F5 Inc (FFIV) 2025 Q3 法說會逐字稿

Good afternoon and welcome to F5 Inc, third quarter fiscal 2025 financial results conference call. (Operator Instructions) Also, today's conference call is being recorded. If anyone has any objections, please disconnect at this time.

下午好，歡迎參加 F5 Inc 2025 財年第三季財務業績電話會議。（操作員指示）此外，今天的電話會議正在錄音。如果有人有異議，請立即斷開連接。

I will now turn the call over to your host, Ms Suzanne DuLong. Thank you. You may begin.

現在我將電話轉給主持人 Suzanne DuLong 女士。謝謝。你可以開始了。

Hello, and welcome. I'm Suzanne DuLong, F5's Vice President of Investor Relations. We are here with you today to discuss our third quarter fiscal year 2025 financial results. Francois Locoh-Donou, F5's President and CEO; and Cooper Werner, F5's Executive Vice President and CFO, will be making prepared remarks on today's call.

您好，歡迎光臨。我是 F5 投資人關係副總裁 Suzanne DuLong。今天我們在這裡與您討論 2025 財年第三季的財務表現。F5 總裁兼執行長 Francois Locoh-Donou 和 F5 執行副總裁兼財務長 Cooper Werner 將在今天的電話會議上發表準備好的演講。

Other members of the F5 executive team are also here to answer questions during the Q&A session. A copy of today's press release is available on our website at f5.com, where an archived version of today's audio will be available through October 27, 2025.

F5 執行團隊的其他成員也在現場，在問答環節解答問題。今天的新聞稿副本可在我們的網站 f5.com 上查閱，今天的音訊存檔版本將保留至 2025 年 10 月 27 日。

We will post the slide deck accompanying today's webcast to our IR site at the conclusion of today's call. To access the replay of today's webcast by phone 877-660-6853 and or 201-612-7415 and use meeting ID 13754228.

我們將在今天的電話會議結束時將今天的網路廣播的幻燈片發佈到我們的 IR 網站上。若要觀看今天網路廣播的重播，請撥打電話 877-660-6853 或 201-612-7415，並使用會議 ID 13754228。

The telephonic replay will be available through midnight Pacific Time, July 31, 2025. For additional information or follow-up questions, please reach out to me directly at s.dulong@f5.com.

電話回放將持續到 2025 年 7 月 31 日太平洋時間午夜。如需更多資訊或後續問題，請直接透過 s.dulong@f5.com 與我聯絡。

Our discussion today will contain forward-looking statements, which include words such as believe, anticipate, expect and target. These forward-looking statements involve uncertainties and risks that may cause our actual results to differ materially from those expressed or implied by these statements.

我們今天的討論將包含前瞻性陳述，其中包括相信、預期、期望和目標等詞語。這些前瞻性陳述涉及不確定性和風險，可能導致我們的實際結果與這些陳述表達或暗示的結果有重大差異。

We've summarized factors that may affect our results in the press release announcing our financial results and in detail in our SEC filings. In addition, we will be referencing non-GAAP metrics during today's discussion.

我們在宣布財務表現的新聞稿中總結了可能影響我們業績的因素，並在提交給美國證券交易委員會的文件中進行了詳細說明。此外，我們將在今天的討論中參考非 GAAP 指標。

Please see our full GAAP to non-GAAP reconciliation in today's press release and in the appendix of our earnings slide deck. Please note that F5 has no duty to update any information presented in this call. During today's call, Francois will speak to our Q3 highlights and our strategy and growth opportunities. Cooper will then review the details of our Q3 results and our outlook.

請參閱今天的新聞稿和收益幻燈片附錄中的完整 GAAP 與非 GAAP 對帳表。請注意，F5 沒有義務更新本次通話中提供的任何資訊。在今天的電話會議中，弗朗索瓦將談論我們第三季的亮點以及我們的策略和成長機會。庫柏隨後將審查我們第三季業績的詳細資訊和展望。

I will now turn the call over to Francois.

現在我將電話轉給弗朗索瓦。

Thank you, Suzanne, and hello, everyone. Our exceptional Q3 results highlight the strength of our business and F5's strong alignment with important secular trends. customers are modernizing their data centers, adopting hybrid multi-cloud architectures and scaling to meet growing application capacity and performance needs.

