Arista Networks, Inc. (ANET) 2026 Q3 法說會逐字稿

Welcome to the third-quarter 2026 Arista Networks financial results earnings conference call. (Operator Instructions) As a reminder, this conference is being recorded and will be available for replay from the Investor Relations section on the Arista website following this call.

歡迎參加 Arista Networks 2026 年第三季財務業績電話法說會。(接線員指示) 提醒各位，本次會議將被錄音，並將於會後在 Arista 官網「投資人關係」專區提供重播。

Mr. Rudolph Araujo, Arista's Head of Investor Advocacy, you may begin.

Arista 投資人倡議主管 Rudolph Araujo 先生，您可以開始。

Good afternoon, everyone, and thank you for joining us. With me on today's call are Jayshree Ullal, Arista Networks' Chairperson and Chief Executive Officer; and Chantelle Breithaupt, Arista Networks' Chief Financial Officer.

各位下午好，感謝各位加入我們。今天與我一同出席的有 Arista Networks 董事長兼執行長 Jayshree Ullal，以及 Arista Networks 財務長 Chantelle Breithaupt。

This afternoon, Arista Networks issued a press release announcing its fiscal third-quarter results for the period ending September 30, 2026. If you want a copy of this release, you can find it on our website.

今天下午，Arista Networks 已發布新聞稿，公布截至 2026 年 9 月 30 日止期間的第三季財報結果。如需取得該新聞稿副本，請至我們的網站查閱。

During the course of this conference call, Arista Networks management will make forward-looking statements, including those relating to our financial outlook for the fourth quarter of the 2026 fiscal year, longer-term business model and financial outlook for 2027 and beyond, our total addressable market and strategy for accessing those market opportunities across AI, campus, data center and other segments, inventory levels and management lead times, purchase commitments, component supply and product innovation, which are subject to the risks and uncertainties that we discuss in detail in our documents filed with the SEC specifically in our most recent Form 10-Q and Form 10-K and which could cause actual results to differ materially from those anticipated by these statements.

在本次電話會議進行期間，Arista Networks 管理團隊將作出前瞻性陳述，包括與我們 2026 會計年度第四季的財務展望、2027 年及以後較長期的商業模式與財務展望、我們的總可服務市場（TAM）以及在 AI、校園網路、資料中心與其他區隔中切入市場機會的策略、庫存水位與管理交期、採購承諾、零組件供應與產品創新等相關內容；上述事項均受我們在提交給美國證券交易委員會（SEC）的文件（特別是最近一期 Form 10-Q 與 Form 10-K）中所詳述之風險與不確定性影響，且可能導致實際結果與這些陳述所預期者出現重大差異。

These forward-looking statements apply as of today, and you should not rely on them as representing our views in the future. We undertake no obligation to update these statements after this call.

這些前瞻性陳述僅適用於今日，您不應依賴其作為我們未來觀點的代表。本公司不承擔於本次電話會議後更新這些陳述的任何義務。

This analysis of our Q3 results and our guidance for Q4 2026 and FY 2027 is based on non-GAAP financial measures and exclude stock-based compensation expense, intangible asset amortization and other nonrecurring items as well as the income tax effect of these non-GAAP exclusions, including the recognition of direct excess tax benefits associated with stock-based awards.

我們對第三季業績的分析，以及對 2026 年第四季與 2027 會計年度的指引，係以非 GAAP 財務衡量指標為基礎，並排除以股份為基礎之薪酬費用、無形資產攤銷及其他非經常性項目，以及上述非 GAAP 排除項目的所得稅影響（包括與股份獎酬相關之直接超額稅務利益的認列）。

A full reconciliation of our selected GAAP to non-GAAP historical results is provided in our earnings release. A reconciliation of our FY27 non-GAAP guidance to the most directly comparable GAAP measures is not available without unreasonable effort.

我們在 GAAP 與非 GAAP 之間的部分歷史結果對帳，已於財報新聞稿中提供完整調節表。在不付出不合理努力的情況下，無法提供 2027 會計年度非 GAAP 指引與最直接可比 GAAP 指標之對帳。

With that, I will turn the call over to Jayshree.

接下來，我將把電話交給 Jayshree。

Thank you, Rudy, and welcome, everyone, to our third-quarter 2026 earnings call.

謝謝你，Rudy，也歡迎各位參加我們 2026 年第三季法說會。

Arista just delivered a standout quarter of $3.66 billion in revenue, which is translating to our annual growth of 58.5%. This marks our largest ever year-over-year revenue growth in the history of being a public company.

Arista 剛交出一個表現突出的季度，營收達 36.6 億美元，折算年增率為 58.5%。這是我們成為上市公司以來，史上最大的年對年營收成長。

Before we dive into the financial specifics, I would like to emphasize our rapid Arista 2.0 success; a strategy we embarked upon just three years ago. It has generated meaningful customer outcomes to become the central network nervous system for us.

在深入財務細節之前，我想強調我們快速推進的 Arista 2.0 成功；這是我們三年前才啟動的策略。它帶來了具體且有意義的客戶成果，並使我們成為核心的網路「神經系統」。

In the 2000 era, the industry had accepted a status quo, plagued by fragile, overly complex, disparate and vulnerable operating systems. Arista pioneered the universal cloud network and the AI Spine architecture based upon our rock-solid high-quality foundation of our differentiated EOS, that one network operating system with a published subscribed state and data-driven foundation has governed us with five As: Agile security, Automation, Availability, Analytics and AI principles.

在 2000 年代，產業曾接受一種現狀：作業系統脆弱、過度複雜、彼此割裂且易受攻擊。Arista 率先提出通用雲端網路（universal cloud network）與 AI Spine 架構，建立在我們差異化 EOS 的堅實高品質基礎之上；這套單一網路作業系統具備已發布的訂閱狀態與資料驅動的基礎，並以「五個 A」原則引領我們：敏捷安全（Agile security）、自動化（Automation）、可用性（Availability）、分析（Analytics）與 AI 原則。

By extending our single EOS operating system from the data center, all the way from the Universal Spine to many forms of edge lead functions, we deliver structural and secure network economy.

透過將單一 EOS 作業系統從資料中心延伸，從 Universal Spine 一路擴展到多種邊緣引領功能，我們提供具結構性且安全的網路經濟效益。

Our Arista 2.0 transformation from data centers to centers of data can be characterized as the age of network intelligence, learning from vast repositories of stream data. It has driven our strategy across locations be a client to branch, to campus, the data to cloud and AI centers for our customers.

我們的 Arista 2.0 轉型，從「資料中心」走向「資料的中心」，可被描述為網路智慧的時代——從龐大的串流資料儲存庫中學習。這也推動了我們跨場域的策略：從用戶端到分支、到校園，從資料到雲端，以及為客戶打造 AI 中心。

Based on our network data lake architecture, we offer our customers automated and conversational troubleshooting agentic companions. Devoid of manual hunting through text-heavy logs, our systems proactively begin identifying and isolating issues for network workloads, workflows, and work streams.

基於我們的網路資料湖架構，我們為客戶提供自動化且可對話的故障排除代理型（agentic）夥伴。不再需要在大量文字化日誌中手動搜尋，我們的系統會主動開始識別並隔離網路工作負載、工作流程與作業流中的問題。

In 2026, the Network Centric acceleration is well underway. Our customers have increasingly cordless the back end and the front end of AI and data centers. Arista is long-chain standard-spaced Ethernet. Customers have welcomed our open approach to modernizing cloud campus data centers, WAN centers and now AI centers with a lower total cost of ownership.

在 2026 年，以網路為中心（Network Centric）的加速發展已全面展開。我們的客戶日益將 AI 與資料中心的後端與前端進行無線化（cordless）。Arista 是長鏈、標準間距的乙太網路。客戶歡迎我們以開放方式，現代化雲端校園資料中心、WAN 中心，以及如今的 AI 中心，並降低總持有成本（TCO）。

Recently, we were recognized with the leader status from Forrester for the highest strength of innovation and data center portfolio. We were also recognized by Gartner Magic Quadrant leadership for campus wired and wireless and data center segments.

近期，我們獲 Forrester 評為領導者，並在創新實力與資料中心產品組合方面獲得最高評等。我們也在 Gartner 魔力象限中，於校園有線與無線，以及資料中心等區隔獲得領導者肯定。

As agentic AI models increase in realism, compute diversity and density, we have the choice of many more training and inference accelerators. The industry, we believe, is shifting from compute-centric vertical stack to a holistic network centric approach. A risk is committed to unleashing and breaking down the barriers between compute and networking.

隨著代理型 AI 模型在真實感、運算多樣性與密度上持續提升，我們可選擇的訓練與推論加速器也更多。我們認為，產業正從以運算為中心的垂直堆疊，轉向更整體、以網路為中心的方法。一項風險在於：致力於釋放並打破運算與網路之間的藩籬。

Arista customers can now construct an open network platform capable of handling heterogeneous models and diverse accelerators at scale. Our ethylene portfolio of at least 25 switches for scale-up scale and scale across is optimally designed to improve XPU utilization, job completion, and token latency for these intense training and inference workloads.

Arista 客戶如今可建構一個開放式網路平台，能在規模化情境下處理異質模型與多元加速器。我們的 ethylene 產品組合包含至少 25 款交換器，涵蓋 scale-up 與 scale-out/scale-across，並以最佳化設計提升 XPU 使用率、作業完成時間，以及在高強度訓練與推論工作負載下的 token 延遲。

In October of 2026 at the Open Compute Project conference, Arista showcased our rack scale-up designs partnering closely with several AI accelerators in field with our Arista network diagnostics layer. Based on ESUN, Ethernet scale-up [rack options], these include orthogonal designs, cable intensive back plans and heterogeneous cross-track designs co-designed with many of our partners, including ARM, Microsoft Maya, AMD's Helios, Qualcomm, Meta's MTIA and many more ahead.

在 2026 年 10 月的 Open Compute Project 大會上，Arista 展示了我們的機櫃級 scale-up 設計，並與多家已在現場部署的 AI 加速器緊密合作，同時結合我們的 Arista 網路診斷層。基於 ESUN、Ethernet scale-up [rack options]，其中包括正交設計、纜線密集的背板，以及異質 cross-track 設計；這些皆與多家合作夥伴共同設計，包括 ARM、Microsoft Maya、AMD 的 Helios、Qualcomm、Meta 的 MTIA，以及更多即將加入的夥伴。

We expect scale-up trials to begin in 2027, driving Ethernet to larger ratings of 126 to 144 within racks and across 1,000 across racks with production expected for scale-up in 2028.

我們預期 scale-up 試驗將於 2027 年開始，推動乙太網路在機櫃內提升至 126 到 144 的更大規模，並在跨機櫃達到 1,000 的規模；預計 2028 年開始進入 scale-up 量產。

We are witnessing the migration from highly centralized training compute to disaggregated and distributed inference prefill and decode demanding much more intelligent networking and co-location of compute, storage and KB cash resources rather than simply focusing on raw feeds and speeds and flops of compute.

我們正見證從高度集中式的訓練運算，遷移至解耦且分散式的推論 prefill 與 decode；這對更智慧的網路，以及運算、儲存與 KB cash 資源的共置提出更高要求，而不再只是聚焦於原始的吞吐、速度與運算 FLOPS。

A well-designed network is becoming a foundational cornerstone for optimal token deficiency. Arista's decades of EOS engineering in both multiplanar and distributed model routing at scale, be it ARP, address resolution protocols, or fifth tables forward information base or access list with rapid reconversions comes to the rescue.

