規模（hyperscale）與新型雲端（Neo cloud）營運商愈來愈希望合作夥伴能提供跨多個世代的連線解決方案，並具備深度的系統層級整合能力。Credo 的差異化在於我們的垂直整合方法，涵蓋核心 SerDes 技術、矽晶與系統層級解決方案、韌體與遙測軟體，以及營運執行。接下來我將更詳細說明我們的各項業務。首先，關於主動式電纜（Active Electrical Cables）。

Our AEC business remains a core growth engine for the company, and we continue to see substantial long-term opportunity ahead. As AI clusters scale, reliability and power efficiency have become primary design constraints. AECs have become the preferred solution for in-rack connectivity and for many multi-rack deployments up to seven meters. Credo zero-flap AECs deliver up to 1,000 times greater reliability than commodity laser-based optical modules while consuming much less power. In environments where cluster downtime can cost millions of dollars and delay AI deployment schedules, network reliability matters more than ever.

我們的 AEC 業務仍是公司的核心成長引擎，我們也持續看到可觀的長期機會。隨著 AI 叢集擴張，可靠性與能效已成為主要設計限制。AEC 已成為機架內連線，以及許多最長達 7 公尺的多機架部署的首選方案。Credo 的 zero-flap AEC 相較於一般以雷射為基礎的光模組，可靠性可高達 1,000 倍，同時耗電更低。在叢集停機可能造成數百萬美元成本並延誤 AI 部署時程的環境中，網路可靠性比以往任何時候都更為重要。

We continue to see strong customer adoption across hyperscaler and Neo cloud operators, both at 100-gig per lane and emerging 200-gig per lane deployments. Our vertically integrated model positions us well for continued leadership as both lane speeds and cluster complexity increase. We also remain on track with our PCIe Gen 6 AEC family where customer engagement and design activity continue to strengthen. Now turning to optics. We believe fiscal '27 represents an inflection point for Credo's optical business.

我們持續看到超大規模客戶與新型雲端營運商的強勁採用動能，涵蓋每通道 100G 以及新興的每通道 200G 部署。隨著通道速率與叢集複雜度提升，我們的垂直整合模式使我們具備持續領先的有利位置。我們在 PCIe Gen 6 AEC 產品系列方面也持續按計畫推進，客戶互動與設計活動持續增強。接著談到光學。我們相信 2027 會計年度將是 Credo 光學業務的轉折點。

First, at an Optical DSP component level, we see momentum in both design wins and revenue contribution. We're looking forward to continued growth in this product family, and we've received excellent customer feedback on the solutions we announced last quarter, both Robin, a highly optimized DSP at 100 gig per lane and Cardinal, a leading-edge DSP at 200 gig per lane.

首先，在光學 DSP 元件層面，我們在設計導入（design wins）與營收貢獻方面都看到動能。我們期待此產品系列持續成長，且客戶對我們上季宣布的解決方案給予極佳回饋，包括 Robin（每通道 100G 的高度最佳化 DSP）以及 Cardinal（每通道 200G 的先進 DSP）。

Next, the acquisition of Dust Photonics, which closed last week, significantly expands our optics position with highly differentiated silicon photonics PIC technology. Dust brings strong design win momentum and a portfolio spanning 800 gig and 1.6T solutions, along with a road map to 3.2 terabits per second and beyond. Importantly, their architecture enables simplified optical designs with substantially fewer lasers.

其次，上週完成交割的 Dust Photonics 收購案，透過高度差異化的矽光子 PIC 技術，顯著擴大我們在光學領域的布局。Dust 帶來強勁的設計導入動能，以及涵蓋 800G 與 1.6T 解決方案的產品組合，並具備通往每秒 3.2 兆位元（terabits per second）及更高規格的產品藍圖。重要的是，其架構可實現更簡化的光學設計，並大幅減少雷射數量。

In addition to enabling better reliability, power efficiency and cost, laser count reduction can ease the industry supply chain limitations. The Dust Silicon Photonics road map also provides a direct path to CPO and NPO architectures, allowing us to address a broad range of customer requirements as AI deployments evolve in the scale-up network domain.

除了提升可靠性、能效與成本表現之外，降低雷射數量也有助於緩解產業供應鏈限制。Dust 的矽光子產品藍圖亦提供通往 CPO 與 NPO 架構的直接路徑，使我們能在 AI 部署於 scale-up 網路領域持續演進時，滿足廣泛的客戶需求。

Based on current customer engagements, initial revenue for CPO and NPO designs is expected in our fiscal '28. Finally, our Zero-Flat optics platform continues to gain strong traction as customers increasingly prioritize network reliability. With the addition of SiPho PIC technology to our ZeroFlat optics platform, we can now control a significantly larger portion of the optical stack, extending visibility deeper into optical link behavior and performance.

根據目前的客戶洽談情況，我們預期 CPO 與 NPO 設計的初始營收將在本公司 2028 會計年度出現。最後，隨著客戶日益將網路可靠性置於優先地位，我們的 Zero-Flat 光學平台持續獲得強勁動能。隨著將 SiPho PIC 技術加入我們的 ZeroFlat 光學平台，我們如今能夠掌控光學堆疊中顯著更大的一部分，將可視性延伸至更深層的光鏈路行為與效能。

The tighter DSP to PIC integration enables richer telemetry, enhanced diagnostics and more intelligent system-level optimization. By combining optimized hardware and our pilot software with Switch level SDK integration, ZeroFlap optics continuously monitor link health and autonomously detect and mitigate link instability conditions before impacting the cluster. Results have shown a meaningful improvement in network reliability, time to cluster stability and long-term uptime. In summary, we're very enthusiastic about the prospects of our optical portfolio. In fiscal '27, we expect our optical DSPs, SiPho PICs and ZeroFlap optics will each contribute more than $100 million of revenue and in total, more than $600 million of revenue, with this expected ramp accelerating in the second half of the year.

更緊密的 DSP 與 PIC 整合可帶來更豐富的遙測、更強化的診斷，以及更智慧的系統層級最佳化。透過結合最佳化硬體與我們的試點軟體，並整合至交換器層級的 SDK，ZeroFlap 光學可持續監測鏈路健康狀態，並在影響叢集之前，自主偵測與緩解鏈路不穩定狀況。結果顯示，網路可靠性、叢集達到穩定所需時間以及長期正常運行時間均有顯著改善。總結而言，我們對光學產品組合的前景感到非常振奮。在 2027 會計年度，我們預期光學 DSP、SiPho PIC 與 ZeroFlap 光學各自將貢獻超過 1 億美元營收，合計超過 6 億美元營收，且預期此一爬坡將在下半年加速。

Based on customer and market feedback, we believe this portfolio will deliver sustained rapid growth in future years. Now regarding retimers. Our retimer business also continues to gain momentum. We're seeing strong growth for retimers at 100 gig and 200 gigabits per second per lane as well as increasing traction for our PCIe Gen 6 retimers. Our Blue Heron 200-gig per lane retimer was purpose-built for scale-out and emerging scale-up networks.