謝謝你，蘇珊娜，大家好。我們出色的第三季業績凸顯了我們業務的實力以及 F5 與重要長期趨勢的緊密結合。客戶正在對其資料中心進行現代化改造，採用混合多雲架構並進行擴展以滿足不斷增長的應用程式容量和效能需求。

Our Q3 results demonstrate F5's position at the forefront of these transformative shifts. We delivered 12% total revenue growth, including 26% growth in product revenue our strongest in 14 years. This performance is a testament to our team's execution, our continued innovation and the enormous trust the largest enterprises and service providers across the globe place in F5.

我們的第三季業績證明了 F5 在這些變革性轉變中處於領先地位。我們實現了總收入成長 12%，其中產品收入成長 26%，為 14 年來的最高水準。這項業績證明了我們團隊的執行力、我們持續的創新以及全球最大的企業和服務提供者對 F5 的巨大信任。

F5's unique ability to deliver and secure every app, every API anywhere on premises, in the cloud, at the edge and across hybrid multi-cloud environments is a significant competitive advantage. It powered strong demand across both hardware and software deployment models in Q3.

F5 具有在本地、雲端、邊緣和混合多雲環境中交付和保護每個應用程式、每個 API 的獨特能力，這是其顯著的競爭優勢。它推動了第三季硬體和軟體部署模型的強勁需求。

Our systems revenue grew 39%, fueled by data center modernization, increased capacity requirements and adoption of our latest generation of hardware. Our software revenue grew 16%, driven by hybrid multi-cloud architecture adoption which is contributing to continued strong software subscription renewals and expansions.

我們的系統收入成長了 39%，這得益於資料中心的現代化、容量需求的增加以及我們採用最新一代硬體。我們的軟體收入成長了 16%，這得益於混合多雲架構的採用，這有助於軟體訂閱續訂和擴展的持續強勁成長。

We continue to operate the business with discipline, combined with our strong revenue performance and continued operating margin leverage, we delivered exceptional EPS results in the quarter, reflecting 24% growth year-over-year. Looking ahead, our Q4 pipeline reflects continued strong demand in support of future forward architectures, data center modernization and an increasing range of application delivery and security services.

我們繼續嚴格遵守紀律地經營業務，加上強勁的收入表現和持續的營業利潤率槓桿，我們在本季度取得了出色的每股收益業績，同比增長 24%。展望未來，我們的第四季度產品線反映了對未來前沿架構、資料中心現代化以及不斷增加的應用程式交付和安全服務範圍的持續強勁需求。

As a result, we expect Q4 revenue in a range of $780 million to $800 million, implying approximately 9% revenue growth for FY25. Cooper will provide a detailed overview of our Q3 performance and our outlook shortly.

因此，我們預計第四季度營收將在 7.8 億美元至 8 億美元之間，這意味著 2025 財年營收將成長約 9%。Cooper 很快就會提供我們第三季業績和展望的詳細概述。

Before that, I will highlight early traction with our F5 application delivery and security platform. I will also recap recent centric innovation and speak to some of our latest wins supporting AI workloads and infrastructure. We expect the long-term structural shifts driving data center modernization will persist continuing to reshape how companies invest in their IT infrastructure.

在此之前，我將重點介紹我們的 F5 應用程式交付和安全平台的早期發展。我還將回顧最近的中心創新，並談談我們在支援人工智慧工作負載和基礎設施方面取得的一些最新成果。我們預計，推動資料中心現代化的長期結構性轉變將持續下去，繼續重塑企業對 IT 基礎設施的投資方式。

Managing a skyrocketing number of apps and APIs across increasingly distributed environment is creating enormous challenges for IT teams and preparing for and implementing AI is only making it worse. At F5, we have made it our mission to dramatically simplify this complex array of operational and cybersecurity challenges with the F5 application delivery and security platform.

在日益分散的環境中管理數量激增的應用程式和 API 給 IT 團隊帶來了巨大的挑戰，而準備和實施 AI 只會讓情況變得更糟。在 F5，我們的使命是透過 F5 應用程式交付和安全平台大幅簡化一系列複雜的營運和網路安全挑戰。

While there are platforms for endpoints for network access and for cloud workloads, before the F5 ADSP, there was no platform that fully converges high-performance traffic management with advanced application and API security capabilities across hybrid and multi-cloud environment at scale.