一個設計良好的網路正成為達成最佳 token 缺失（deficiency）的基礎基石。Arista 數十年在 EOS 的工程投入，涵蓋多平面（multiplanar）與分散式模型的規模化路由——無論是 ARP（位址解析協定）、或是第五表（fifth tables）的轉送資訊庫（FIB）、或是存取清單（ACL）在快速重新收斂（reconvergence）方面——都能在此發揮關鍵作用。

Our ethylene products are inherently designed to actively understand parameters at a deep level. including latency and bandwidth profiles of multi-tenancy to intelligently scale workloads for optimal placement between local and remote domains. The need for visibility into ports packages, flows, buffer utilization and part latency across the entire scale out and scale across fabric validates real AI system level performance to avoid idling of precious compute cycles.

我們的 ethylene 產品在本質上即被設計為能在深層次主動理解各項參數，包括多租戶的延遲與頻寬輪廓，以智慧方式擴展工作負載，並在本地與遠端網域之間進行最佳化配置。對整個 scale-out 與 scale-across fabric 的連接埠封包、流量、緩衝區使用率與部分延遲的可視性需求，驗證了真實的 AI 系統層級效能，以避免寶貴的運算週期閒置。

It assures immediate and network-wide insight to the root cause as well as rapid recovery of the AI system with smart system upgrades across millions of accelerators.

它可確保立即且全網路範圍內洞察根本原因，並透過在數百萬個加速器上進行智慧系統升級，讓 AI 系統得以快速復原。

When we initially began this transformational AI journey, we spoke about four or five AI customers with approximately 50,000 to 100,000 NVIDIA GPUs, and in fact, we're largely staring at a proprietary network cycle such as InfiniBand.

當我們最初開始這段具變革性的 AI 旅程時，我們談到大約四到五家 AI 客戶，擁有約 50,000 到 100,000 顆 NVIDIA GPU；而事實上，我們很大程度上面對的是一個專有的網路週期，例如 InfiniBand。

In 2026, we are in a super cycle of AI having already surpassed 100 ethylene customers. By the end of '27, we are forecasting greater than 10 gigawatts of deployment across 10 or more different types of compute accelerators. What a stark inflection in just three years.

在 2026 年，我們正處於 AI 的超級週期，且已經超過 100 家乙烯客戶。到 2027 年底，我們預測在 10 種或更多不同類型的運算加速器上，部署規模將超過 10 吉瓦。短短三年就出現如此鮮明的拐點。

Central to this AI super cycle is also the expansion of Arista AVA, our autonomous virtual assist to support efficient AI in the enterprise, by providing that open programmable interface within the US and utilizing the model context protocol, MCP, AVA allows emerging AI agents to reason to interact with the network stack and perform advanced multi-domain event correlation across the entire enterprise.

這個 AI 超級週期的核心之一，也包括 Arista AVA 的擴展；AVA 是我們的自主虛擬助理，透過在美國提供開放的可程式化介面並採用模型情境協定（MCP），AVA 讓新興的 AI 代理能夠推理並與網路堆疊互動，並在整個企業範圍內執行進階的跨多領域事件關聯分析，以支援企業高效率導入 AI。

With exponential rise of generated AI attacks and the ever-increasing cyber threats, securing the network is key and integral and our responsibility.

隨著生成式 AI 攻擊呈指數級上升以及網路威脅不斷加劇，確保網路安全是關鍵、不可或缺，也是我們的責任。

I would like to invite Ken Duda, our Founder, President and CTO, to shed some light on this problem. Ken?

我想邀請我們的創辦人、總裁兼 CTO Ken Duda，來為這個問題提供一些見解。Ken？

Thanks, Jayshree. It's hard to overstate the impact of AI on networking.

謝謝你，Jayshree。AI 對網路的影響之大，幾乎難以言喻。

First, a good news. AVA, which is Arista's autonomous virtual assistant, fueled by real-time data AVA is incredibly powerful. It seamlessly navigates networking and security infrastructure from campus to WAN to data center to cloud continuously monitoring and troubleshooting, isolating root causes without manual labor into tasks.

首先，一個好消息。AVA，也就是 Arista 的自主虛擬助理，在即時資料的驅動下，AVA 的能力極其強大。它能無縫穿梭於從園區到 WAN、到資料中心、到雲端的網路與資安基礎設施，持續監控與疑難排解，在不需要人工介入的情況下隔離根本原因並完成相關任務。

Having seen what AVA can do and imagine managing my own infrastructure without it, and I expect many of our customers will feel the same way.

看過 AVA 能做到的事之後，再想像在沒有它的情況下管理我自己的基礎設施，幾乎不可想像；我也預期許多客戶會有同樣的感受。

The bad news, of course, is the massive challenge AI poses for us and our customers by discovering and exploiting security vulnerabilities. This threat landscape demands secure and robust networking.

當然，壞消息是：AI 透過發現並利用安全漏洞，對我們與客戶帶來巨大的挑戰。這樣的威脅態勢要求我們具備安全且強韌的網路。

Over the past few months, we've been collaborating with Anthropic, Google Open AI and many others to integrate new AI-enabled security capabilities from frontier models into our existing software security pipeline. We are being invited early as a key infrastructure provider into partnerships like Project Class win, hold in AI-driven vulnerability discovery and assessment into our established security vulnerability management process.

在過去幾個月裡，我們一直與 Anthropic、Google、OpenAI 以及許多其他夥伴合作，將前沿模型所帶來的新型 AI 賦能安全能力整合進我們既有的軟體安全管線。作為關鍵基礎設施供應商，我們也在早期就受邀參與像 Project Class win 這樣的合作夥伴計畫，將 AI 驅動的漏洞發現與評估納入我們既有的安全漏洞管理流程。

True network security has become a must-have, not a nice to have. And Arista's EOS operating system is the ideal architectural foundation based on four key principles.

真正的網路安全已成為必需品，而非可有可無。而 Arista 的 EOS 作業系統，基於四項關鍵原則，是理想的架構基礎。

First, a single high-quality operating system. Arista has maintained one of the lowest CDE counts less than 100 in the industry over the last two decades, while others are rising exponentially due in part to the simple fact that multiple networking operating systems tend to have more vulnerabilities than just one.

第一，單一且高品質的作業系統。在過去二十年裡，Arista 一直維持業界最低之一的 CDE 數量（少於 100）；而其他廠商的數量則呈指數上升，部分原因很簡單：多套網路作業系統往往比單一作業系統有更多漏洞。

Second, limiting vulnerability impact through process utilization EOS keeps each software agent, management playing work controller planes architecturally separate from each other and also from hardware forwarding. So traffic keeps moving even if there is a defect in one component. EOS state sharing architecture also resists lateral movement. So a defect in one component typically cannot be exploited to attack other agents within the switch.

第二，透過流程利用來限制漏洞影響。EOS 在架構上讓每個軟體代理、管理（播放）工作控制平面彼此分離，並且也與硬體轉送分離。因此，即使某個元件有缺陷，流量仍能持續轉送。EOS 的狀態共享架構也能抵禦橫向移動。因此，一個元件的缺陷通常無法被利用來攻擊交換器內的其他代理。

Third, we provide a seamless upgrade process. Process isolation also means that we can ship a targeted fix to the effective component and install the fix without a maintenance window or a network disruption.

第三，我們提供無縫的升級流程。流程隔離也意味著我們可以針對受影響的元件提供精準修補，並在不需要維護視窗或造成網路中斷的情況下安裝修補。

And fourth, delivering a compliance dashboard to stay current. When a security advisory lens, Arista CloudVision brings real-time visibility to turns a stressful advisory day into a manageable one.

第四，提供合規儀表板以保持最新狀態。當出現安全性公告時，Arista CloudVision 可提供即時可視性，將令人緊張的公告日轉變為可管理的一天。

In addition to providing secure switches, we also provide Zero Trust networking or ZTN technology to help our customers build a security structure. I'd especially like to highlight Macro Segmentation Services, or MSS, is ZTN feature allowing for dynamic security enforcement preventing undesired interactions among users, devices, and applications.

除了提供安全的交換器之外，我們也提供零信任網路（ZTN）技術，協助客戶建立安全架構。我特別想強調巨集分段服務（Macro Segmentation Services，MSS），這是一項 ZTN 功能，可進行動態安全強制，防止使用者、裝置與應用程式之間不希望發生的互動。

Second, we provide Network Detection and Response, or NDR services that profile, analyze and proactively detect sophisticated security threats.

第二，我們提供網路偵測與回應（Network Detection and Response，NDR）服務，用於建立輪廓、分析並主動偵測複雜的安全威脅。

And finally, once workload traffic leads the confines of the local data center, it becomes even more vulnerable to malicious actors. So we offer protection with wire speed ubiquitous link and total infliction native in every port with zero performance (inaudible).

最後，一旦工作負載流量離開本地資料中心的範圍，就更容易受到惡意行為者的攻擊。因此，我們以線速、無所不在的連結與每個埠原生的總（inaudible）提供防護，且零效能損失（聽不清）。

I hope this was a helpful summary at how Arista is helping customers take advantage of AI while improving security to at the new challenges.

希望這能對 Arista 如何協助客戶在善用 AI 的同時提升安全性、以因應新的挑戰，提供一個有幫助的總結。

Back over to you, Jayshree.

交還給你，Jayshree。

Thank you, Ken. It was useful for me, too.

謝謝你，Ken。這對我也很有幫助。

With Arista's uncompromised ZTN approach, we can systematically address the security hazards and parabolic attacks. Our CVE count, a vulnerability count is often 10 to 50 times better than alternatives. Our customer momentum this year has been significant to the enterprise, in the providers across data center campus and AI because they are addressing security proactively.

透過 Arista 毫不妥協的 ZTN 方法，我們可以系統性地處理安全風險與拋物線式的攻擊。我們的 CVE 數量（漏洞數）往往比替代方案好 10 到 50 倍。今年我們在企業與供應商客戶上的動能非常顯著，涵蓋資料中心、園區與 AI，因為他們正在主動因應安全議題。

I would like to invite Todd Nightingale, our President and Chief Operating Officer, to describe more about this. Over to you, Todd.

我想邀請我們的總裁兼營運長 Todd Nightingale，來進一步說明。交給你，Todd。

Thank you, Jayshree.

謝謝你，Jayshree。

We're seeing phenomenal growth across our enterprise business, driven by customers building mission-critical networks, which benefit from Arista's industry-leading quality and reliability. The entire industry response to AI-powered attack exposure, customers are motivated by the fact that keeping their infrastructure current with security updates will protect, segment and defend against vulnerabilities beyond the network.

我們看到企業業務出現驚人的成長，動能來自客戶建置任務關鍵型網路，而這些網路受益於 Arista 業界領先的品質與可靠性。面對整個產業對 AI 驅動攻擊曝險的回應，客戶之所以積極，是因為讓其基礎設施持續更新安全修補，將能在網路之外也提供保護、分段與防禦漏洞的能力。

Arista on a strategic deal at an Asian web services company as they shifted their focus strategically on the speed, scale and resilience in the new AI era. Reliability was the key to the engagement as their teams are frustrated by long-standing quality issues with their existing vendor.

Arista 在一家亞洲網路服務公司拿下一筆策略性交易，因為他們在新的 AI 時代將策略重心轉向速度、規模與韌性。可靠性是這次合作的關鍵，因為他們的團隊長期受既有供應商的品質問題所困擾。

Arista was able to demonstrate reliability and scalability while leveraging standards-based technology and a powerful e-VPN VXLAN architecture on our 100- and 400-gig platforms. This deployment will operate as the front end for their AI deployment moving forward.

Arista 能夠在採用以標準為基礎的技術、並在我們的 100G 與 400G 平台上運用強大的 e-VPN VXLAN 架構的同時，展現可靠性與可擴展性。此部署未來將作為其 AI 部署的前端。

In the health care sector, leading hospitals are selecting Arista for strategic data center deployment. One important hospital group chose Arista is following an exhaustive Zero Trust working evaluation. Their primary business goal was to protect sensitive patient data through advanced network segmentation while enabling seamless integration to cloud environments.