根據客戶與市場回饋，我們相信此產品組合將在未來年度帶來持續的高速成長。接下來談重定時器（retimer）。我們的重定時器業務也持續累積動能。我們看到每通道 100G 與 200Gbps 重定時器的強勁成長，同時也看到我們 PCIe Gen 6 重定時器的採用動能提升。我們的 Blue Heron 每通道 200G 重定時器是為橫向擴展（scale-out）與新興縱向擴展（scale-up）網路量身打造。

We're seeing increased interest and demand as the device combines support for the broad range of 200-gig per lane protocols, including Ethernet, UA Link and ESUN. As AI infrastructure becomes increasingly complex and protocol diversity expands, we believe our system-level expertise and software integration capabilities position us well for continued share gains. Now moving to our emerging growth categories. We also continue to make strong progress across our newer growth vectors, including active LED cables and OmniConnect. Our ALC solutions will extend the reliability and power profile of AECs into row scale optical connectivity by replacing traditional lasers with MicroLED technology.

隨著該裝置支援廣泛的每通道 200G 協定（包括 Ethernet、UA Link 與 ESUN），我們看到市場興趣與需求增加。隨著 AI 基礎設施日益複雜、協定多樣性擴大，我們相信我們的系統層級專業能力與軟體整合能力，將使我們在持續提升市占方面具備良好定位。接著談我們的新興成長類別。我們在較新的成長向量上也持續取得強勁進展，包括主動式 LED 纜線（active LED cables）與 OmniConnect。我們的 ALC 解決方案將以 MicroLED 技術取代傳統雷射，將 AEC 的可靠性與功耗特性延伸至機櫃列（row scale）的光學連接。

This creates a highly differentiated connectivity category capable of delivering AEC class reliability with optical reach of up to 30 meters. Our OmniConnect family expands our solutions inward towards the silicon. Our first gearbox solution Weaver, will address growing memory bandwidth and density challenges by enabling substantially higher memory I/O density and more flexible architectures. Customer engagement remains strong, especially around next-generation inference designs. We continue to expect production ramps for both ALC and OmniConnect solutions beginning in our fiscal '28.

這創造出一個高度差異化的連接類別，能提供 AEC 等級的可靠性，並具備最遠達 30 公尺的光學傳輸距離。我們的 OmniConnect 產品家族將我們的解決方案向內延伸至更靠近矽晶。我們的第一款齒輪箱（gearbox）解決方案 Weaver，將透過實現顯著更高的記憶體 I/O 密度與更具彈性的架構，來因應日益增加的記憶體頻寬與密度挑戰。客戶互動仍然強勁，特別是在下一代推論設計方面。我們仍預期 ALC 與 OmniConnect 解決方案將自本公司 2028 會計年度開始進入量產爬坡。

And in conclusion, the data center connectivity market continues to evolve as AI scales toward gigawatt class deployments and increasingly dense architectures, network reliability becomes even more critical. Even isolated link instabilities can impact cluster bring-up times, GPU utilization and overall system availability. That's why reliability has been Credo's North Star over the past several years. It drives our system-level philosophy, our telemetry-first software architecture and the investments across the full spectrum of our solutions. Fiscal '26 was another transformative year for Credo, and yet we believe we are still in the early innings of the opportunity ahead.

最後總結，隨著 AI 規模朝向 GW（吉瓦）等級部署與更高密度架構發展，資料中心連接市場持續演進，網路可靠性變得更加關鍵。即便是孤立的鏈路不穩定，也可能影響叢集啟動時間、GPU 使用率與整體系統可用性。這也是為什麼在過去幾年，可靠性一直是 Credo 的北極星。它驅動我們的系統層級理念、以遙測優先的軟體架構，以及我們在全系列解決方案上的投資。2026 會計年度對 Credo 而言是另一個具轉型意義的一年，但我們相信，面前的機會仍處於早期階段。

With that, I'll turn the call over to Dan Fleming for a detailed financial review and our outlook for Q1 fiscal '27.

接下來，我將把電話會議交給 Dan Fleming，進行更詳細的財務回顧，並說明我們對 2027 會計年度第一季的展望。

Thank you, Bill, and good afternoon. I will first provide a financial summary of our fiscal year '26, then review our Q4 results and finally, discuss our outlook for Q1 and provide some color on our expectations for fiscal year '27. Revenue for fiscal year '26 was another record at $1.3 billion, up 206% year-over-year. Gross margin for the year was 68.1%, up 310 basis points year-over-year. Our operating margin improved by 2,144 basis points as we continued to generate considerable top line leverage driven by growth in our products while growing operating expenses considerably slower than revenue.

謝謝你，Bill，大家下午好。我將先提供 2026 會計年度的財務摘要，接著回顧第四季業績，最後討論我們對 2027 會計年度第一季的展望，並補充一些我們對 2027 會計年度的預期。2026 會計年度營收再創新高，達 13 億美元，年增 206%。全年毛利率為 68.1%，年增 310 個基點。隨著我們在產品成長帶動下持續產生可觀的營收槓桿，且營業費用增速顯著低於營收增速，我們的營業利益率提升了 2,144 個基點。

That step-up in profitability flowed through to the bottom line as we reported earnings per share of $3.46 for the year, a $2.76 improvement or up 392% over the prior year. In fiscal year '26, Credo not only delivered the dramatic growth, which we had forecast, but we also demonstrated the considerable earnings power in our business model. Moving on to the fourth quarter. In Q4, we reported revenue of $437 million, up 7% sequentially and above the high end of our guidance range. Year-over-year, revenue grew 157% and notably, Q4 alone exceeded our total fiscal year '25 revenue.

獲利能力的提升也反映在最終盈餘上：我們全年每股盈餘為 3.46 美元，較前一年增加 2.76 美元，或年增 392%。在 2026 會計年度，Credo 不僅實現了我們所預測的顯著成長，也展現了我們商業模式可觀的獲利能力。接著談第四季。第四季我們營收為 4.37 億美元，季增 7%，並高於我們財測區間的上緣。年增方面，營收成長 157%；值得注意的是，單第四季的營收就超過我們 2025 會計年度全年的營收。

Q4 marks another revenue record, driven by substantial year-over-year growth across four domestic customers. Our top four end customers each came in at or greater than 10% of revenue in Q4. As a reminder, customer mix will vary from quarter-to-quarter. We continue to expect that three to four customers will be greater than 10% of revenue in the coming quarters and fiscal year, and we continue to make progress in diversifying our revenue base across hyperscalers, Neo clouds and other customers. Our team delivered Q4 non-GAAP gross margin of 68.3%, above the high end of our guidance range.