雖然存在用於網路存取和雲端工作負載的端點平台，但在 F5 ADSP 之前，還沒有一個平台能夠在混合和多雲環境中大規模地將高效能流量管理與先進的應用程式和 API 安全功能完全融合。

The F5 ADSP is unique in that it enables customers to say, consolidate multiple delivery and security on solutions for applications and APIs on F5, leveraging a single platform with best-in-class capabilities. B, consistently deploy delivery and security across their hybrid multi-cloud environment by hardware, software and SaaS form factors.

F5 ADSP 的獨特之處在於，它使客戶能夠利用具有一流功能的單一平台，整合 F5 上應用程式和 API 的多種交付和安全解決方案。B、透過硬體、軟體和 SaaS 形式在其混合多雲環境中一致部署交付和安全。

And C, dramatically simplify operations and reduce manual efforts by leveraging AI-powered analytics, insights and policy management. With these powerful benefits, it's no surprise that initial customer response is very positive.

C、透過利用人工智慧分析、洞察和策略管理，大幅簡化操作並減少人工工作量。憑藉這些強大的優勢，最初的客戶反應非常積極也就不足為奇了。

In Q3, we won a number of deals that demonstrate the power and the benefit of F5's platform approach. For example, North American health insurer chose F5 distributed cloud services and big IP, consolidating multiple security and delivery vendors across both its on-premises and SaaS environment, leveraging our unique hybrid multi-cloud approach, F5 is significantly reducing operational complexity while also providing enhanced security and improved API visibility.

在第三季度，我們贏得了許多交易，證明了 F5 平台方法的強大功能和優勢。例如，北美健康保險公司選擇了 F5 分散式雲端服務和大 IP，在其本地和 SaaS 環境中整合了多個安全和交付供應商，利用我們獨特的混合多雲方法，F5 顯著降低了營運複雜性，同時還提供了增強的安全性和改進的 API 可見性。

In another example, an airline in our EMEA region selected F5 to modernize its application delivery and security posture. Already an F5 BIG-IP customer they faced escalating application growth and challenges securing a rapidly growing base of APIs across multi-cloud environment.

另一個例子是，我們 EMEA 地區的一家航空公司選擇 F5 來實現其應用交付和安全態勢的現代化。作為 F5 BIG-IP 客戶，他們面臨著不斷增長的應用程式成長以及在多雲環境中保護快速增長的 API 基礎的挑戰。

After extensive competitive evaluation, they consolidated both their on-premises and SaaS environments on F5, replacing their incumbent SaaS provider and deploying F5 distributed cloud services with comprehensive WAP and secure multi-cloud networking. F5 is enabling the airline to securely connect and manage more than 100 B2B applications across multi-cloud environment, ensuring scalability, security and operational efficiency.

經過廣泛的競爭評估後，他們在 F5 上整合了他們的內部部署和 SaaS 環境，取代了他們現有的 SaaS 供應商，並部署了具有全面 WAP 和安全多雲網路的 F5 分散式雲端服務。F5 使該航空公司能夠在多雲環境中安全地連接和管理 100 多個 B2B 應用程序，確保可擴展性、安全性和營運效率。

In a final example, a South American service provider consolidated all of its multi-cloud and on-premise security services on F5 through discovery during the sales process of what was initially a WAF-only opportunity, our team learned the full scale of the customers challenges. The customer embraced the benefits of a unified F5 approach, consolidating their WAF bot protection and API security on F5 distributed card services.

最後一個例子是，一家南美服務提供者透過在最初僅限 WAF 的機會的銷售過程中的發現，將其所有多雲和內部安全服務整合到 F5 上，我們的團隊全面了解了客戶面臨的挑戰。客戶接受了統一 F5 方法的好處，在 F5 分散式卡片服務上整合了他們的 WAF 機器人保護和 API 安全。

F5 is streamlining their operations, eliminating multi-vendor complexity and reducing cost by consolidating across all of its on-premises and multi-cloud environment. These examples underscore the significant value F5 delivers to customers through our ADSP.