在醫療保健領域，領先的醫院正在選擇 Arista 進行策略性的資料中心部署。其中一個重要的醫院集團是在完成一項徹底的零信任工作評估後選擇 Arista。他們的主要業務目標是透過進階網路分段來保護敏感的病患資料，同時實現與雲端環境的無縫整合。

Arista secured this win by delivering an automated high-performance fabric that perfectly aligned with their trust nothing verify everything security mandate. By centralizing routing on Arista's R3 spines, we provided a powerful streamlined layer for traffic management tailored to modern health care workloads with a robust lease spine topology.

Arista 透過交付自動化的高效能 Fabric，且與其「不信任任何人、驗證一切」的安全要求完美一致，因而拿下此案。透過在 Arista 的 R3 spine 上集中路由，我們提供了一個強大且精簡的流量管理層，針對現代醫療工作負載量身打造，並具備強健的 leaf-spine 拓撲。

We deployed CloudVision as a service alongside the Arista ABB framework for cloud-based automation and deep visibility. And by integrating Arista's segmentation service MSS, the customer can now dynamically automate the entire system.

我們部署了 CloudVision as a Service，並搭配 Arista ABB 架構，以實現雲端自動化與深度可視性。此外，透過整合 Arista 的分段服務 MSS，客戶如今可以動態自動化整個系統。

In the hospitality sector, and an award-winning regional resort and casino selected Arista to modernize its expanding operations. realizing their legacy network was a bottleneck for innovation, they architected a truly unified standards-based architecture across their data center and campus environments to support new guest experiences and rapidly scale their services.

在飯店與旅遊（hospitality）產業中，一家屢獲殊榮的區域度假村與賭場選擇 Arista 來現代化其持續擴張的營運。意識到其既有網路已成為創新的瓶頸，他們在資料中心與園區環境中打造真正統一、以標準為基礎的架構，以支援新的賓客體驗並快速擴展其服務。

Arista's routing and switching platforms alongside campus access switches are delivering quality and reliability key to performing in mission-critical applications like their demanding casino gaming floor network. They utilize CloudVision and ABD to automate device provisioning and gain the deep real-time visibility required to optimize network performance.

Arista 的路由與交換平台以及園區接入交換器，正提供在其高要求的賭場遊戲樓層網路等關鍵任務應用中所需的品質與可靠性。他們使用 CloudVision 與 ABD 來自動化設備佈署，並取得優化網路效能所需的深度即時可視性。

In all three of these examples, Arista's differentiated EOS provided unmatched quality and reliability and integrated seamlessly with their overarching security ecosystem partners.

在這三個案例中，Arista 具差異化的 EOS 提供了無與倫比的品質與可靠性，並與其整體安全生態系夥伴無縫整合。

Additionally, our TAC delivered the flawless experience and peace of mind to derisk these modern enterprise migrations to Arista. 2026 has really been an exciting year, building momentum across our entire enterprise business around the world in data center, campus, and AI buildouts.

此外，我們的 TAC 提供了零瑕疵的體驗與安心感，降低這些現代企業遷移至 Arista 的風險。2026 年確實是令人振奮的一年，我們在全球整體企業業務於資料中心、園區以及 AI 建置方面持續累積動能。

Thank you.

謝謝。

Thank you, Todd. Wow, that momentum is indeed unprecedented. I haven't experienced anything like this until you arrived. We are pleased with Avitas accelerated net working progress across all our sectors.

謝謝你，Todd。哇，這股動能確實前所未見。在你加入之前，我從未經歷過類似的情況。我們對 Avitas 在各個領域加速的網路進展感到滿意。

Our commitment to fulfilling this accelerated demand in the second half of 2026 is really playing out well with our improved execution in supply chain, partnering closely with our vendors. I want to give actual kudos to you and to your entire manufacturing team led by Urena for the tireless work in progress despite the wide head industry-wide supply chain constraints and shortages.

我們致力於在 2026 年下半年滿足這股加速的需求，並透過改善供應鏈執行、與供應商緊密合作，已展現出良好成效。儘管整個產業面臨廣泛的供應鏈限制與短缺，我仍要實際向你以及由 Urena 領導的整個製造團隊致上肯定，感謝你們不懈的努力與持續推進。

With this backdrop, we are now increasing our annual guidance for the fourth time in 2026 to $13.3 billion, which represents a 48% annual growth.

在此背景下，我們在 2026 年第四次上調全年指引至 133 億美元，代表年增 48%。

At Arista, we find ourselves in a position to address this ever-increasing total accessible market, not just available market, which now stands at north of $200 billion in 2030, therefore, representing a multiyear growth opportunity for us. Our first look at 2027 guidance range puts us in the 38% to 40% range for growth annually to approximately $18.5 billion.

在 Arista，我們正處於能夠因應這個持續擴大的「可觸及總市場」（TAM），而不僅是「可服務市場」的位置；該市場到 2030 年已超過 2,000 億美元，因此為我們帶來多年期的成長機會。我們對 2027 年指引區間的初步展望，年成長率約在 38% 至 40%，營收約達 185 億美元。

Within this envelope, we expect our AI fabrics target to be the fastest growing segment and to double to approximately $7 billion, while our cognitive campus represents a healthy growth of 60%, much faster than the market to approximately $2 billion in 2027.

在此範圍內，我們預期 AI fabrics 目標將是成長最快的區塊，並將倍增至約 70 億美元；同時，我們的認知型園區（cognitive campus）在 2027 年預計將健康成長 60%，遠快於市場，達到約 20 億美元。

We do have an exciting time as we harness the power of cloud and AI revolution. It has certainly surpassed my expectations as we exceeded our original $10 billion goal this year to enter revenue dollars in the teens for both 2026 and 2027. We do so in the typical Arista way, bringing our unique combination of innovation, quality, profitability, efficiency, and zero debt, making our financials very pristine.

當我們掌握雲端與 AI 革命的力量之際，確實正迎來令人振奮的時刻。今年我們超越原先 100 億美元的目標，讓 2026 與 2027 年的營收進入「十幾」億美元區間，這確實超出我的預期。我們以典型的 Arista 方式達成：結合創新、品質、獲利能力、效率與零負債，使我們的財務狀況非常健全。

Our results would not be possible without the cohesive Arista leadership and team, our elite customers and partners and, of course, our support of shareholders.

若沒有凝聚一致的 Arista 領導團隊與同仁、我們頂尖的客戶與合作夥伴，當然還有股東的支持，我們不可能取得這樣的成果。

And with that, over to you, Chantelle, to shed more light on our impressive financials.

接下來交給你，Chantelle，請你進一步說明我們亮眼的財務表現。

Thank you, Jayshree.

謝謝你，Jayshree。

This quarter, we delivered record high revenue growth since our IPO at 58.5%, partnered with robust operating margin above 49%. At the same time, we invested ahead of demand in R&D, inventory and supply commitments to support the AI opportunity in front of us. Congratulations to the Arista team for being named to the S&P 100 in September, a great achievement within a short time frame.

本季我們交出自 IPO 以來最高的營收成長率 58.5%，並搭配超過 49% 的強勁營業利益率。同時，我們在研發、庫存與供應承諾上超前於需求進行投資，以支援眼前的 AI 機會。恭喜 Arista 團隊於 9 月入選 S&P 100，能在短時間內達成這項成就非常了不起。

Let's dive into the financial details. To start off, core revenues in Q3 were approximately $3.66 billion, up 58.5% year over year and above our guidance of $3.3 billion. This significant growth was driven by our continued strength in both AI and enterprise customers.

讓我們深入財務細節。首先，第三季核心營收約為 36.6 億美元，年增 58.5%，並高於我們 33 億美元的指引。這項顯著成長來自我們在 AI 與企業客戶兩方面持續展現的強勁動能。

International revenues for the quarter came in at $682.5 million or 18.6% of total revenue, down from 23% last quarter. This quarter-over-quarter change was primarily influenced by a shift in the geographic mix of sales to our large global customers.

本季國際營收為 6.825 億美元，占總營收 18.6%，低於上季的 23%。這項季對季變化主要受到我們大型全球客戶銷售地理組合轉變的影響。

Year-over-year double-digit growth was achieved in all three geographic segments. The overall gross margin in Q3 was 64%, down from 65.2% in the prior year with supply chain efficiency working to partially offset product cost inflation and up from 63.4% in the prior quarter, driven by end customer mix.

三個地理區隔皆達成年對年雙位數成長。第三季整體毛利率為 64%，較去年同期的 65.2% 下降，供應鏈效率部分抵銷了產品成本通膨；同時較前一季的 63.4% 上升，主要受終端客戶組合帶動。

Operating expense for the quarter were $545.3 million or 14.9% of revenue, up from the last quarter of $411 million. Our R&D spending came in at $408.5 million or 11.2% of revenue, up from the last quarter at $278.1 million.

本季營業費用為 5.453 億美元，占營收 14.9%，高於上季的 4.11 億美元。研發支出為 4.085 億美元，占營收 11.2%，高於上季的 2.781 億美元。

Arista continues to demonstrate its commitment and focus on network innovation with additional investment in liquid cooling, high vatic switching and leading AI software.

Arista 持續展現對網路創新的承諾與專注，並加碼投資於液冷、高 vatic 交換以及領先的 AI 軟體。

Sales and marketing expense was $110.9 million or 3% of revenue, down slightly from 3.6% of revenue last quarter, representative of the highly efficient Arista go-to-market methodology. Our G&A costs came in at $25.9 million or 0.7% of revenue, up slightly from $23.1 million last quarter as we continue to focus on our critical corporate systems and processes.

銷售與行銷費用為 1.109 億美元，占營收 3%，較上季占比 3.6% 略降，反映 Arista 高效率的市場推進（go-to-market）方法論。一般與行政（G&A）費用為 2,590 萬美元，占營收 0.7%，較上季的 2,310 萬美元略增，因我們持續聚焦於關鍵的公司系統與流程。

Our operating income for the quarter was $1.8 billion or 49.1% of revenue, demonstrating our ability to invest in innovation while expanding profitability. Other income expense for the quarter was a favorable $138.6 million, and our effective tax rate was 20.9%.

本季營業利益為 18 億美元，占營收 49.1%，展現我們在擴大獲利能力的同時仍能投資創新。本季其他收益／費用為有利的 1.386 億美元，有效稅率為 20.9%。

Overall, this resulted in net income for the quarter of $1.5 billion or 41.9% of revenue. Our diluted share number was 1.278 billion shares, resulting in a diluted earnings per share for the quarter of $1.20, up significantly with a growth of 55.8% from $0.77 in the prior year period.

整體而言，本季淨利為 15 億美元，占營收 41.9%。稀釋後股數為 12.78 億股，使本季稀釋後每股盈餘為 1.20 美元，較去年同期的 0.77 美元大幅提升，成長 55.8%。

Now turning to the balance sheet. Cash, cash equivalents and marketable securities ended the quarter at approximately $14.4 billion, up from $13.3 billion at the end. Of the $1.5 billion stock repurchase program approved in May of 2025, $817.9 million remain available for repurchase in future quarters. The actual timing and amount of future repurchases will be dependent on market and business conditions, stock price and other factors.

接著看資產負債表。現金、約當現金與有價證券於季末約為 144 億美元，高於期末的 133 億美元。在 2025 年 5 月核准的 15 億美元庫藏股回購計畫中，仍有 8.179 億美元可於未來季度回購。未來回購的實際時點與金額將取決於市場與業務狀況、股價及其他因素。

Now turning to operating cash performance for the third quarter. We generated approximately $1.2 billion of cash from operations. This was driven by a robust earnings performance, coupled with an intentional inventory build to meet our growing customer demand.

接著看第三季的營運現金表現。我們由營運活動產生約 12 億美元現金。這主要由強勁的獲利表現所帶動，並搭配為滿足日益成長的客戶需求而有意進行的庫存增加。

DSOs came in at 59 days, down from 68 days in Q2 due to the timing of customer shipments and invoicing. Our inventory turns dropped slightly from 1.7% in Q2 to 1.6% for the quarter. We ended the quarter with $3.2 billion in inventory, up from $2.5 billion last quarter.