第四季再度創下營收新高，主要來自四家國內客戶的顯著年增成長所帶動。我們的前四大終端客戶在第四季各自占營收比重皆達到或超過 10%。提醒一下，客戶組合會隨季度而變動。我們仍預期在接下來幾個季度與本會計年度中，將有三到四家客戶的營收占比超過 10%；同時，我們也持續在超大規模客戶（hyperscalers）、新型雲端（Neo clouds）及其他客戶之間推進營收來源多元化。我們團隊第四季非 GAAP 毛利率為 68.3%，高於我們財測區間的上緣。

Total non-GAAP operating expenses in the fourth quarter were $81.7 million, above the high end of our guidance range due to our strong R&D investment and up 6% sequentially. Our non-GAAP operating income was $216.7 million in Q4 compared to non-GAAP operating income of $201.8 million in Q3. Our non-GAAP operating margin was 49.6% in the quarter, flat sequentially. Our bottom line once again demonstrated the substantial leverage we are delivering in the business. Our non-GAAP net income was $226.7 million in the quarter, a record high and a 9% sequential increase compared to non-GAAP net income of $208.8 million in Q3.

第四季非 GAAP 營業費用總額為 8,170 萬美元，由於我們強勁的研發投資而高於財測區間上緣，並季增 6%。第四季非 GAAP 營業利益為 2.167 億美元，相較第三季非 GAAP 營業利益 2.018 億美元。本季非 GAAP 營業利益率為 49.6%，與前一季持平。我們的最終獲利再次展現出公司所帶來的顯著槓桿效益。本季非 GAAP 淨利為 2.267 億美元，創歷史新高，較第三季非 GAAP 淨利 2.088 億美元季增 9%。

Our Q4 non-GAAP net income more than tripled year-over-year and was 33% higher than our revenue in the fourth quarter of last year, which most clearly demonstrates the magnitude of our top line growth, strong gross margins and disciplined approach to managing operating expenses. Our non-GAAP net margin was 51.9% in the quarter. Cash flow from operations in the fourth quarter was a record $182.2 million, up $16 million sequentially. CapEx was $4.8 million in the quarter, and free cash flow was $177.5 million, up more than $37.8 million from the third quarter. We ended the quarter with cash and equivalents of $1.4 billion, an increase of $141.8 million from the third quarter, driven by our strong free cash flow.

我們第四季非 GAAP 淨利較去年同期成長逾三倍，且比去年第四季營收高出 33%，最能清楚展現我們營收高速成長、強勁毛利率，以及在營業費用管理上的紀律所帶來的成果。本季非 GAAP 淨利率為 51.9%。第四季營運活動現金流為創紀錄的 1.822 億美元，季增 1,600 萬美元。本季資本支出（CapEx）為 480 萬美元，自由現金流為 1.775 億美元，較第三季增加超過 3,780 萬美元。本季期末現金及約當現金為 14 億美元，較第三季增加 1.418 億美元，主要由強勁的自由現金流所帶動。

We remain well capitalized to continue investing in our growth opportunities while maintaining a substantial cash buffer. Our Q4 ending inventory was $250.8 million, up $42.9 million sequentially. Now turning to our guidance. We currently expect revenue in Q1 of fiscal '27 to be between $465 million and $475 million. We expect Q1 non-GAAP gross margin to be within a range of 67% to 69%.

我們仍維持充足的資本水準，以便在保有可觀現金緩衝的同時，持續投資於我們的成長機會。我們第四季期末存貨為2.508億美元，較前一季增加4,290萬美元。現在轉到我們的財測指引。我們目前預期27會計年度第一季營收介於4.65億至4.75億美元之間。我們預期第一季非GAAP毛利率將落在67%至69%的區間。

We expect Q1 non-GAAP operating expenses to be between $86 million and $90 million and we expect Q1 diluted weighted average share count to be approximately 199 million shares. These expectations are based on the current tariff regime, which remains fluid. We were pleased to see fiscal year '26 play out better than we had expected. The rapid shift to AI workloads continue to drive new and broad-based customer engagement, and we executed well to grow at more than twice the rate we expected at the beginning of the year. As we begin fiscal year '27, we expect mid-single-digit sequential growth in the first half with an inflection beginning in the second half.

我們預期第一季非GAAP營業費用介於8,600萬至9,000萬美元之間，並預期第一季稀釋後加權平均股數約為1.99億股。這些預期係基於目前的關稅制度，而該制度仍在變動中。我們很高興看到26會計年度的表現優於我們原先的預期。AI工作負載的快速轉移持續帶動新的且更廣泛的客戶參與，我們也執行得宜，使成長速度超過年初預期的兩倍以上。隨著我們進入27會計年度，我們預期上半年將呈現中個位數的季增率，並在下半年開始出現拐點。

That inflection is bolstered by more than $600 million in optical revenue with zero-flap optics, silicon photonics PICs and optical DSPs each contributing more than $100 million driving more than 80% year-over-year total revenue growth for the full year. We expect non-GAAP gross margin in fiscal year '27 to be broadly consistent with fiscal year '26 levels. We expect non-GAAP operating expenses to increase approximately 50% year-over-year, well below our revenue growth rate as we continue to invest in R&D to support the new product development and address the significant growth opportunities ahead. As a result, we expect our non-GAAP net margin to be in the vicinity of 50%.

該拐點將由超過6億美元的光學營收所支撐，其中零翻蓋（zero-flap）光學、矽光子PIC以及光學DSP各自貢獻超過1億美元，推動全年總營收年增超過80%。我們預期27會計年度的非GAAP毛利率將大致與26會計年度水準一致。我們預期非GAAP營業費用年增約50%，顯著低於我們的營收成長率，因為我們將持續投資研發以支援新產品開發，並掌握前方重大的成長機會。因此，我們預期非GAAP淨利率將在約50%左右。

And with that, I will open it up for questions.

接下來，我們開放提問。

(Operator Instructions)

(接線員指示)

Tom O'Malley, Barclays.

Tom O'Malley，Barclays。

I just wanted to dive in into the guidance. Is there any optical revenue ZF Optics or Dust Photonics in that guidance? And if so, how much? And then you mentioned $100 million across a couple of different vectors in the optics business for this year. One of them was over $100 million in DSPs.

我想深入了解一下指引。這個指引裡是否包含任何光學營收，例如ZF Optics或Dust Photonics？如果有，大概是多少？另外你提到今年光學業務在幾個不同方向合計有超過1億美元的貢獻。其中一項是DSP超過1億美元。

Have you seen a change in the attitude of the market with procuring discrete DSPs? I assume you mean discrete DSP sales versus in the module. And just that change over the last couple of months, do you see that accelerating? And just clarify if those are being sold to customers directly or inside of modules.

你們是否看到市場在採購獨立式DSP方面的態度有所改變？我猜你指的是獨立式DSP的銷售，而不是在模組裡的。就過去幾個月的變化而言，你覺得這種趨勢會加速嗎？另外也請釐清，這些是直接賣給客戶，還是包含在模組內一起銷售。

Yes, Tom, for the full year outlook that we provided, which is revenue growing at 80%-plus year-over-year, a component of that is our entire optical portfolio, which we described three different legs of, which we expect to each be over $100 million. So as you say, it's the discrete optical DSPs over $100 million in the year and ZF Optics over $100 million in the year and the silicon photonics picks over $100 million per year. So it is in the overall yearly guidance. And if you kind of break that down or dive into that a little bit deeper, what you'll see is based on that guidance, if you look at absolute dollars year-over-year expectation for fiscal '27, about half of that growth in absolute dollars is coming from our optical portfolio and about half of that is coming from our existing copper portfolio, predominantly AECs, but also retimers.