F5 正在透過整合其所有內部部署和多雲環境來簡化其運營，消除多供應商複雜性並降低成本。這些例子強調了 F5 透過我們的 ADSP 為客戶提供的巨大價值。

While there is still work ahead to unlock the full potential of the platform, we are confident that our ongoing innovation will enable us to deliver even greater benefits to our customers. Over recent quarters, we have been unveiling a suite of groundbreaking customer-centric innovations that showcase the power of the F5 ADSP while also highlighting how we are leveraging advanced AI technologies to enhance customer experiences and drive business growth.

儘管要充分發揮平台的潛力還有很多工作要做，但我們相信，我們不斷的創新將使我們能夠為客戶帶來更大的利益。在最近幾個季度，我們推出了一系列突破性的以客戶為中心的創新，展示了 F5 ADSP 的強大功能，同時也強調了我們如何利用先進的人工智慧技術來增強客戶體驗並推動業務成長。

These innovations fall into two categories. The first is AI for ADC. These are innovations that leverage AI to reduce the operational complexity of delivering and securing applications in a hybrid multi-cloud world. These innovations empower customers to unlock the full potential of F5, simplifying deployment and accelerating their ability to scale F5 solutions seamlessly.

這些創新分為兩類。第一個是針對ADC的AI。這些創新利用人工智慧來降低在混合多雲世界中交付和保護應用程式的操作複雜性。這些創新使客戶能夠充分發揮 F5 的潛力，簡化部署並加速無縫擴展 F5 解決方案的能力。

The second is ADC for AI. These innovations highlight how we are applying our strength in delivery and security to enable AI-driven applications. During Q3, we continued to make advances in both of these categories.

第二個是針對AI的ADC。這些創新凸顯了我們如何利用我們在交付和安全方面的優勢來實現人工智慧驅動的應用程式。在第三季度，我們在這兩個類別中繼續取得進展。

In AI for ADC, we introduced advancements to our F5 AI assistant. Our goal is to make the F5 AI assistant an indispensable member of every NetOps and SecOps team.

在 ADC 的 AI 中，我們引入了 F5 AI 助理的進步。我們的目標是讓 F5 AI 助理成為每個 NetOps 和 SecOps 團隊不可或缺的成員。

We are systematically building it around a simple, powerful loop, understand what's happening, prioritize what matters and act to solve it. To dramatically improve our customers ability to understand potential threats and streamline prioritization, we are integrating technology from our Fletch acquisition.

我們正在圍繞一個簡單而強大的循環系統地構建它，了解正在發生的事情，優先考慮重要的事情並採取行動來解決它。為了大幅提高客戶了解潛在威脅和簡化優先順序的能力，我們正在整合收購 Fletch 的技術。

This integration will enable the F5 AI assistant to provide a real-time contextualized zoo of the threat landscape, along with proactive actionable recommendations that help cut through the noise. But insight without action is incomplete.

透過此次整合，F5 AI 助理將能夠提供即時的威脅情勢情境化動物園，以及有助於消除噪音的主動可行建議。但沒有行動的洞察力是不完整的。

That's why we enabled the assistant to act starting with its new ability to generate eye rules from a simple pump. This translates customer intent into secure performance code, making it easier for more customers to harness and leverage F5's unmatched programmability.

這就是為什麼我們讓助手從其透過簡單的幫浦產生眼球規則的新功能開始行動。這將客戶意圖轉化為安全的效能程式碼，使更多客戶更容易利用和利用 F5 無與倫比的可程式性。

This complete loop from deep understanding to automated action is how we move our customers from reactive firefighting to proactive control. In ADC for AI, we announced expanded capabilities for our F5 AI gateway to prevent data leaks and deliver cutting-edge AI data protection.

從深刻理解到自動化行動的完整循環就是我們如何讓客戶從被動救火轉向主動控制。在 AI ADC 中，我們宣布擴展 F5 AI 網關的功能，以防止資料外洩並提供尖端的 AI 資料保護。

Additionally, we introduced new functionality for F5 BIG-IP SSL Orchestrator to classify and defend encrypted data in motion and block unapproved AI use. Through industry-leading technology partnerships, F5 is also providing integrated secure and streamlined solutions to support complex emerging AI ecosystem.