應收帳款週轉天數（DSO）為 59 天，較第二季的 68 天下降，原因在於客戶出貨與開立發票的時點。庫存週轉率由第二季的 1.7% 略降至本季的 1.6%。本季季末庫存為 32 億美元，高於上季的 25 億美元。

Inventory level fluctuations are expected to continue as we work through the balancing of components timing and availability. This could result in quarters of elevated inventory balances affecting the timing of cash flow from operations ahead of customer deployments.

隨著我們持續處理元件時程與供應可得性的平衡，預期庫存水位的波動仍將延續。這可能導致某些季度出現較高的庫存餘額，進而影響在客戶部署之前來自營運活動的現金流入時點。

Our purchase commitments at the end of the quarter were $11.4 billion, up from $9.7 billion at the end of Q2. As mentioned in prior quarters, this expected activity mostly represents purchases for chips related to new products and AI deployments.

本季末我們的採購承諾為114億美元，高於第二季末的97億美元。如同先前季度所提及，這些預期活動多數代表與新產品及AI部署相關的晶片採購。

Our total deferred revenue balance was approximately $6.8 billion, down from $6.9 billion in the prior quarter. The majority of the deferred revenue balance is product related. Our product deferred revenue decreased $86 million sequentially versus last quarter.

我們的遞延收入總餘額約為68億美元，低於上一季的69億美元。遞延收入餘額的大部分與產品相關。我們的產品遞延收入較上季按季減少8,600萬美元。

We remain in a period of ramping our new products, winning new customers and expanding new use cases, including AI. These trends have resulted in increased customer-specific acceptance clauses and an increase in the volatility of our product deferred revenue balances.

我們仍處於新產品放量、贏得新客戶並擴展新使用情境（包括AI）的階段。這些趨勢導致客戶特定的驗收條款增加，並使我們的產品遞延收入餘額波動性上升。

As mentioned in prior quarters, the deferred balance can move significantly on a quarterly basis, independent of underlying business drivers. Accounts payable days were 48 days, down from 57 days in Q2, reflecting the timing of inventory receipts and payments.

如同先前季度所提及，遞延餘額可能在季度間大幅變動，且不一定與基本業務驅動因素同步。應付帳款天數為48天，低於第二季的57天，反映庫存收貨與付款時點的影響。

Capital expenditures for the quarter were $100.7 million. Our construction work to build expanded facilities in Santa Clara remains on track, and we expect construction to be completed in the coming months.

本季資本支出為1.007億美元。我們在聖塔克拉拉擴建設施的工程進度仍按計畫推進，預期將於未來數月內完工。

As we look ahead, the momentum is strong. We are grateful to our customers and proud of our team's continuous ability to rise to meet the demands of the networking industry.

展望未來，動能依然強勁。我們感謝客戶的支持，也以團隊持續挺身而出、滿足網路產業需求的能力為榮。

For the fourth quarter, we are forecasting a range of revenue between $3.9 billion and $4 billion. This positions fiscal year 2026 revenue just over $13.3 billion at the midpoint. Our fourth-quarter gross margin outlook is 63% to 64% and operating margin is between 48% and 49%.

針對第四季，我們預估營收區間為39億至40億美元。以中位數計算，這將使2026會計年度營收略高於133億美元。我們對第四季毛利率的展望為63%至64%，營業利益率為48%至49%。

Diluted EPS is forecasted to be between $1.26 and $1.30 with approximately 1.28 billion diluted shares, and the effective tax rate is expected to be approximately 21%.

稀釋後每股盈餘（EPS）預估介於1.26至1.30美元，稀釋後股數約為12.8億股，且有效稅率預期約為21%。

Looking further ahead to 2027, the momentum continues. For fiscal year 2027, the revenue growth outlook is between 38% and 40%, equating to approximately $18.5 billion the midpoint. Gross margin continues to be in the range of 62% to 64% and an operating margin of approximately 49%. Our effective tax rate is expected to be approximately 21%.

再往前看至2027年，動能仍將延續。就2027會計年度而言，營收成長展望介於38%至40%，以中位數計算約為185億美元。毛利率仍維持在62%至64%的區間，營業利益率約為49%。有效稅率預期約為21%。

In closing, there is a lot to be excited about here at Arista. We will continue to be focused on our innovation, execution and delighting customers in the Arista way. Now back to you, Rudy, for Q&A.

總結來說，Arista有許多令人振奮的進展。我們將持續專注於創新、執行力，並以Arista的方式讓客戶滿意。接下來把時間交還給你，Rudy，進入問答。

Thank you, Chantelle.

謝謝你，Chantelle。

We will now move to the Q&A portion of the Arista earnings call. (Event Instructions) Regina, please take it away.

我們現在進入Arista財報電話會議的問答環節。（活動說明）Regina，請開始。

(Operator Instructions) Michael Ng, Goldman Sachs.

（接線員指示）Michael Ng，高盛。

I wanted to ask about your comment around the greater than 10 gigawatts of deployments across 10 plus compute accelerators for next year. Could you just talk a little bit about whether that is an industry number or an Arista-specific number in terms of those deployments?

我想就你提到明年在10個以上運算加速器上、部署超過10吉瓦的說法提問。你能否多談一點，這個部署數字是指整體產業的數字，還是Arista自身的數字？

And how does the diversification of compute, potentially create more opportunities for Arista networking relative to just having a handful of compute suppliers?

另外，運算的多元化如何可能相較於只有少數幾家運算供應商，為Arista網路帶來更多機會？

Yes. Thanks, Michael. As you know, I think much of AI, even in our world, started initially being compute-centric, and we're seeing that change to network-centric. And so as part of the accelerated networking strategy, we are supporting three types of AI fabrics, scale up, scale out and scale across.

是的，謝謝你，Michael。如你所知，我認為AI在很大程度上，即使在我們的領域，一開始也是以運算為中心，而我們正看到它轉向以網路為中心。因此，作為加速網路策略的一部分，我們支援三種類型的AI fabric：scale up、scale out，以及scale across。

So our guess is by the end of '27, we will deploy largely in '27 is where the exponential ramp happened across or greater than 10 gigawatts.

因此我們的推估是，到27年底，我們將在27年進行大規模部署，也就是指數式放量發生的時間點，規模達到或超過10吉瓦。

Now -- some of that -- much of that gigawatt may be new, but remember, in scale across, we can also have existing gigawatts being deployed. So we're talking about new distributed 10 gigawatts, but they may already be multiple gigawatts deployed across. So the rationale here is, while we are new to scale up, we are touching a variety of XPUs and scale out and scale across, across the rack and across data centers.

現在——其中一部分——許多吉瓦可能是新增的，但請記得，在scale across中，我們也可能部署既有的吉瓦規模。所以我們談的是新增、分散式的10吉瓦，但也可能已經有多個吉瓦在scale across上部署。因此這裡的邏輯是，雖然我們在scale up方面是新進者，但在scale out與scale across方面，我們接觸到各式各樣的XPU，涵蓋機櫃內以及跨資料中心。

And we already know with our scale across, we are touching multiple XPUs, four or five of them. With the advent of scale up, we will add another 5 in 2027.

而且我們已經知道，在scale across方面，我們正在接觸多種XPU，大約四到五種。隨著scale up的到來，我們將在2027年再增加另外5種。

Meta Marshall, Morgan Stanley.

Meta Marshall，摩根士丹利。

Congrats on the quarter. I guess just Jayshree, if you could give an update just on how you see that mix of scale across and scale out and scale up just as a mix of your business, clearly, scale up, nothing today. I think last quarter, you had said you expected 1/3 of the demand this year to be from scale across just as you've continued to increase guidance. And as you look into '27, just what do you see those mixes being?

恭喜本季表現。我想請教Jayshree，你能否更新一下你如何看待scale across、scale out與scale up在你們業務中的組合占比？顯然目前scale up還沒有。我記得上季你提到，今年需求中預期有三分之一來自scale across，但你們持續上調指引。展望到27年，你認為這些組合占比會是什麼樣子？

Yes. I think in '27, Meta, thank you for the kind wishes. We've been working hard for this day. The scale-up will still be in early trials. So I don't think there's a material number associated with scale up in '27. I do think it could contribute in '28.

是的。我想在27年，Meta，謝謝你的祝福。我們為這一天努力了很久。Scale-up仍將處於早期試驗階段。因此我不認為在27年scale up會有具體且重大的數字貢獻。我確實認為它可能在28年開始貢獻。

So then that -- if scale-up is virtually zero, then we're basically left with most of the use case being scale out and scale across. And as you've heard us say in the past, scale across, we think, will be at least 1/3, we'll probably go higher in '27.

因此——如果scale-up幾乎為零——那我們基本上就剩下多數使用情境會是scale out與scale across。而且如同我們過去所說，我們認為scale across至少會占三分之一，在27年可能會更高。

And so the split this year is roughly 70-30. The split next year may be -- I can't venture to guess, but I think the 70% may come down and the 30% would go up when it comes to scale across because it's a very strong use case for us. Arista is highly differentiated with our 7800 scale across platform.

因此今年的拆分大約是70-30。明年的拆分我不便猜測，但我認為在scale across方面，70%可能會下降、30%會上升，因為這對我們而言是非常強勁的使用情境。Arista在我們的7800 scale across平台上具備高度差異化。

And unsurprisingly, the word scale means a lot to us there and our customers, scale for security, scale for performance, scale for reliability and there's just no room for error when you're connecting million of XPUs. So we feel very comfortable that these are our two strongest use cases.

而且不意外地，「scale」這個詞對我們與客戶都非常重要：為安全性而擴展、為效能而擴展、為可靠性而擴展；當你在連接數百萬個XPU時，完全沒有犯錯空間。因此我們非常有信心，這兩個是我們最強的使用情境。

I also think it will fuel -- I mean, a lot of this is back in AI fabric. We will come back to revisit the question of front end as well as we move to more inference and the line between back and front becomes thinner.

我也認為這將帶動——我的意思是，很多都回到AI fabric。隨著我們走向更多推論（inference），後端與前端之間的界線會變得更薄，我們也會回頭再檢視前端的問題。

Ben Reitzes, Melius Research.

Ben Reitzes，Melius Research。

Jayshree, a pretty good quarter and guide there. So I just want you've previously referred to potentially 2 10%-plus customers into the fourth quarter. Do you mind talking about whether they're landing in that range. And what the use cases are for? And if you're excited about those customers still and into '27?

Jayshree，這一季的表現和指引都相當不錯。所以我想問，你之前提到第四季可能會有 2 個占比超過 10% 的客戶。方便談談他們是否會落在那個區間嗎？以及他們的使用情境是什麼？還有你是否仍對這些客戶感到興奮，並且一路延續到 2027 年？

Yes, yes, yes. So I guess diversity is the name of the game and the team today, right, Chantelle? We have diverse accelerators. We have diverse models. We want to have diverse 10% customers, too.

是的、是的、是的。所以我想，多元化就是今天的主題，對吧，Chantelle？我們有多元的加速器。我們有多元的模型。我們也希望有多元的 10% 客戶。

So as we said before, our dearest ones are M&M, we fully expect our top prior 10% customers to continue to be with us in 2016, and we hope beyond and we expect to add at least one, maybe two. And a lot of that depends on whether Todd and Nehemia can ship enough product. So we'll let them decide and we'll give them all the way into December 31.

所以如同我們之前說過的，我們最珍視的是 M&M；我們完全預期先前的前幾大 10% 客戶在 2016 年仍會與我們合作，我們也希望能延續更久，而且我們預期至少再新增一個、也許兩個。而這在很大程度上取決於 Todd 和 Nehemia 能否出貨足夠的產品。所以我們就讓他們來決定，我們會一路給他們到 12 月 31 日。

We'll take every day.