是的，Tom，就我們提供的全年展望而言，也就是營收年增80%以上，其中一個組成部分是我們完整的光學產品組合；我們將其描述為三個不同支柱，且我們預期每一個都將超過1億美元。因此如你所說，全年獨立式光學DSP超過1億美元、ZF Optics全年超過1億美元，以及矽光子PIC全年超過1億美元。所以這些都包含在整體的全年指引之中。若你進一步拆解或更深入來看，依據該指引，若以27會計年度年對年絕對金額的成長來看，大約一半的絕對金額成長來自我們的光學產品組合，另一半來自我們既有的銅互連產品組合，主要是AEC，但也包含retimer。

Just to be clear, Tom, the three categories that Dan just described, in total, we are expecting more than $600 million of revenue contribution. So each one will each contribute more than $100 million, but in total, it will be more than $600 million.

Tom，為了更清楚說明，Dan剛才描述的三個類別合計，我們預期營收貢獻將超過6億美元。因此每一項都會貢獻超過1億美元，但合計將超過6億美元。

Tore Svanberg, Stifel.

Tore Svanberg，Stifel。

Yes. Congratulations on another record. Maybe to follow up on the optical revenue for fiscal '27. So Bill, I mean, I get it, right? I mean, $100 million-plus, but you are guiding to $600 million total plus. So which one within the three segments skews higher because obviously, one of them is going to have to be significantly more than $100 million --

是的。恭喜再創新高。我想追問一下27會計年度的光學營收。所以Bill，我的意思是，我理解，對吧？每個都超過1億美元，但你們指引是合計超過6億美元。那麼在這三個區塊中，哪一個會占比較高？因為很明顯其中一個必須顯著高於1億美元--

I think the important message there is that we're going to see growth broadly across the three segments. You're right that there will be one that has the largest revenue. But I think that we can clearly state that all three categories are growing very quickly. In fact, more than the growth that Dan guided for, for the whole year, right? So each one of these categories independently is going to grow more than 80% year-over-year.

我認為這裡最重要的訊息是，我們將在這三個區塊都看到廣泛的成長。你說得對，其中會有一個擁有最高的營收。但我認為我們可以清楚地表示，這三個類別都在非常快速地成長。事實上，成長速度會高於Dan對全年所給的指引，對吧？因此，這三個類別各自獨立來看，年增都會超過80%。

Joseph Cardoso, JPMorgan.

Joseph Cardoso，JPMorgan。

Maybe another one on the guidance, but thinking about it from a first half versus second half dynamic, like how should we think about the largest contribution of growth from a portfolio perspective from a half-over-half basis, including AECs, the optical solutions, et cetera, as we progress into the second half? And then as a second part of that question, like how should we be thinking about the implications to margins as these inflect into the back half?

我也想再問一個關於指引的問題，但從上半年對下半年的動態來思考：隨著我們進入下半年，從產品組合的角度、以半年度對半年度的基礎來看，成長的最大貢獻應該如何理解（包括AEC、光學解決方案等）？另外作為第二部分，當這些在下半年出現拐點時，對毛利率的影響我們應該如何看待？

So similar to how I described the absolute dollar contribution of growth. In the first half of the year, we guided mid-single digits between our Q1 guide, and we can expect something similar in our Q2 guide when that comes. That's largely driven by increases in our current portfolio that has ramped substantially, which is AEC predominant. But in the second half, as we described, a lot of that inflection is being driven by our optical portfolio, which is, again, just to reiterate, the zero flap optics ramp plus silicon photonics ramp, plus our existing optical DSP portfolio gaining significant traction this year versus fiscal '26.

所以，類似我先前描述的絕對金額成長貢獻。在上半年，我們在第一季指引中給的是中個位數成長，而當我們之後提供第二季指引時，也可以預期會是類似的水準。這主要是由我們目前已大幅放量的既有產品組合所帶動，其中以AEC為主。但在下半年，如我們所描述，許多拐點將由我們的光學產品組合所驅動；再次重申，包括零翻蓋（zero-flap）光學的放量、矽光子放量，以及我們既有的光學DSP產品組合在今年相較26會計年度獲得顯著的市場動能。

(Operator Instructions)

(接線員指示)

Sean O'Loughlin, TD Cowen.

Sean O'Loughlin，TD Cowen。

Congrats on the results and the guidance. I wanted to ask about the supply side of things. I think it's required that at least one of us ask about supply on every earnings call now. Back at OFC, we had talked about some of the work you were doing to secure some of the supply for the ZF Optics ramp. Maybe if you could just talk about any updates there? And then if there's anything notable as well on the AEC side, it would be great to hear about the supply situation there.

恭喜這次的成果與指引。我想問供應端的情況。我想現在幾乎每場法說至少都要有人問一次供應鏈。在OFC時，我們談到你們為ZF Optics放量所做的一些供應保障工作。也許你可以談談這方面是否有任何最新進展？另外，如果AEC那邊也有任何值得注意的地方，也希望能聽聽那邊的供應狀況。

Generally speaking, I think that supply chain discussions are going to become stay popular throughout the next year or even longer. There is a significant tightness in the supply chain, and that's why it's important to understand just the capability and the depth that we've got both on the silicon operations team as well as our systems operations team, and that includes AECs and ZF Optics.

一般來說，我認為供應鏈的討論在接下來一年甚至更久都會持續熱門。供應鏈存在顯著的緊俏，因此理解我們在矽相關營運團隊以及系統營運團隊（涵蓋AEC與ZF Optics）所具備的能力與深度就很重要。

So starting with ZF Optics, owning the entire bill of materials basically and being responsible for sourcing that and being able to lean in and basically get capacity commitments based on us leaning forward, leaning into making an investment in each area. We feel confident that we're going to achieve a very aggressive ramp in the second half of our fiscal year and even more than double or triple that the following year. So we feel very good about the commitments that we've got in place across the board from the supply chain that contributes to ZF Optics.

因此，先從 ZF Optics 談起，我們基本上擁有完整的物料清單（BOM），並負責其採購；同時也能在我們更積極投入、在各個環節進行投資的前提下，向供應端爭取產能承諾。我們有信心在本財年下半年實現非常積極的放量爬坡，並且在下一年度甚至能在此基礎上再翻倍或三倍。因此，對於供應鏈各環節就 ZF Optics 所提供的承諾，我們整體感到非常滿意。

And of course, that's beyond the silicon that we're building. Now from a silicon side, we have a very diverse set of products from a perspective of process geometry. Right now, 12-nanometer is a workhorse. A lot of the ADCs that we're shipping are using 12-nanometer for all the 100-gig per lane. We also are transitioning for optical DSPs into 7-nanometer for 100 gig per lane.