此外，我們還為 F5 BIG-IP SSL Orchestrator 引入了新功能，用於對動態加密資料進行分類和保護，並阻止未經批准的 AI 使用。透過業界領先的技術合作夥伴關係，F5 還提供整合的安全且精簡的解決方案，以支援複雜的新興 AI 生態系統。

We recently announced advances with two of our AI-focused partners. In June, building on our prior work with NVIDIA, we announced we are expanding F5's performance, multi-tenancy and security capabilities for large-scale AI infrastructures. We are leveraging BIG-IP next for Kubernetes running natively on NVIDIA BlueField-3 DPUs.

我們最近宣布與兩家專注於人工智慧的合作夥伴取得了進展。今年 6 月，在先前與 NVIDIA 合作的基礎上，我們宣布將擴展 F5 針對大規模 AI 基礎設施的效能、多租戶和安全功能。接下來，我們將利用 BIG-IP 來實現在 NVIDIA BlueField-3 DPU 上原生運行的 Kubernetes。

Feedback from customers who have tested the solution has been very promising. One customer reported an 18% increase in HTTP throughput and 11x improvement in time to first bite and 190x boost in network energy efficiency.

測試過該解決方案的客戶的回饋非常積極。一位客戶報告稱，HTTP 吞吐量增加了 18%，首次使用時間提高了 11 倍，網路能源效率提高了 190 倍。

Another customer reported a 20% improvement in GPU utilization. We believe this level of performance improvement could be meaningful to customers looking to scale and enhance performance of their AI infrastructure and we continue to work toward validating the solution with customers and NVIDIA.

另一位客戶則報告 GPU 利用率提高了 20%。我們相信，這種程度的效能提升對於希望擴展和增強其 AI 基礎架構效能的客戶來說意義重大，我們將繼續努力與客戶和 NVIDIA 一起驗證解決方案。

With customers leveraging F5 to move incredible amounts of data at high speed, to and from data stores for AI modeling and inferencing, our work with storage partners is also critical. In July, we announced that we are deepening our alliance with (inaudible) advancing our AI application and data delivery solution to manage training and inference-driven data growth.

隨著客戶利用 F5 以高速在資料儲存之間移動大量資料以進行 AI 建模和推理，我們與儲存合作夥伴的合作也至關重要。7 月，我們宣布將深化與 (聽不清楚) 的聯盟，推進我們的人工智慧應用和資料傳輸解決方案，以管理訓練和推理驅動的資料成長。

These innovations and partnerships highlight F5's role in enabling AI-driven applications. Today, dozens of leading enterprises are deploying F5 to ensure AI workflows operate flawlessly and scale seamlessly and remain resilient against evolving threats.

這些創新和合作凸顯了 F5 在實現 AI 驅動應用方面的作用。如今，數十家領先企業正在部署 F5，以確保 AI 工作流程完美運作、無縫擴展並保持抵禦不斷演變的威脅的能力。

These AI use cases represent net new insertion points for F5 and leverage technology and expertise built over a decade. Today, customers are leveraging F5 across three primary AI-related use cases.

這些 AI 用例代表了 F5 的全新插入點，並利用了十多年來積累的技術和專業知識。如今，客戶正在三個主要的 AI 相關用例中利用 F5。

The first is AI data delivery. In these use cases, customers are deploying F5 in front of data stores to ensure secure high throughput data ingestion for AI model training and inferencing. F5 is enforcing policy-based controls and eliminating bottlenecks, enabling performance delivery of massive data sets, while safeguarding sensitive information.

第一是AI資料傳輸。在這些用例中，客戶在資料儲存前部署 F5，以確保為 AI 模型訓練和推理提供安全的高吞吐量資料擷取。F5 正在實施基於策略的控制並消除瓶頸，從而實現海量資料集的效能交付，同時保護敏感資訊。

The second is AI runtime security. In these use cases, F5 is protecting AI applications, APIs and models with F5 WAP and AI Gateway. F5 prevents abuse, data leakage and attacks like pump injection while ensuring full visibility and control over how AI models and applications interact across environment.