我們會把每一天都用上。

So we're feeling very good about at least three and maybe four.

所以我們對至少三個、也許四個感覺非常好。

George Notter, Wolfe Research.

George Notter，Wolfe Research。

I was just curious about the big sequential jump in R&D. It was really, really significant. I'm just wondering if that included some onetime items.

我只是好奇研發費用（R&D）季對季大幅跳升，幅度真的非常、非常顯著。我想了解其中是否包含一些一次性項目。

Also, is this the new run rate in R&D going forward? Any perspective there would be great.

另外，這會是未來研發費用的新常態水準嗎？如果能提供一些觀點就太好了。

Okay. So over to you, Ken, why do you spend so much money this year?

好的。那交給你了，Ken，你今年為什麼花這麼多錢？

So much new stock. The opportunity is so many of these AI tolerating systems. We are -- our success with the hyperscaler AI companies depends on our ability to partner. Our ability to partner is based on our understanding of each of their unique requirements and constraints.

這麼多新的股票。機會在於有這麼多能容忍 AI 的系統。我們在超大規模（hyperscaler）AI 公司上的成功，取決於我們的合作能力。而我們的合作能力，建立在我們對他們各自獨特需求與限制的理解之上。

And also, there's a new technology introduction around liquid cooling. This requires a major infrastructure upgrades and technology upgrades on our part. So I think that's what's really driving the expansion spending there.

另外，還有一項關於液冷的新技術導入。這需要我們在基礎設施與技術上做重大升級。所以我認為這才是真正推動那邊支出擴張的原因。

Sorry, George. And just in your comment is the new run rate. I'll still put the framework between 9% and 11% as we work through. So some quarters like we've talked about, we'll hit this range as we have specific investments. And so I would still keep it on that into 11% more to come as we guide next year in more detail.

抱歉，George。另外針對你提到的新常態水準，我仍會把框架放在 9% 到 11% 之間，因為我們還在推進中。所以有些季度，如同我們談過的，當我們有特定投資時，就會落在這個區間。因此我仍會維持到 11% 的說法；更多細節會在我們對明年提供更詳細指引時再說明。

But I noticed the prototype bill was very high this quarter, Ken.

但我注意到這一季的原型機帳單非常高，Ken。

That's right because of the opportunities in the hyperscaler AI space.

沒錯，因為超大規模 AI 領域的機會。

Amit Daryanani, Evercore ISI.

Amit Daryanani，Evercore ISI。

Thanks a lot, and congrats on a really good set of numbers over here. Jayshree, you talked initially about the industry moving towards a more network network-centric architecture. As customers deploy different type of accelerators, does Arista become involved earlier in designing the overall systems rather than connecting them afterwards?

非常感謝，也恭喜你們交出一組非常漂亮的數字。Jayshree，你一開始提到產業正朝向更以網路為中心（network-centric）的架構移動。當客戶部署不同類型的加速器時，Arista 是否會更早介入整體系統的設計，而不是事後再把它們連接起來？

And as you have to work with a more distributed disaggregated compute network, does that change your networking content per deployment? I know you talked about 10 gigawatts. I'm just sort of wondering is the network intensity start to change as well?

另外，當你們必須面對更分散、解耦（disaggregated）的運算網路時，這會改變每次部署中你們的網路內容占比嗎？我知道你提到 10 吉瓦。我只是想了解，網路密度（network intensity）是否也開始改變？

And then maybe for Chantelle, as you think about the 38%, 40% guide for '27, which is extremely impressive, can you just touch on your confidence around the supply chain execution to meet this demand? What's in the bag versus what's to needs to be closed?

然後也許請 Chantelle 回答：當你們思考 2027 年 38%、40% 的指引（非常令人印象深刻）時，能否談談你對供應鏈執行力、以滿足這些需求的信心？哪些是已經確定在手的，哪些還需要去完成（close）？

This is a very good question. I think until now, we've enforced our architecture mostly on scale out and scale across. And a lot of that had to do with distributed power location, moving to regional edge edges a very software defined type of network with the intensity, like I described of package, flows, performance, predictable latency, et cetera.

這是一個非常好的問題。我認為到目前為止，我們的架構主要著力在 scale out 與 scale across。而這在很大程度上與分散式電力位置有關，轉向區域邊緣（regional edge edges）是一種非常軟體定義的網路型態，其強度如我所描述，包含封包、流量、效能、可預測的延遲等等。

As we start participating with new accelerators, we get a greater chance early in the cycle to see the behavior of these exact accelerators. two or three examples I'd like to highlight is our close work with Cerebrus, our work with Microsoft and Maya. These have been going on in the last couple of years and has allowed us to tune our buffers, tune our software performance and maximize the IL-2, if you will.

當我們開始參與新的加速器時，我們在週期早期就有更大的機會觀察這些特定加速器的行為。我想強調兩三個例子：我們與 Cerebrus 的緊密合作、我們與 Microsoft 和 Maya 的合作。這些在過去幾年一直在進行，讓我們得以調校緩衝（buffers）、調校軟體效能，並且把 IL-2（如果你願意這麼稱呼）最大化。

The recognition of those scale-up designs are still in the works, but they're already manifesting itself and scale out and scale across. So you wrap a very good point, not only in the actual work we're doing with these accelerators, but the aggregate impact it has on our network design in terms of scale, performance and reliability.

對於那些 scale-up 設計的認可仍在進行中，但它們已經在 scale out 與 scale across 上展現出來。所以你提到了一個很好的點：不僅是我們與這些加速器實際合作的工作本身，還有它對我們網路設計在規模、效能與可靠性方面所帶來的整體影響。

Yes. And then I think to part two of your question, Amit, and thank you for the kind words at the beginning. I think -- so I am confident in the guide and I am continent in the supply chain. What makes me confident in the supply chain is internally, coding Yohan and the team doing a fantastic job every day being scrappy to make sure we can do what we need to do to deliver to our customers.

是的。然後我想回到你問題的第二部分，Amit，也謝謝你一開始的好話。我想——所以我對這個指引有信心，我也對供應鏈有信心。讓我對供應鏈有信心的是，在內部，Yohan 和團隊每天都做得非常出色、很拚，確保我們能做需要做的事，把產品交付給客戶。

And also, I have confidence in a lot of great conversations and agreements with some of our key suppliers, and that's why you see our purchase commitments increasing because we have that confidence. So internally, externally confident.

此外，我也對與一些關鍵供應商之間許多良好的對話與協議有信心；這也是為什麼你會看到我們的採購承諾增加，因為我們有那樣的信心。所以，內部有信心、外部也有信心。

As far as visibility, I think we have two quarters of visibility as we always have, and we need into the other two. But given our preyfungible portfolio, we're pretty clear where to make that bet. So we're pretty excited about next year.

至於能見度，我認為我們一如既往有兩個季度的能見度，而另外兩個季度我們還需要再爭取。但考量我們的「preyfungible」產品組合，我們相當清楚該把賭注下在哪裡。所以我們對明年非常興奮。

Tim Long, Barclays.

Tim Long，Barclays。

Maybe if we could shift a little bit and talk about the campus and branch. A pretty good year this year and guiding it to 60% next year. Could you talk a little bit about kind of what's enabling you to grow so much faster than the industry?

也許我們可以稍微轉換一下，談談校園（campus）與分支（branch）。今年表現相當不錯，而且你們指引明年成長到 60%。你能否談談是什麼讓你們能夠比產業成長快這麼多？

And is this something that you see as a multiyear type of dynamic? Obviously, you got strength in enterprise data centers. But curious what else is helping this much higher than industry growth.

而且你是否認為這會是一種跨多年（multiyear）的動態？顯然你們在企業資料中心很強。但我也好奇，還有哪些因素在推動你們遠高於產業的成長。

Yes. Thanks for the question. In a lot of ways, I think the enterprise app market is coming to us. Arista has always led the way in quality and reliability. And over the past few years, just practically every significant enterprise has shifted gears and really considered their campus networks, their branch now works mission critical with the highest need for reliability.

是的，謝謝你的問題。在很多方面，我認為企業應用市場正在向我們靠攏。Arista 一直在品質與可靠性方面領先。而在過去幾年，幾乎每一家重要的企業都已經換檔，並且真正重新審視他們的校園網路與分支網路，如今這些都在執行關鍵任務（mission critical），對可靠性的需求最高。

This is only accelerated because of the glass wing and ethos effect where people are really focused deeply not only on providing the highest level of reliability, but being able to do that as a rapidly patch and adjust their infrastructure over time, and the differentiation with EOS and the ability for Arista software to be able to do this with practically no downtime with extremely, extremely high uptime.

這之所以能加速，完全是因為 Glasswing 與 Ethos 效應：大家不僅非常專注於提供最高等級的可靠性，也能隨時間快速修補並調整其基礎架構；再加上 EOS 的差異化，以及 Arista 軟體幾乎可在零停機的情況下做到這一點、並維持極高（極高）的上線時間。

It's highly differentiated in the market. And I really think that's the core of what's driving market share gains here.

在市場上具有高度差異化。我真的認為，這正是推動我們在此取得市占率提升的核心因素。

I think the other thing that makes us an incremental process is that campus refresh cycles are pretty long, believe in customers understand the Arista advantage may take some time to adopt us because of the schedule in which they depreciate and upgrade their year. So I think we can expect this to be a multiyear cycle.

我認為另一個讓我們呈現漸進式成長的原因是，校園網路的汰換週期相當長；即使客戶理解 Arista 的優勢，也可能因為其折舊與年度升級的時程，而需要一些時間才會採用我們。因此我認為，我們可以預期這會是一個為期多年的循環。

David Vogt, UBS.

David Vogt，瑞銀（UBS）。

So maybe, Chantelle, one for you, and I'll wrap two together in one question. Can you kind of help us understand sort of your purchase commitment dynamic and your deferred revenue dynamic against strong revenue guide that you provided for 2027? So obviously, we recognize that you're confident in the number, but deferred revenue, I understand that it's volatile and it varies quarter to quarter.

那麼也許，Chantelle，這題給你，我把兩個點合併成一個問題。你能否幫我們理解一下：在你們對 2027 年給出強勁營收指引的背景下，你們的採購承諾（purchase commitment）動態，以及遞延收入（deferred revenue）動態應該怎麼看？顯然，我們理解你們對這個數字很有信心，但遞延收入我也理解它具有波動性，且會隨季度而變化。

But just help us understand how to think about that metric in the context of a 38% to 40% guide, particularly also with purchase commitments going up a couple of billion dollars sequentially.

但請協助我們理解：在 38% 到 40% 的指引情境下，該如何看待這個指標；尤其是採購承諾也連續一季增加了數十億美元。

Yes. Thank you. Thanks, David. I think there's a few things. I don't know if they correlate directly, but they're all in purpose and designed to serve the customers as we need to, exiting Q4 and into 2027. Purchase commitments from that perspective, some things additionally we added.

是的，謝謝。謝謝你，David。我想有幾點。我不確定它們是否直接相關，但它們的目的都是一致的，都是為了在我們走出第四季並邁向 2027 年時，能按客戶需求提供服務。就採購承諾而言，我們也額外加入了一些項目。

Now we have memory in there a memory agreement, which has been that. So that's a multiyear agreement that you'll see as the Q comes out.

現在我們把記憶體也納入其中，包含一份記憶體協議，就是那個。因此那是一份多年期協議，你會在 Q（文件/揭露）出來時看到。

I think on the purchase commitment side, you can probably expect that to go up a little bit more as we exited Q4 as we get ready for this 2027 guide we have. As far as the deferred revenue, deferred revenue actually had the largest churn in and out and still remained flat and still delivered a 58.5% revenue growth quarter, which I think is a fantastic outcome.