當然，這還不包含我們正在打造的矽晶片部分。就矽晶片而言，從製程幾何的角度來看，我們有非常多元的產品組合。目前 12 奈米是主力製程。我們出貨的許多 ADC 都採用 12 奈米，用於每通道 100G。我們也正將光學 DSP 轉向 7 奈米，以支援每通道 100G。

And we've got a program in flight in 5-nanometer that will be significant volume. And then for all of our 200-gig per lane products across the portfolio, we're using 3-nanometer. And I can say that's probably pretty consistent throughout the industry that if we talk about the 1.6T market that I don't think we're going to see anybody in 5-nanometer. We'll see the bulk of the production happen in three and possibly a shift to two over time to respond to the need for lower power. For us, we're going to deliver very low power in 3-nanometer, as you would expect.

我們也有一個正在進行中的 5 奈米專案，未來將有可觀的出貨量。而在我們產品組合中所有每通道 200G 的產品，則採用 3 奈米。我可以說，這在整個產業大致上也相當一致：如果談到 1.6T 市場，我不認為會看到有人使用 5 奈米。大部分量產會在 3 奈米進行，並且可能隨時間推移逐步轉向 2 奈米，以回應更低功耗的需求。對我們而言，我們將在 3 奈米上交付非常低的功耗，這也符合大家的預期。

Now there's been a lot of discussion about 3-nanometer overall capacity and supply chain issues. We've got an extremely close relationship with our supply chain partners. And I think there's a strong understanding that clusters don't get built without the small complementary connectivity chips that we build and that we embed into our system-level products as well. So I feel quite good. I mentioned on the last call that I feel good about our position, and it's really based on the fact that this is something that over the last five years, we've invested heavily and not just in resources, but also dollars in making sure our supply chain partners understand our commitment.

目前市場對於 3 奈米整體產能與供應鏈問題有很多討論。我們與供應鏈夥伴維持極為緊密的合作關係。我認為大家也很清楚：若沒有我們所打造、並嵌入系統級產品中的那些小型互補性連接晶片，叢集（clusters）是無法建置的。因此我感覺相當踏實。我在上一次電話會議提到我對我們的供應位置有信心，而這主要是基於過去五年我們投入了大量資源，不僅是人力，也投入了資金，確保供應鏈夥伴理解我們的承諾。

Quinn Bolton, Needham.

Quinn Bolton，Needham。

I just wanted to ask maybe cut the revenue a different way. Do you guys expect any meaningful revenue in fiscal '27 from scale up? Or do you expect the vast, vast majority still to be scale out? And if scale up is beyond '27, when would you think scale up becomes potentially a meaningful contributor to revenue?

我想換個角度切分營收來問。你們是否預期在 2027 財年會有任何具實質意義的 scale-up 營收？還是你們預期絕大多數仍將來自 scale-out？如果 scale-up 要到 2027 年之後才會發生，那你們認為 scale-up 何時可能成為營收的有意義貢獻者？

I would say fiscal '27, there is scale up revenue, but it is going to represent the beginning of the ramp. I think we see fiscal '28 as being more substantial. But it really boils down to the different architectures that are being considered. It boils down to each customer and their own individual strategy around how do they solve that bandwidth opportunity or that bandwidth challenge for scale-up. And so we very much view that all of the hyperscalers and Neo Clouds for that really represent their own individual markets. And so we're very, very focused customer by customer, and our portfolio is going to be deployed differently for each one.

我會說在 2027 財年會有 scale-up 營收，但那將代表放量爬坡的起點。我們認為 2028 財年會更為顯著。但這其實取決於正在被考量的不同架構。也取決於每位客戶各自的策略：他們要如何解決 scale-up 的頻寬機會或頻寬挑戰。因此我們非常清楚地看待：所有 hyperscalers 與 Neo Clouds 對我們而言，基本上各自代表獨立的市場。所以我們會以客戶為單位高度聚焦，而我們的產品組合也會因不同客戶而以不同方式部署。

Karl Ackerman, BNP Paribas.

Karl Ackerman，法國巴黎銀行（BNP Paribas）。

Bill, I was hoping you could think about the demand opportunity for CPU-based AI servers for inferencing tasks. Should we assume a similar NIC-to-ToR attach rate for custom ASICs and Agentic CPU servers relative to the GPU-based servers you address today?

Bill，我希望你能談談以 CPU 為基礎的 AI 伺服器在推論（inferencing）任務上的需求機會。相較於你們目前所服務的 GPU 伺服器，我們是否應假設：客製化 ASIC 與 Agentic CPU 伺服器的 NIC-to-ToR（網卡到機架頂端交換器）搭載率（attach rate）會是類似的？

I think from an expectation standpoint, when we talk about training and we talked about inference, and now we're talking about Agentic, all of them are deployed in different ways, and all of them represents a great connectivity opportunity for us. I think from the standpoint of Agentic, I think is really a great opportunity from the standpoint of more front-end connections. And depending on architectures, it could go beyond that as well.

我認為從預期的角度來看，當我們談訓練（training）、談推論（inference），現在又談 Agentic，它們的部署方式各不相同，但對我們而言都代表很大的連接（connectivity）機會。就 Agentic 而言，我認為它在更多前端連接的需求上是一個非常好的機會。而且依不同架構而定，機會也可能不僅止於此。

Sebastien Naji, William Blair.

Sebastien Naji，William Blair。

You guys have announced a number of partnerships with smaller Neo clouds. It seems like Credo is diversifying even beyond your core five hyperscaler customers. So could you maybe help frame for investors how big an opportunity these Tier 2 cloud customers could be for Credo? Could they represent 10% or even 20% of total revenue over the next few years? Or is that too aggressive just given how strong hyperscaler demand is? Any color there would be great.

你們宣布了多項與較小型 Neo clouds 的合作夥伴關係。看起來 Credo 正在多元化，甚至超越你們核心的五大 hyperscaler 客戶。因此你能否協助投資人理解：這些 Tier 2 雲端客戶對 Credo 可能有多大的機會？未來幾年它們是否可能占總營收的 10% 甚至 20%？或者考量 hyperscaler 需求如此強勁，這樣的假設是否過於激進？若能提供一些看法會很有幫助。

Absolutely. This is one of the biggest trends that we're encouraged by this emergence of the Neo cloud ecosystem. So historically, as you mentioned, a lot of the AI infrastructure investment was concentrated within five and now six hyperscalers. We're seeing a great growth in the number of Neo cloud providers that are building AI infrastructure platforms to support a broad range of applications from model developers to enterprises to sovereign AI and inference workloads, even Agentic. These operators are maybe the perfect customer for Credo in a sense that they move.

當然可以。我們非常受到鼓舞的一個重大趨勢，就是 Neo cloud 生態系的崛起。如你所提，過去許多 AI 基礎設施投資集中在五家、現在是六家 hyperscalers。我們看到越來越多 Neo cloud 供應商正在建置 AI 基礎設施平台，以支援從模型開發者、企業、主權 AI（sovereign AI）到推論工作負載，甚至 Agentic 等廣泛應用。這些營運商在某種意義上可能是 Credo 的完美客戶，因為他們行動非常快。

They move very quickly. They are deploying very optimized architectures. And they put a huge emphasis on network performance, reliability and also time to deployment. All of these things make them financially more successful if they could do a good job of it. So we think that the Neo Clouds are going to represent a growing and more meaningful opportunity for Credo over the coming years. And I do think you're right that we can -- if we look at that group of customers collectively, I definitely think it could be on the order of 20%.