第二是AI運作時安全。在這些用例中，F5 使用 F5 WAP 和 AI Gateway 保護 AI 應用程式、API 和模型。F5 可防止濫用、資料外洩和泵浦注入等攻擊，同時確保對 AI 模型和應用程式在整個環境中的互動方式進行完全可見性和控制。

The third is AI factory load balancing. In these cases, F5 is optimizing traffic both across and within AI factories. In AI, performance is measured in terms of token throughput by intelligently distributing AI traffic to maximize GPU utilization, F5 AI factory load balancing directly increases token generation, reduces time to first token and lowers cost per token.

第三是AI工廠負載平衡。在這些情況下，F5 正在優化 AI 工廠內部和跨工廠的流量。在人工智慧領域，效能以令牌吞吐量來衡量，透過智慧分配人工智慧流量來最大化 GPU 利用率，F5 人工智慧工廠負載平衡直接增加令牌生成，減少第一個令牌的時間並降低每個令牌的成本。

In Q3, we secured several new AI wins. In an AI data delivery use case, a government institution in our EMEA region needed to dramatically accelerate data ingestion for its AI platform, but its existing infrastructure could not meet the throughput and traffic modification demand.

在第三季度，我們獲得了幾項新的 AI 勝利。在 AI 資料傳輸用例中，我們 EMEA 地區的政府機構需要大幅加快其 AI 平台的資料提取速度，但其現有基礎設施無法滿足吞吐量和流量修改需求。

The core challenge was processing data at 100 gigabits per second while performing intensive SSL decryption and inserting critical data for audit trails. To solve this, they placed high-performance F5 BIG-IP hardware in front of their S3 data stores.

核心挑戰是以每秒 100 千兆位元的速度處理數據，同時執行密集的 SSL 解密並插入用於審計追蹤的關鍵數據。為了解決這個問題，他們在 S3 資料儲存前面放置了高效能 F5 BIG-IP 硬體。

BIG-IP is able to perform SSL decryption at line rate while preserving critical data for auditing and policy enforcement. This combination of hardware-based decryption, intelligent traffic modification and robust traffic management allowed the institution to meet its demanding performance goals and build a scalable foundation for future AI workloads.

BIG-IP 能夠以線速執行 SSL 解密，同時保留用於稽核和策略執行的關鍵資料。這種基於硬體的解密、智慧流量修改和強大的流量管理的結合使該機構能夠滿足其苛刻的效能目標，並為未來的人工智慧工作負載建立可擴展的基礎。

In an AI factory load balancing use case, an e-commerce provider based in our APAC region strategically deployed F5 BIG-IP and NGINX in key layers of their GPU as a service cloud platform. They use high-performance BIG-IP hard work to manage and secure the main entry point to their infrastructure, while F5 NGINX is embedded directly in front of each AI cluster to enforce fine grain authentication and authorization.

在 AI 工廠負載平衡用例中，我們亞太地區的一家電子商務供應商在其 GPU 即服務雲端平台的關鍵層策略性地部署了 F5 BIG-IP 和 NGINX。他們使用高效能 BIG-IP 硬體來管理和保護其基礎設施的主要入口點，同時 F5 NGINX 直接嵌入在每個 AI 叢集前面，以實施細粒度的身份驗證和授權。

This ensures that as their cloud scale, performance and security scale with it, enabling them to provide a foundational AI resource for their customers. Collectively, these successes underscore F5's expanding leadership in the hybrid multi-cloud landscape and the tangible value our platform approach delivered to customers and powering them to simplify operations, enhance security and accelerate innovation across their environment.

這確保了他們的雲端規模、效能和安全性隨之擴展，使他們能夠為客戶提供基礎的人工智慧資源。總的來說，這些成功凸顯了 F5 在混合多雲領域不斷擴大的領導地位，以及我們的平台方法為客戶帶來的實際價值，幫助他們簡化營運、增強安全性並加速整個環境中的創新。

Now I will turn the call to Cooper to elaborate on our Q3 results and our Q4 outlook. Cooper?

現在我將把電話轉給 Cooper，詳細說明我們的第三季業績和第四季展望。庫柏？

Thank you, Francois, and hello, everyone. I will review our Q3 results before I elaborate on our Q4 and FY25 outlook. As Francois noted, we delivered a strong Q3, growing revenue 12% to $780 million, reflecting a mix of 50% global services revenue and 50% product revenue.