我想在採購承諾方面，隨著我們走出第四季、並為我們的 2027 指引做準備，你大概可以預期它還會再增加一些。至於遞延收入，遞延收入其實是進出變動（churn）最大的一項，但仍然維持持平，同時我們仍交出 58.5% 的季度營收成長，我認為這是非常出色的結果。

So there, you're seeing things exiting getting released and things to get booked back to the deferred balance, which I think is a really healthy sign of the demand that's fueling that 38% -- excuse me, to 40% revenue growth next year.

所以你會看到有些項目在期末釋放（recognize/release），也有些項目再回到遞延餘額中入帳（book）。我認為這是需求健康的訊號，而這些需求正是支撐明年 38%——抱歉，是 40%——營收成長的動能。

So I think they're all related purchase comes grow to serve that demand in the first day flat to deliver that outcome to the P&L, and we're pretty excited to see how that continues through 2027.

所以我認為它們彼此都有關聯：採購承諾成長是為了滿足那個需求；遞延收入維持持平，並能把結果反映到損益表（P&L）上；我們也很期待看到這個趨勢在 2027 年如何延續。

Simon Leopold, Raymond James.

Simon Leopold，Raymond James。

Appreciate that. I wanted to get one quick clarification. You gave us targets for 2027 for campus in AI. Just wanted to check on what the baselines are for 2026. And the question was really on a topic that I think Andy has been vocal about, but your thoughts about the evolution of switching with the inclusion of either CPO or NPL, what does that mean to Arista?

感謝說明。我想快速釐清一點。你們給了 2027 年校園與 AI 的目標。想確認一下 2026 年的基準（baseline）是什麼。另外一個問題是我認為 Andy 也多次談過的主題：你們怎麼看交換器（switching）的演進，包含導入 CPO 或 NPL 的情況，這對 Arista 意味著什麼？

Okay. Just to clarify, the baseline upon 38% to 40% is the '26 guide we gave of 13 --

好的。先釐清一下：38% 到 40% 的基準，是我們給出的 2026 年指引 13——

Yes, $13.3 billion. Yes. And within that AI is $3.5 billion and campus is $1.25 billion.

是的，133 億美元。是的。而在其中，AI 是 35 億美元，校園是 12.5 億美元。

Yes. We're staying with the target numbers because we'll only know actuals after we ship everything end of the year. And to your second question on Optics. Optics is very integral to Arista strategy. Andy Bessen, our Chief Architect, lives and read this. It doesn't mean we build everything, but we actually systematically put it together.

是的。我們維持這些目標數字，因為只有在年底把所有產品出貨後，我們才會知道實際數字。至於你第二個關於光學（Optics）的问题。光學是 Arista 策略中非常關鍵的一環。我們的首席架構師 Andy Bechtolsheim（Andy Bessen），長期投入並深耕這一塊。這不代表我們什麼都自己做，但我們會以系統化方式把它整合起來。

We, first of all, strongly believe in copper when you can and optics a few months because when you go beyond 1 or 2-meter radius, the flexibility of pluggable optics and lasers is probably the most widely deployed and accepted. Andy was a pioneer of OSFP that gave us a very wide density of thermal limits and widely used.

首先，我們強烈相信：能用銅纜就用銅纜；而在需要光學的情境下，當距離超過 1 或 2 公尺半徑時，可插拔光學與雷射的彈性，可能是部署最廣、接受度最高的方案。Andy 是 OSFP 的先驅之一，OSFP 讓我們在散熱限制下仍能提供非常高的密度，且被廣泛採用。

And now our new coherent optics LPO models is something that goes long distance without requiring DSP. And then to meet higher front-panel density, we're very excited about extra rent pluggable Optics that can -- that is being supported by over 150 vendors. So we think the world is going to remain in a very majority fashion on pluggable optics.

而現在，我們新的相干光學 LPO 模型，可以在不需要 DSP 的情況下支援長距離傳輸。為了滿足更高的前面板密度，我們也對「外部雷射可插拔光學」（external laser pluggable optics）感到非常興奮——這項技術已獲得超過 150 家供應商支援。因此我們認為，世界仍將以「可插拔光學」為主流。

Having said that, I think there's always room for complementing the pluggables with a strong adoption at higher speeds to what we call near packaged optics or open CPO. We don't want to build proprietary co-package optics and to solve these constraints, Arista is pioneering this open CPO. That instead of sovereign components, we can use socketed optical engines and picaybers to allow these modules to be pretested, et cetera.

話雖如此，我認為在更高速的情境下，仍有空間以更強的採用來補充可插拔方案，也就是我們所稱的近封裝光學（near packaged optics）或開放式 CPO（open CPO）。我們不想打造專有的共封裝光學（co-packaged optics）；為了解決這些限制，Arista 正在推動開放式 CPO：也就是不使用封閉的（sovereign）元件，而是使用可插槽（socketed）的光學引擎與（picaybers），讓這些模組能預先測試等等。

So we are big fans of these. We think they'll come on to the market in the '28 to '30 time frame. It's not imminent. But meanwhile, long live pluggable optics.

所以我們是這些技術的忠實支持者。我們認為它們會在 2028 到 2030 年的時間框架進入市場。短期內不會立刻發生。但同時，可插拔光學長存。

Joseph Cardoso, JPMorgan.

Joseph Cardoso，摩根大通（JPMorgan）。

Congrats on the results as well. Maybe the question that I wanted to touch on and maybe it kind of dovetails off of a clarification on the last one. When we take a look at the initial full year guide for 2020, particularly as it relates to maybe the non-AI noncampus business. I think growth is being implied to moderate from the strong 2026 levels.

也恭喜你們的業績表現。我想問的問題，可能也延伸自上一題的釐清。當我們看你們對 2020 年的初始全年指引時，特別是與非 AI、非校園業務相關的部分。我認為市場隱含的成長率，似乎會從 2026 年的強勁水準放緩。

I was just wondering if you can kind of touch on the underlying assumptions or factors that may be driving a little bit more moderation here into 2027 and whether that's maybe just tough compares or conservatism in terms of supply, et cetera, just trying to understand directionally what's the drivers of the 2027 forecast that relates to the maybe core business?

我想請你談談：有哪些底層假設或因素，可能驅動 2027 年成長較為溫和？這是否只是比較基期較高，或是供應等方面採取較保守的假設等等；我想方向性地理解，與核心業務相關的 2027 年預測驅動因素是什麼？

Thank you, Joe. Well, listen, when some things are growing 60% and 100%, and then you say the overall growth is 38% to 40%, then obviously, some of the things are not growing as fast, right? We are very comfortable that everything is going to grow double digits, some more, some less.

謝謝你，Joe。嗯，聽著，當有些項目成長 60% 或 100%，而你又說整體成長是 38% 到 40%，那很顯然，有些項目的成長速度就不會那麼快，對吧？我們非常有信心所有項目都會維持雙位數成長，有些更高、有些較低。

And we're also very comfortable that given a first guide literally over a year before that all sectors are doing well for us. AI, enterprise, specialty cloud providers, front-end networks. So we started out 2026, wondering with front end to grow -- so it's a cyclical thing.

而且我們也非常有信心：以第一版指引來說，這是提前一年以上給出的，所有領域對我們而言都表現良好。AI、企業、專業雲端服務供應商、前端網路。所以我們在 2026 年初開始時，還在思考前端是否會成長——這是一個循環性的事情。

When the back end grows, the front end may go a little slower and the bet slows down the front end, may grow a little faster. So I wouldn't read any conservatism into these numbers. These are widely higher and optimistic than we had once projected. So enjoy them for a moment and hopefully, they'll get better.

當後端成長時，前端可能會稍微慢一些，而這項押注會讓前端放慢，但後端可能會成長得稍微快一些。所以我不會從這些數字中解讀出任何保守的意味。這些數字比我們曾經預估的要高得多，也更樂觀。所以先享受一下，希望它們還會更好。

James Fish, Piper Sandler.

James Fish，Piper Sandler。

Fantastic quarter and guide here to start for '27 ladies. And can pose gentlemen and two ladies yes. And Todd, I didn't forget about you, Todd. But look, I appreciate the early outlook, but you guys always talk about one to two quarters of visibility. You tend to be a very prudent team.

這一季表現非常出色，對於從現在開始到'27年的指引也很棒，各位女士。也可以請各位先生以及兩位女士提問，是的。Todd，我沒有忘記你，Todd。不過我很感謝你們這麼早就提供展望，但你們一向都說只有一到兩季的能見度。你們通常是一個非常審慎的團隊。

And so maybe going back to, I think it was Amit's question, I guess the confidence in providing this durability of this scale, this early and why we're going to see this kind of strength persist why Arista for scale up in 2028.

所以也許回到我想是Amit的問題：我想問的是，你們為什麼能在這麼早的時間點，就有信心提供這種規模的持續性（durability），以及為什麼我們會看到這種強勁表現延續、為什麼Arista在2028年仍能在scale-up領域擴張？

Well, I'll just go back and thank you for the kind words.

好，我先回應一下，也謝謝你的好話。

The kind words for ladies and gentlemen.

對各位女士與先生的好話。

That's right to all of us. I think that -- so let's go back and reiterate the confidence level. There was a lot of thought obviously put into a guide that's strong this early. So there's a lot of attention with Ken and Jayshree and myself and team. So we had to get comfortable first, and we're very comfortable with all the things we've done to see what we need to go ahead.

沒錯，對我們所有人。我想——我們回到並重申一下信心程度。顯然，我們在這麼早就給出強勁指引這件事上，投入了很多思考。因此Ken、Jayshree、我以及團隊都非常關注。所以我們必須先讓自己感到安心，而我們對於為了推進下一步所做的一切都非常有把握。

And again, the two quarters visibility, two years of relationship and technical conversations. So somewhere between the second half lands and it's the usual kind of cadence we have. So nothing has changed there.

再說一次，兩季的能見度，背後其實是兩年的合作關係與技術層面的對話。所以大概落在下半年，這就是我們一貫的節奏。因此這方面沒有任何改變。

I think from -- and then the second part of your question was about scale up going into 2027, 2028, Jayshree?

我想——你問題的第二部分是關於在2027、2028年進入scale-up的擴張，Jayshree？

Yes. So listen, I think scale-up has been a tough market for us because it doesn't have all the software advantages that we bring, but it absolutely has the hardware skills and all the non-Nvidia accelerators that Arista is in the best position to connect to.

是的。聽著，我認為scale-up對我們來說一直是個艱難的市場，因為它沒有我們所帶來的所有軟體優勢，但它絕對具備硬體能力，而且Arista在連接所有非Nvidia加速器方面處於最佳位置。

So this movement from stand-alone basic switches to that scale is something that Andy, Alex, you hold book, the entire team has been working on and codeveloping for a couple of years. It's hard work. Just as we are staring at InfiniBand and now we can do, of course, it's going to be Ethernet.

因此，從獨立的基礎交換器走向那種規模的架構，是Andy、Alex、你（們）以及整個團隊過去幾年一直在努力並共同開發的方向。這是艱苦的工作。就像我們過去盯著InfiniBand，而現在我們當然可以做到——它將會是乙太網路（Ethernet）。

We're looking forward to a time where we were staring at proprietary UA Link or NVLink and we're going to go, of course, it's going to be ESUN, but these type of migrations to do are multiyear. And so we're very confident not only in our '27 guide, but a promising '28 ahead as well.

我們也期待有一天，當我們曾經盯著專有的UA Link或NVLink時，會回頭說：當然，會是ESUN；但這類遷移需要多年時間來完成。因此，我們不僅對'27年的指引非常有信心，也對接下來有前景的'28年同樣有信心。

And the reason we are is because even though we have two quarters of visibility, we have two years of work behind us.

我們之所以如此，是因為即使我們只有兩季的能見度，但我們背後已經累積了兩年的工作。

Tal Liani, Bank of America.

Tal Liani，美國銀行（Bank of America）。

Ladies and gentlemen, sorry, I'll start from the beginning. Thank you, I missed the call last time, and you raised the guidance substantially, and now you are doing it again.