他們推進速度非常快。他們部署高度最佳化的架構。而且他們非常重視網路效能、可靠性以及部署時程（time to deployment）。如果他們能把這些事情做好，財務表現就會更成功。因此我們認為，Neo Clouds 在未來幾年將為 Credo 帶來持續成長、且更具意義的機會。我也確實認同你的看法：如果把這一群客戶合在一起看，我絕對認為其占比可能達到約 20% 的量級。

Suji Desilva, ROTH Capital.

Suji Desilva，ROTH Capital。

Congrats on the strong results here. Trying to get some understanding of the relative growth maybe of the copper-based products versus optical. Copper is still a large part of the mix, but optical is coming out strong here. The next few years, just wanted to ask maybe, would you expect that possibly optical with all the tailwinds can cross over into the kind of half of the revenues to half half? Or is that not the right expectation for a multiyear growth path?

恭喜你們交出強勁的成績。我想了解一下以銅互連產品相對於光學產品的成長幅度。銅仍占很大比重，但光學這邊看起來表現很強。展望未來幾年，我想問：你們是否預期在各種利多推動下，光學營收可能上升到接近一半、與銅各占一半（50/50）？或者對於多年期的成長路徑而言，這不是正確的預期？

I think that eventually, we could see optical and copper products in our portfolio to eventually, we could get to 50-50 for sure. I think if you look at the market size in general, and we're doing extremely well in AECs. And there's as I mentioned earlier, I think there's long-term growth case to be made. We feel very strongly about where the market is going and our position in the market and the products that we're bringing forward, especially as the market transitions from 800-gig to 1.6T. So it's going to be a lot of great dynamics within the AEC market.

我認為最終我們的產品組合中，光學與銅產品確實有可能走到 50/50。如果看整體市場規模，我們在 AEC（主動式電纜）表現非常出色。而且如我先前提到，我認為可以建立一個長期成長的論述。我們對市場走向、我們在市場中的定位，以及我們正在推出的產品都非常有信心，特別是在市場從 800G 轉向 1.6T 的過程中。因此 AEC 市場將會出現許多很好的動態。

But look, if we look at the overall market size for optical, just with pluggables alone, and we can have that pluggable versus CPO or NPO long-term discussion. But if we look at the long-term market for pluggables, it's significantly larger than the AEC pluggable market. So yes, I think we can make a case where we will achieve 50-50. And I can make the case just based on market size and penetration with the market that optical could actually achieve copper in the upcoming years.

但你看，如果我們只看光學的整體市場規模、僅就可插拔（pluggables）而言，我們當然也可以做可插拔對 CPO 或 NPO 的長期討論。但如果我們看可插拔的長期市場，它明顯大於 AEC 可插拔市場。所以是的，我認為我們可以提出一個論點，認為我們將達到 50-50。而且我也可以僅基於市場規模與滲透率來論證，在未來幾年光學實際上可能追上銅互連。

Blayne Curtis, Jefferies.

Blayne Curtis，Jefferies。

I want to ask about the four 10% customers. I don't know if you're willing to give the numbers, but maybe just speak to some of the lumpiness that you've seen in the past and whether that will impact the first half of this year. And then if you can also address just those customers moving to the next product and kind of the growth trajectory you said three or four would contribute. Just relative confidence on that would be great.

我想問一下那四個占比 10% 的客戶。我不確定你是否願意提供具體數字，但也許可以談談你們過去看到的一些訂單/出貨波動（lumpiness），以及這是否會影響今年上半年。另外，如果你也能說明一下這些客戶轉向下一代產品的情況，以及你提到其中三到四家會帶來貢獻的成長軌跡。如果能談談你對此的相對信心程度就更好了。

Yes. So as we mentioned in our prepared remarks, we had four 10% customers in Q4. They range from a high of 34% down to 10%. The three largest customers in Q4 were the same customers that were the largest in Q3 as well. So it's 34% for the largest, 27% for the second largest, 16% for the third.

是的。如同我們在事先準備的發言中提到的，我們在第四季有四個占比 10% 的客戶。占比從最高的 34% 到 10% 不等。第四季的前三大客戶與第三季的前三大客戶相同。也就是最大客戶 34%、第二大 27%、第三大 16%。

And notably, the fourth customer, which was at 10%, they were not previously a 10% customer during fiscal '26. As you mentioned, Blayne, there is some quarter-over-quarter variability with any given customers. This quarter, I would say it wasn't huge variability that was driven, but there could be some of that going forward. We have seen that certainly in the past. But one of the key takeaways is we expect increasing diversification in fiscal '27, and it's really based on our current customer engagements.

值得注意的是，第四大客戶占比為 10%，在 26 會計年度之前他們並不是占比 10% 的客戶。如你所說，Blayne，任何單一客戶都可能出現季對季的變動。就本季而言，我會說並沒有由此驅動很大的變動，但未來可能會出現一些這樣的情況。我們過去確實也看過這種情形。但其中一個關鍵重點是，我們預期在 27 會計年度客戶結構會更加多元化，而這主要是基於我們目前與客戶的合作進展。

And we continue to diversify with multiple large hyperscale customers. And as Bill just described, with Neo clouds, we expect them to become an important and sizable contributor in revenue in the upcoming quarters and years.

而且我們也持續與多家大型超大規模（hyperscale）客戶推進多元化。正如 Bill 剛才所描述的，對於 Neo clouds，我們預期它們在接下來幾季與未來幾年將成為營收上重要且可觀的貢獻來源。

Vijay Rakesh, Mizuho.

Vijay Rakesh，Mizuho。

Good to see a strong continued ramp here. Just a question back on the 1.6T. Obviously, that should drive a good content increase, ASP increase for you with the 200-gig per lane. When do you see that ramping? Is that more like into next year? Can you give some color on that? And on the ZF Optics side, do you see that also gaining traction on the 1.6T side? And if you can give us some color on how many customers in ZF-Optics now?

很高興看到這裡持續強勁的放量成長。我想再回到 1.6T 的問題。顯然，隨著每通道 200G，你們的內容價值（content）與 ASP 都應該會有不錯的提升。你們什麼時候看到它開始放量？會比較像是明年嗎？能否提供一些說明？另外在 ZF Optics 方面，你們是否也看到在 1.6T 上獲得更多採用？以及能否分享一下目前 ZF-Optics 有多少客戶？

So first, on the transition from 800-gig to 1.6T. Of course, the timing is going to shift a lot about the deployment of Rubin and really each of the customers' individual strategy. Some will be very delayed. Some will be first to deploy. But I think the exact timing of the transition will move somewhat as the platforms evolve.

先談從 800G 轉向 1.6T。當然，時程會很大程度取決於 Rubin 的部署進度，以及各家客戶各自的策略。有些會延後很多。有些會率先部署。但我認為，隨著平台演進，這個轉換的確切時點會有所變動。

But the underlying bandwidth requirements, I don't think change. It's important to recognize that customers are designing for flexibility. In many cases, system architectures that support 1.T bandwidth when they come to market, typically, we'll see customers building flexibility on the rest of the ecosystem of connectivity. So we see a lot of our customers designing two physical ports for each GPU. And in the case of 1.6T, one option would be to deploy in each port, four lanes of 200-gig signaling.

但底層的頻寬需求，我認為不會改變。重要的是要認知到，客戶在設計時會追求彈性。在許多情況下，當支援 1.T 頻寬的系統架構上市時，我們通常會看到客戶在其餘連接生態系上也建立彈性。因此我們看到許多客戶為每顆 GPU 設計兩個實體埠。而在 1.6T 的情境下，其中一個選項是在每個埠部署 4 條 200G 訊號通道（lanes）。

Another risk-averse way of going to market would be to deploy two ports with eight lanes of 100-gig. And so for us, that allows customers to manage deployment timing while preserving future upgrade paths. So not everything has to come at once. There's a huge trade-off on overall supply chain as well as maturity as it relates to different connectivity products. For our perspective, we participate in both.

另一種較為保守、降低風險的上市方式，是部署兩個埠、每個埠 8 條 100G 通道。對我們而言，這讓客戶能在保留未來升級路徑的同時，管理部署時程。所以不需要所有事情一次到位。在整體供應鏈以及不同連接產品的成熟度方面，也存在很大的取捨。就我們的角度，我們兩邊都會參與。

We have strong positions in 100-gig per lane deployments today, and we can continue supplying that at very high volume. We also invested heavily in 200-gig per lane. So we're ready to go from the standpoint of AECs and ZF Optics. Now as it relates to optical DSPs, yes, we've got several customer engagements. And from a cyclo PIC perspective, same thing.

我們目前在每通道 100G 的部署上有很強的地位，並且可以持續以非常高的出貨量供應。我們也在每通道 200G 上投入了大量資源。因此就 AEC 與 ZF Optics 而言，我們已經準備就緒。至於光學 DSP，是的，我們已經有數個客戶合作案在進行中。從 cyclo PIC 的角度也是一樣。

But we don't ultimately control the outcome of when those transceivers go to market. And so it's hard to talk about the precise timing of the transition, but we're bullish on the long-term opportunity, whether it's 800-gig or 1.6T. The dynamics as we go to 1.6T, as we've talked about, is we'll -- as the rest of the industry will experience an uplift in ASPs. So hopefully, that answers your question. If I could just summarize at a high level, I think our fiscal '27 will have relatively light revenue as it relates to 200-gig per lane, and that is just a function of the fact that the industry hasn't seem to get there quite yet, and there's always a rumor about delays.

但我們最終無法控制那些收發器何時上市。因此很難談論轉換的精確時點，但我們對長期機會持樂觀態度，無論是 800G 或 1.6T。正如我們談過的，走向 1.6T 的動態是，我們——以及產業其他公司——都會經歷 ASP 的上升。希望這回答了你的問題。如果我用高層次做個總結，我認為我們 27 會計年度與每通道 200G 相關的營收會相對較少，這只是因為產業目前似乎還沒真正走到那一步，而且市場上總是有延遲的傳聞。

If we are going to make a shift and the industry is ready, we're ready right now. We're absolutely confirmed from the standpoint of being ready for production on all of our copper as well as optical system-level products as well as component level products.

如果我們要進行轉換、而且產業已準備就緒，我們現在就已經準備好了。從量產準備度來看，我們在所有銅互連、光學系統層級產品以及元件層級產品方面，都已完全確認具備量產能力。

Janco Venter, Arete Research.

Janco Venter，Arete Research。

Look, you're very well capitalized right now, as you presented in the prepared remarks, you've undergone strategic investments with Dust Photonics. You've added SiPho to the portfolio. being integrated into your Optics portfolio as well as being a stand-alone opportunity, which you're putting to work in the coming fiscal year. So can you perhaps help us understand how you're thinking about putting the capital on the balance sheet to work and how you will be approaching further strategic investments?

你們目前資本相當充裕，正如你們在事先準備的發言中所呈現的；你們也透過 Dust Photonics 進行了策略性投資。你們把 SiPho 納入產品組合，既整合進你們的光學（Optics）產品線，同時也作為一個獨立的機會，並且你們將在下一個會計年度加以運用。所以你們能否協助我們理解，你們如何思考運用資產負債表上的資本，以及你們將如何進一步進行策略性投資？

Yes. Let me first state that we have no current plans to raise any additional capital at this point, nor do we plan to do a share buyback, goes without saying. One of our goals that we've laid out in the past, particularly when we did an ATM a couple of quarters ago, is that we want to maximize our strategic flexibility for things such as now we've done three acquisitions. And just to note that the Dust Photonics acquisition closed last week, as I'm sure you saw, the net cash amount that you'll see in Q1 that went out the door in Q1 was about $750 million. And of course, we ended Q4 with $1.4 billion in cash.

是的。我先說明，目前我們沒有任何計畫要再募集額外資本，也不打算進行庫藏股回購，這不言而喻。我們過去提出的一個目標，特別是在幾季前我們做 ATM（隨市發行）時，就是希望最大化我們的策略彈性，例如我們現在已經完成了三項收購。另外補充一下，Dust Photonics 的收購已在上週完成，我相信你也看到了；你們在第一季會看到的、第一季流出的淨現金金額約為 7.5 億美元。當然，我們在第四季結束時持有 14 億美元現金。

But that's against the backdrop of being in a very strong cash flow position where cash flow from operations is approaching $200 million per quarter now. So as we exit Q1, we're very comfortable in our cash position. And we expect there will be times where there will be opportunistic acquisitions that we may entertain from this point forward, but nothing immediately that we're looking at.

但這是在我們現金流狀況非常強勁的背景之下，目前營運活動產生的現金流已接近每季 2 億美元。因此在我們結束第一季之際，我們對自身的現金部位非常有信心。我們預期從現在起，未來有時會出現一些機會型收購，我們可能會考慮，但目前沒有任何我們正在立即評估的項目。

(Operator Instructions)

（接線員指示）

Tore Svanberg, Stifel.

Tore Svanberg，Stifel。

Just a follow-up, Bill. So if you think about the three optical businesses in fiscal '27, I mean, two of them, the PIC and the DSP, they've been sort of -- I mean, they've been ramping, but ramping over time, whereas ZF Optics obviously, is more of a hockey stick ramp, at least that's the way I would position it. When you think about Weaver and ALC starting to ramp in fiscal '28, how should we think about those ramps in relation to the other 3? Just so we sort of get a feel for the trajectory of both those new product categories in fiscal '28?