各位女士與先生，抱歉，我從頭開始。謝謝，我上次錯過了電話會議，你們大幅上調了指引，而現在你們又再做一次。

And on the other hand, I heard Chantelle explanation about deferred revenues, but the sequentials are slowing down. If I look at total Ferd and current deferreds are slowing down. But you are feeling more comfortable with next year and you're guiding well, well above the Street.

另一方面，我聽到Chantelle對遞延收入的解釋，但季比（sequential）增速正在放緩。如果我看總遞延與流動遞延（current deferred），都在放緩。但你們對明年反而更有把握，而且給出的指引遠遠高於市場共識。

So the question is, when you look at orders versus recognizing from deferred because of the availability of components because of releasing kind of deferred revenues because of completion of build-out, et cetera. So what are the drivers for -- like same question about from different angles, what are the drivers for the growth? Is it orders? Or is there some part there of recognition of historical orders?

所以問題是，當你們看訂單（orders）相對於從遞延中認列（recognizing from deferred），考量到零組件供應可得性、因建置完成而釋放部分遞延收入等等。那麼成長的驅動因素是什麼——同一個問題從不同角度問：成長的驅動因素是什麼？是訂單嗎？還是其中有一部分是歷史訂單的認列？

And then is deferred revenue actually losing some of its predicting ability? Just because I see some decoupling between the deferred revenue trends and revenue growth trends.

另外，遞延收入是否正在失去部分預測能力？因為我看到遞延收入趨勢與營收成長趨勢之間似乎出現某種脫鉤。

Okay. Let me start and then, maybe Chantelle, will you take over? So I think, first you tell -- we missed you last quarter because we don't get this level of detail from anyone else on exactly what are the puts and takes. I just broadly look at you and say, hey, it's great demand.

好。我先回答，然後也許Chantelle你再接著說？我想，首先——我們上季很想念你，因為我們從其他人那裡得不到這麼細的細節，能把各種增減因素（puts and takes）講得這麼清楚。我通常只是大方向看著你，說：嘿，需求很強。

And if you want to dissect demand into how much is fitting into deferred, how much is fitting into our orders, how much is getting shipped into revenue, you're turning demand into a physics problem that I don't have the answer for. Because I don't predict or forecast deferred.

如果你想把需求拆解成：有多少進到遞延、有多少進到我們的訂單、有多少出貨轉成營收——你等於把需求變成一個物理題，而我沒有答案。因為我不會去預測或預估遞延。

Deferred is a strong function of a new use case or a new product that eventually is not a new use case or a new product. So we had millions of dollars come out of deferred if this use case is satisfied. We had millions of dollars go in as deferred.

遞延很大程度取決於新的使用情境或新產品，而最終它會不再是新的使用情境或新產品。所以當某個使用情境被滿足時，我們會有數百萬美元從遞延中釋出。同時也會有數百萬美元以遞延的形式進來。

And generally, that number, if you look at it over the last few quarters or if you look at it over the entire year, it's trending up but deferred is never going to be infinite. It is going to come out and other new deferred will go in or the product becomes familiar and less new, right?

一般來說，如果你看過去幾季或看整年，它的趨勢是向上的，但遞延不可能無限增加。它會被認列出來，而其他新的遞延會再進來，或者產品變得更熟悉、不再那麼新，對吧？

So it's important to just look at this big picture and say, we got demand. We've got demand in AI, in enterprise, in campus, in data centers, and we're just in the best position we've been in a long, long time in my 18-year career.

所以重要的是從大局來看，並說：我們有需求。我們在AI、企業、校園（campus）、資料中心都有需求，而在我18年的職涯中，我們正處於很久以來最好的位置。

Over to you, Chantelle, you want to --

交給你了，Chantelle，你要——

The only thing I think it was very well said, and I think that's the right perspective to take on this, so we don't get too lost in the means of a quarter-to-quarter basis. We're very confident in the demand in the guide. Burgas a balance sheet account that will reflect whatever happens in the quarter. So that is what it is.

我覺得唯一要補充的是，你說得非常好，而且我認為這就是看待這件事的正確角度，這樣我們才不會在逐季的層面迷失在細節裡。我們對指引所反映的需求非常有信心。遞延收入是資產負債表科目，會反映當季發生的各種情況。所以它就是那樣。

The only thing I would add, like I said before, Tal, we did it do last quarter, was that we anticipate the end of this year from a deferred balance to be higher than we ended last year, which shows that the P&L did what the P&L needed to do without deferred. And so if that helps you feel more comfortable with that and that's the only thing I do with.

我唯一想補充的是，就像我之前說的，Tal——我們上季也提過——我們預期今年年底的遞延餘額會高於去年年底，這顯示損益表（P&L）在沒有依賴遞延的情況下，也完成了它該完成的表現。如果這能讓你更安心，那這就是我唯一要補充的。

Ryan Koontz, Needham & Company.

Ryan Koontz，Needham & Company。

I wanted to follow up on some of your scale-up commentary and ask about some of the technical design changes a little bit here with obviously, including more optics and including liquid cooling. And how will that change your fulfillment model? What things are you including in your sale versus partner to reference third parties?

我想延伸追問你們一些關於scale-up的評論，並請教一些技術設計上的變化：很明顯包含更多光學元件（optics）以及液冷（liquid cooling）。這將如何改變你們的履約／交付（fulfillment）模式？你們在銷售中會包含哪些項目，哪些會交由合作夥伴或轉介第三方？

And maybe you can expand a little bit on your product plans around scale up and how we should think about the product changes?

也許你可以再多談一點你們在 scale-up 方面的產品規劃，以及我們應該如何看待產品變化？

Yes. Actually, that's a really good question, Ryan. We're still figuring it out ourselves. Because Arista's value in the rack scale is still the switch. But however, we have to design the whole rack and then work with different partners differently on whether they want the whole rack from us or the rest gets shipped through our manufacturing partners, while we basically ascertain the switch.

是的。其實這是個非常好的問題，Ryan。我們自己也還在摸索。因為 Arista 在機櫃級（rack scale）的價值核心仍然是交換器。但同時，我們必須設計整個機櫃，並且依不同合作夥伴的需求，以不同方式合作：他們是希望整個機櫃都由我們提供，還是其餘部分透過我們的製造合作夥伴出貨，而我們基本上負責確認／提供交換器。

So each different stocks for different books because we've got three different designs going on with the orthogonal, the cable black plan and the cross rack designs, each one a little bit different. Give us a chance in the next few quarters to share with you what's going on with each one. There isn't one answer to your question.

所以不同的方案各有不同的配置，因為我們同時在推進三種不同設計：orthogonal、cable black plan，以及 cross rack 設計，每一種都略有差異。接下來幾個季度會給我們機會，與各位分享每一種設計的進展。你的問題沒有單一答案。

The one last thing I do want to add to that is, but what's very clear is there's a lot of switches simming and water. A lot of it's liquid cool.

最後我還想補充一點：很明顯有很多交換器都在「浸沒式與水冷」環境中運作。其中很大一部分是液冷。

Karl Ackerman, BNP Paribas.

Karl Ackerman，法國巴黎銀行（BNP Paribas）。

Chantelle or Jayshree, I'd like to ask a question that was similar to peers, but in a slightly different way. In the context of your 2027 revenue outlook, can you speak to whether your revenue outlook is predicated on new permitting or financing of power shell infrastructure to achieve these forecasts?

Chantelle 或 Jayshree，我想問一個和同業類似、但角度稍有不同的問題。在你們 2027 年營收展望的脈絡下，能否談談你們的營收展望是否建立在需要新增許可（permitting）或為電力殼層（power shell）基礎設施取得融資，才能達成這些預測？

And as you address that, can you speak to whether you require an aggressive drawdown of your deferred revenue balance to achieve your outlook? I think I know the answer, but we love your view.

另外，在回答這點時，也請談談你們是否需要大幅動用（drawdown）遞延收入（deferred revenue）餘額，才能達成你們的展望？我想我知道答案，但我們很想聽聽你們的看法。

What do you think our answer is?

你覺得我們的答案會是什麼？

Yes. It's so nice to sometimes ask a question ourselves.

是的，有時候自己問自己問題也挺不錯的。

I don't think you need to, but I would love clarification.

我認為你們不需要，但我希望能得到釐清。

Okay.

好的。

Yes, we're not -- we don't guide deferred. So I think that was the answer you were expecting to secondhand. I would say from the perspective of the guide next year, -- we are looking at what we know ahead of us, two quarters visibility taking a lien on the second half of the next year in the sense of what we've had conversations with some of our largest customers. the run rate we have in Cloud's business in a sense of enterprise and all that comes to other, but we don't guide deferred.

是的，我們不——我們不提供遞延收入的指引。所以我想這就是你預期的答案。我會說，就明年指引的角度來看——我們是根據我們已知的前方狀況來看，具備兩個季度的能見度，並且在某種程度上對明年下半年做出判斷，這是基於我們與一些最大客戶的對話。以及我們在雲端業務的 run rate（運行水準），包括企業端等其他因素，但我們不提供遞延收入指引。

This question was are we counting on data center.

這個問題是：我們是否仰賴資料中心（相關因素）。

The first part, the second part was are we preparing to draw down deferred to hit the number?

第一部分；第二部分是：我們是否準備透過動用遞延收入來達成數字？

Yes.

是的。

So we don't guide deferred, but I would say that we see a very strong demand signal and demand environment back to the momentum theme, but Jayshree, do you want to comment.

所以我們不提供遞延收入指引，但我會說我們看到非常強的需求訊號與需求環境，回到「動能」這個主題；不過 Jayshree，你要不要補充一下。

Right, exactly. So I think while we don't guide deferred, let me be clear that we do see plenty of new use cases and new products. 2027, like '26 is going to be filled with lots of new products.

對，完全正確。所以我想，雖然我們不提供遞延收入指引，但我想說清楚：我們確實看到很多新的使用情境與新產品。2027 年會像 2026 年一樣，充滿大量新產品。

Gianmarco Conti, Deutsche Bank.

Gianmarco Conti，德意志銀行（Deutsche Bank）。

Congrats on a full set of great numbers. As SU-144 moves towards production in '28, how should we think about gross margin economics relative to your existing switching portfolio. And how much customer commits do you already have behind those programs?

恭喜交出一整套很漂亮的數字。隨著 SU-144 在 28 年走向量產，我們應該如何看待其毛利率經濟性相對於你們既有交換器產品組合的表現？另外，在這些專案背後，你們目前已經有多少客戶承諾（commitments）？

You're talking about '27 or '28?

你是在問 27 年還是 28 年？

It's about gross in for --

是關於毛利——

Look, since we went public, our long-term gross margin has always been in the 60% to $65. I think we've tightened the rain now to more like 62% to 64%. We don't see any change.

你看，自從我們上市以來，我們的長期毛利率一直都在 60% 到 65% 之間。我想我們現在把區間收斂到大約 62% 到 64%。我們看不到任何變化。

Daniel Zhu, Bernstein.

Daniel Zhu，Bernstein。

Well, congrats on a great quarter. I just want to ask how concentrated the customer base will scale across. Should we think about this as really being driven by the cloud titans for synchronous training? Or are we also seeing telcos deployments even among smaller customers or for infant use cases?

恭喜這一季表現很棒。我想問的是，客戶基礎在 scale-across 方面會有多集中？我們應該把它視為主要由雲端巨頭為了同步訓練（synchronous training）所驅動嗎？還是我們也看到電信商（telcos）的部署，即使是在較小客戶或較早期（infant）的使用情境？

Yes. No, I think there are both small use cases and colos for scale across, which would be a smaller footprint switching and routing platforms like the 7280, but the concentration of our large customers also is high with our larger customers opting for the very large AI spine across data centers and requiring that kind of 576 port density and terabit scale and multi-gigawatt deployments for securities and performance.

是的。不，我認為兩者都有：既有較小的使用情境，也有 colo（共置機房）在做 scale-across，這通常會使用像 7280 這類較小規模的交換與路由平台；但我們的大型客戶集中度也確實很高，因為較大的客戶會選擇在資料中心之間建置非常大型的 AI spine（骨幹）互連，並且需要那種 576 埠密度、太位元（terabit）等級的規模，以及多 GW（gigawatt）等級的部署，以滿足可靠性與效能。

So we do see both. But I think the revenue concentration favors the 7800. The location concentration of multiple sites favors the 7280.

所以我們兩者都看得到。但我認為營收集中度更偏向 7800。而多站點的地點集中度則更偏向 7280。

Sean O'Loughlin, TD Cowen.

Sean O'Loughlin，TD Cowen。

Congrats on the fantastic results, ladies and gentlemen. To just check in broadly on where we are actually in the front-end cycle. I think we've spent so much time talking about scale out, scale across back-end deployments and the focus on 800 gig and above in those deployments that maybe we collectively have lost track of where we are in the sort of more traditional front-end compute.

各位女士先生，恭喜有非常出色的成果。我想更宏觀地確認一下：我們在前端（front-end）週期中實際處於什麼位置。我覺得我們花了太多時間在談 scale-out、scale-across 的後端部署，以及在那些部署中對 800G 及以上的聚焦，以至於我們可能集體忽略了較傳統的前端運算（front-end compute）目前的狀況。

I'm just wondering if you could comment on that and what you see in terms of market trends on the front-end specifically?

想請你們評論一下，特別是在前端方面，你們看到哪些市場趨勢？

Yes. No, Sean, it's a very good question. I think there are -- it's like a sine wave. When everybody is focused on the back end, then there's -- in those customers, they're thinking less about the front end. And then sometimes when they're focusing a lot on the front end, they're not worrying as much about the back.

是的。不，Sean，這是個非常好的問題。我認為這——有點像正弦波。當大家都把焦點放在後端時，那些客戶就會比較少去思考前端。而有時候當他們非常專注在前端時，就不會那麼擔心後端。

If you look at it over a period of time, like six quarters to two years, you can capture both -- but in a given year, we had one 10% customer, for example, put a lot of thought into the multi-hundred-gig refresh of their front end before enhancing their back end. And so that particular year, the front end is very high, but the back end was low.

如果你把時間拉長來看，例如六個季度到兩年，你就能同時涵蓋兩者——但在某一年裡，我們曾有一個占比 10% 的客戶，例如，在強化後端之前，先花了很多心力在前端的多百 G（multi-hundred-gig）更新換代上。因此在那一年，前端非常強，但後端相對較弱。

The following year may be the reverse. So if you tend to look at it over multiple quarters and years they even out. But if you go to look at it in a concentrated period of time, you'll usually see more of one than the other. Right now, in general, we're seeing more back end and AI fabrics, and I think the front end will catch up later.

隔一年可能就相反。所以如果你用多個季度與年度來觀察，它們會趨於平均。但如果你只看一段集中的期間，通常會看到其中一端更突出。目前整體而言，我們看到更多是後端與 AI fabrics（AI 網路織構），我認為前端之後會追上來。

Antoine Chkaiban, New Street Research.

Antoine Chkaiban，New Street Research。

Congratulations on a very strong set of results. Ken, maybe one for you, I want to ask about AVA with all the recent high-profile infrastructure security breaches that we've been hearing about. It looks like that technology can address a very large opportunity.

恭喜你們交出非常強勁的一組成果。Ken，也許這題問你：我想問 AVA。最近我們聽到不少高知名度的基礎設施資安事件。看起來這項技術可以切入一個非常大的機會。

So how should we think about sizing it? How do you plan to monetize AVA? And is that going to fall into products? Or could you drive actually a meaningful acceleration in your services revenue?

那麼我們應該如何思考其規模配置？你們打算如何將 AVA 變現？這會歸入產品收入嗎？或者你們是否能實際推動服務營收出現有意義的加速？

We are at early days in understanding all of that. We have just turned AVA on as an alpha feature for customer trial within our CloudVision service. And we are currently working on pricing models. I don't think I have any forecast or data to share with you at this time on that.

我們目前仍處於早期階段，正在理解這一切。我們剛剛在 CloudVision 服務中，將 AVA 作為 alpha 功能開啟，供客戶試用。我們目前也正在制定定價模型。我想我現在沒有任何預測或數據可以與你分享。

But I think it's very promising, Antoine, because back again to Ken's point that he repeatedly makes with me is the architecture of the US is so sound that adding AVO CloudVision our capability to look at deeper details and get more analysis starts with the foundation, right?

但我認為這非常有前景，Antoine，因為回到 Ken 一再對我強調的觀點：美國的架構非常穩健，因此在此基礎上加入 AVO CloudVision、我們深入查看更多細節並取得更多分析的能力，是從根基開始的，對吧？

So we feel very privileged that we can sit here and make a tell, well, oh my god, our vulnerabilities doubled from 30 to 67 and we fix them. And therefore, we can go apply AVA. Everybody else is dealing with a game of wakamol where they've got thousands of vulnerabilities and no AVA can possibly keep track of it. So I think we feel in a very good position to monetize AVA in the next few years.

所以我們覺得非常幸運，能坐在這裡說，天啊，我們的弱點從 30 個翻倍到 67 個，然後我們把它們修復了。因此，我們就可以去套用 AVA。其他人都在玩打地鼠的遊戲，他們有成千上萬個弱點，任何 AVA 都不可能追蹤得過來。所以我認為，我們在未來幾年將 AVA 變現方面處於非常有利的位置。

Ben Bollin, Cleveland Research.

Ben Bollin，Cleveland Research。

Jayshree and Todd, I suppose, I'm interested in your thoughts on where lead times are across the portfolio if we kind of carve it up into enterprise-centric and hyperscale centric? Just what those look like today?

Jayshree 和 Todd，我想我有興趣聽聽你們對於整體產品組合交期的看法，如果我們把它切分為以企業為主與以超大規模客戶為主兩部分。目前看起來各自的交期是什麼狀況？

And I was hoping, Todd could talk maybe a little bit more big picture about the go-to-market effort, the execution within enterprise share capture and how that's building and where that goes from here?

另外我也希望 Todd 能從更宏觀的角度，多談一些 go-to-market（市場進入）推進工作、在企業市占擴張上的執行狀況，這些如何建立起來，以及接下來會往哪裡走？

Yes, on the lead times, and it goes both for the component lead times on the finished good lead times. In the enterprise side of our business, we tend to have shorter lead times, somewhat shorter. And that's driven by the fact that those switching model is in the absolute cutting edge. They're not leading brand-new technologies requiring the highest process, silicon, et cetera.

是的，關於交期，這同時涵蓋元件交期與成品交期。在我們企業端業務方面，交期通常較短，算是相對短一些。這主要是因為那些交換器型號並非處於最尖端。它們不需要導入全新技術、也不需要最高階製程的矽晶片等。

And we're able to -- they also tend to be at a lower price point, we're able to hold a little bit more buffered inventory there. So we tend to have lower component lead times there and lower lead times to our customers on the AI side, on the ultra-advanced hyperscale data center lead times tend to be a little bit longer.

而且它們通常也在較低的價格帶，我們能在那裡保有稍多一些的緩衝庫存。因此在那一端，我們的元件交期較低、交付給客戶的交期也較短；而在 AI 端、在超先進的超大規模資料中心方面，交期往往會稍微長一些。

As far as the buildout on the go-to-market side, we were really -- we have been just really pleased with the momentum we've seen this year and the herd that's being built into the whole go-to-market on the kind of thought leadership side and the awareness side, we saw great momentum on the -- from webinars and from events.

至於 go-to-market 端的建置，我們真的——我們對今年看到的動能感到非常滿意；在整體 go-to-market 的思想領導與市場認知層面所形成的聲量與聚集效應，我們從網路研討會與活動中看到了很好的動能。

In fact, we have our largest customer event coming up in December, which we're very excited about. And on the sales execution side, it's been just phenomenal efficiency from that team, and we're looking forward to continuing to build out next year.

事實上，我們在 12 月將舉辦規模最大的客戶活動，我們對此非常期待。而在銷售執行面，該團隊展現了極高的效率，我們也期待明年持續擴大建置。

I just want to add, Todd, also I see that you and Chris Schmidt, Postea, you really heightened the investment not only in the United States, but with key verticals and international as well. So I think all of the go-to-market investments we've been making over the last year and in fact, the last three years is really paying off.

我也想補充一下，Todd，我看到你和 Chris Schmidt、Postea，你們不僅在美國，也在關鍵垂直領域以及國際市場大幅提高了投資。所以我認為，我們過去一年、甚至過去三年所做的所有 go-to-market 投資，確實正在開花結果。

Absolutely.

完全同意。

Regina, we have time for one last question.

Regina，我們還有時間再問最後一個問題。

Mehdi Hosseini, Susquehanna.

Mehdi Hosseini，Susquehanna。

I have two quick follow-ups. Going back to your revenue guide for '27 is $8.5 billion. Should we assume that all the sell have been planned for?

我有兩個簡短的追問。回到你們對 2027 年的營收指引是 85 億美元。我們是否應該假設所有的矽晶片都已經規劃到位？

And then the second question, if you could elaborate on the mix between cloud and enterprise. I just want to get a better understanding how enterprises are investing as it relates to hyperscalers.

第二個問題，如果可以的話，請你們進一步說明雲端與企業之間的組合占比。我想更清楚了解企業在投資上相對於超大規模雲端業者的情況。

Sorry, you asked is all the silicon planned for already in the $18.5 billion.

抱歉，你問的是在 185 億美元裡，所有的矽晶片是否都已經規劃好了嗎？

Should we -- since you are providing additional assessment on 2027 revenue target, should we assume that you have already planned with your suppliers and selling in secure for both compute and memory? And there are some that argue that ties is going to get worse in '27.

也就是說，既然你們提供了對 2027 年營收目標的進一步評估，我們是否應該假設你們已經與供應商完成規劃並確保供應，涵蓋運算與記憶體？而且有人認為 2027 年的供應緊張狀況會更嚴重。

I just want to make sure that there is no downside to this target as it relates to inventory management.

我只是想確認，就庫存管理而言，這個目標不會有下行風險。

Understand, understand. Yes. We work very, very closely. We're sort of proud of our strategic partnerships, and we've been investing heavily in long-term strategic agreements in order to secure supply, not just in silicon, but PCB capacity memory, which Chantelle mentioned earlier.

了解，了解。是的。我們與合作夥伴非常、非常緊密地合作。我們對我們的策略夥伴關係感到自豪，並且一直在大力投資長期策略協議，以確保供應，不僅是矽晶片，也包括 PCB 產能與記憶體，Chantelle 先前也提到過。

So we're incredibly -- we have very high confidence in the ability to deliver supply to the forecast. Of course, we would not want to place everything to five quarters lead time for all of next year. So not every order is placed.

因此我們對於能夠依照預測交付供應，具有非常高的信心。當然，我們不會希望為了明年所有需求，把所有事情都以五個季度的交期一次性下單。所以並不是每一筆訂單都已經下了。

This concludes Arista Networks' third-quarter 2026 earnings call. We have posted presentations that provide additional information on our results, which you can access on the Investors section of our website.

以上為 Arista Networks 2026 年第三季財報電話會議。我們已發布簡報，提供更多關於本次業績的補充資訊，你可以在我們網站的投資人（Investors）專區取得。

Thank you for joining us today and for your interest in Arista.

感謝各位今天的參與，以及對 Arista 的關注。

Thank you for joining, ladies and gentlemen. This concludes today's call, and you may now disconnect.

各位女士、先生，感謝加入。今天的電話會議到此結束，您現在可以掛線。