Bill，我再追問一下。如果你把 2027 會計年度的三項光學業務放在一起看，我是說，其中兩項——PIC 和 DSP——它們一直是那種……我是說，它們在爬坡，但屬於隨時間逐步爬坡；而 ZF Optics 顯然更像是曲棍球棒式的爬坡，至少我是這樣定位的。當你想到 Weaver 和 ALC 在 2028 會計年度開始爬坡時，我們應該如何看待它們的爬坡相對於另外三項的關係？只是想讓我們大致感受一下，這兩個新產品類別在 2028 會計年度的成長軌跡？

Sure. I will say that to give a little more color on your first question earlier, the ASPs on the discrete components, optical DSPs and cyclo PICs, those ASPs are typically two-digit ASPs. On the ZF Optics, we're going to see three-digit ASPs. And so I didn't mean to be a bit elusive or not answer your question. But I think the math clearly says that as we ramp ZF Optics, that's going to be clearly our largest revenue contributor for our optical portfolio that the potential there is great. And what we're delivering is a solution to network reliability, which is something that is absolutely higher and higher priority for the customers that we work with. Now of course, I forgot the rest of the question.

當然。我先補充一點，回到你先前第一個問題，離散元件、光學 DSP 與 cyclo PIC 的 ASP（平均售價）通常是兩位數的 ASP。而在 ZF Optics 上，我們會看到三位數的 ASP。所以我並不是想要閃躲或沒有回答你的問題。但我認為從數學上很清楚：隨著我們拉升 ZF Optics 的出貨，它將明顯成為我們光學產品組合中最大的營收貢獻者，潛力非常大。而我們交付的是一個提升網路可靠性的解決方案，這對我們合作的客戶而言，正變得愈來愈優先。當然，我忘了你後面問的問題是什麼。

Yes, Omni and ALC and how should we think about those?

是的，Omni 和 ALC，我們應該如何看待它們？

I got too focused on the near term. Sorry about that. Let me put things in perspective for ALCs. So that is a system-level pluggable product, just like an AEC or just like a ZF Optic. What it does is it brings the same AEC dynamics of power efficiency, cost efficiency.

我剛剛太專注在短期了。抱歉。我先把 ALC 的情況放到脈絡裡。它是一個系統級的可插拔產品，就像 AEC 或 ZF Optic 一樣。它帶來與 AEC 相同的動態：電力效率、成本效率。

It's got a better form factor than AECs, and it's got a much longer length. And so for certain customers, that's going to be a really natural product to get from a design in and beginning of production. So the dynamic there could be very much similar to AECs and ZF Optics in the sense that we can see large revenue quickly. Now as it relates to Weaver, just let's use our first customer, Positron, as an example. They are redefining memory attachment and bandwidth for inference.

它的外形規格比 AEC 更好，而且可達更長的長度。因此對某些客戶而言，這會是一個非常自然、容易從設計導入並開始量產的產品。所以其動能可能會非常類似 AEC 與 ZF Optics，也就是我們能在短時間內看到可觀的營收。至於 Weaver，我們以第一位客戶 Positron 為例。他們正在重新定義推論的記憶體連接與頻寬。

So they announced their first product with a total memory LPDDR size of two terabytes. So that's more than 10 times any other inference engine that's been announced in the market. So game changing from the standpoint of performance. Now if two terabytes are deployed, that requires many, many Weaver chips. And what I've said in the past is that the revenue contribution per GPU can be between $2,000 and $3,000. And so you can see that, that product category will ramp very quickly as well as we've got customers that go into volume with their inference GPUs that utilize that solution.

他們已宣布第一款產品，其總記憶體 LPDDR 容量達到 2TB。這比市場上已宣布的任何其他推論引擎都高出 10 倍以上。因此從效能角度來看是顛覆性的。如果部署 2TB，就需要非常非常多的 Weaver 晶片。我過去也提過，每顆 GPU 的營收貢獻可能介於 2,000 到 3,000 美元。因此你可以看到，隨著客戶採用該解決方案的推論 GPU 進入量產，這個產品類別也會非常快速地爬坡。

Jim Schneider, Goldman Sachs.

Jim Schneider，Goldman Sachs。

Given the accelerating growth you expect sequentially in the back half of this year, presumably driven by the optical portfolio, can you talk about any kind of changes you're expecting to see in terms of how your customers deploy AECs in environments? Or is that growth just going to remain very strong off of an elevated base?

鑑於你們預期今年下半年營收將出現逐季加速成長，推測主要由光學產品組合帶動，你能否談談你們預期客戶在環境中部署 AEC 的方式會有哪些變化？或者這個成長只是會在已經偏高的基期之上，仍然維持非常強勁？

From the standpoint of say, AEC customer diversity or growth, we're definitely seeing that AEC adoption is broadening across the industry. And we very much feel like we're in the early innings or stages of penetration. So we're seeing a broad breadth of adoption. We've mentioned Neo clouds as a new customer category, really a perfect solution for those players that have architectures that will enable in-rack and multi-rack deployments, there is a trend towards more density. So there is definitely a trend towards more Neo clouds being able to use AECs.

就例如 AEC 客戶多元性或成長而言，我們確實看到 AEC 的採用正在整個產業中擴大。我們也非常認為目前仍處於滲透的早期階段。因此我們看到採用的廣度正在擴張。我們提到「Neo clouds」作為新的客戶類別，對於那些架構能支援機櫃內（in-rack）與跨多機櫃（multi-rack）部署的業者而言，這真的是非常合適的解決方案，而且市場正朝更高密度的方向發展。因此確實存在更多 Neo clouds 能夠使用 AEC 的趨勢。

But I think that as we talked about that category growing, I think we can definitely say Neo clouds will be growing for the foreseeable future. Also, if we look at the hyperscalers, of course, we've mentioned we're deployed and we're high volume with five of six now. And I think that with all of them, we've still got the ability to penetrate further into their networks. I don't think we can point to one where we're fully deployed other than xAI. And so we see great growth opportunity for both hyperscalers and Neo clouds.

但我認為，當我們談到該類別的成長時，我們可以很明確地說，Neo clouds 在可預見的未來都會持續成長。另外，如果我們看超大規模雲端業者（hyperscalers），當然我們也提過，我們現在已在其中六家中的五家完成部署並處於高出貨量狀態。而我認為對所有這些客戶而言，我們仍有能力進一步滲透他們的網路。除了 xAI 之外，我不認為我們能指出任何一家已經完全部署到位。因此我們看到在 hyperscalers 與 Neo clouds 兩者上都有很大的成長機會。

And I think it's really going to be a long-term growth that we expect. I don't think it's going to be as fast growth from a percentage standpoint as what we're going to see in optical, but it's going to be a continued growth driver for the company.

而且我認為這將會是我們所預期的長期成長。我不認為從百分比角度來看，它會像我們在光學業務看到的那麼快，但它會持續成為公司的成長驅動力。

There are no further questions at this time. Mr. Brennan, I turn the call back over to you.

目前沒有其他問題。Brennan 先生，我把電話交回給您。

Thanks so much for the questions. I really appreciate the ongoing interest and support and look forward to the next time we talk.

非常感謝各位的提問。我非常感謝大家持續的關注與支持，也期待下次再與各位交流。

This concludes today's conference call. You may now disconnect.

今天的電話會議到此結束。您現在可以掛線。