謝謝你，弗朗索瓦，大家好。在詳細闡述我們第四季和 2025 財年的展望之前，我將回顧我們第三季的業績。正如弗朗索瓦所指出的，我們第三季業績表現強勁，營收成長 12%，達到 7.8 億美元，其中全球服務收入和產品收入分別佔 50% 和 50%。

Global Services revenue of $392 million grew 1%, while product revenue of $389 million grew 26% year-over-year. Our software revenue grew 16% year-over-year to $208 million as customers continue to expand consumption and adopt additional capabilities across our platform.

全球服務收入 3.92 億美元，年增 1%，產品收入 3.89 億美元，年增 26%。隨著客戶不斷擴大消費並採用我們平台上的更多功能，我們的軟體收入年增 16%，達到 2.08 億美元。

Our subscription-based software revenue grew 19% year-over-year to $185 million, representing 89% of our total software revenue. Perpetual license software totaled $23 million, down slightly year-over-year.

我們的訂閱軟體營收年增 19%，達到 1.85 億美元，占我們總軟體收入的 89%。永久許可軟體總額為 2,300 萬美元，年比略有下降。

Systems revenue totaled $181 million, up 39% year-over-year, with strength driven by tech refresh and data center modernization, including customers preparing for AI as well as competitive takeout's. Revenue from recurring sources contributed 73% of our Q3 revenue.

系統收入總計 1.81 億美元，年增 39%，主要得益於技術更新和資料中心現代化，包括為人工智慧做準備的客戶以及競爭性外送。經常性收入占我們第三季營收的 73%。

Our recurring revenue consists of our subscription-based revenue and the maintenance portion of our global services revenue. Shifting to revenue distribution by region.

我們的經常性收入包括訂閱制的收入和全球服務收入的維護部分。轉向按地區分配收入。

Our teams drove growth across all theaters. Revenue from the Americas grew 13% year-over-year, representing 55% of total revenue. EMEA delivered 6% growth, representing 26% of revenue and APAC grew 21%, representing 19% of revenue.

我們的團隊推動了所有劇院的成長。來自美洲的營收年增13%，佔總營收的55%。歐洲、中東和非洲地區成長 6%，佔營收的 26%，亞太地區成長 21%，佔營收的 19%。

Looking at our major verticals. Enterprise customers represented 70% of Q3's product bookings. Government customers represented 15% of product bookings, including 5% from US Federal. Finally, service providers represented 15% of Q3 product bookings.

看看我們的主要垂直行業。企業客戶佔第三季產品預訂量的 70%。政府客戶佔產品預訂量的 15%，其中美國聯邦政府佔 5%。最後，服務提供者佔第三季產品預訂量的 15%。

Our continued operating discipline contributed to our strong Q3 operating results. GAAP gross margin was 81%. Non-GAAP gross margin was 83.1%. Our GAAP operating expenses were $435 million. Our non-GAAP operating expenses were $381 million.

我們持續的經營紀律促成了我們第三季強勁的經營業績。GAAP毛利率為81%。非公認會計準則毛利率為83.1%。我們的 GAAP 營運費用為 4.35 億美元。我們的非公認會計準則營運費用為 3.81 億美元。

Our GAAP operating margin was 25.2%. Our non-GAAP operating margin was 34.3%, an improvement of more than 80 basis points year-over-year. Our GAAP effective tax rate for the quarter was 10.8%. Our non-GAAP effective tax rate was 14.4%.

我們的 GAAP 營業利潤率為 25.2%。我們的非公認會計準則營業利潤率為 34.3%，比去年同期提高了 80 多個基點。本季我們的 GAAP 有效稅率為 10.8%。我們的非公認會計準則有效稅率為 14.4%。

This includes a discrete benefit associated with the filing of our annual federal income tax return during the quarter. Our GAAP net income for the quarter was $190 million, or $3.25 per share. Our non-GAAP net income was $243 million or $4.16 per share, reflecting 24% EPS growth from the year ago period.

這包括與本季提交年度聯邦所得稅申報表相關的單獨福利。本季我們的 GAAP 淨收入為 1.9 億美元，即每股 3.25 美元。我們的非公認會計準則淨收入為 2.43 億美元，即每股 4.16 美元，較去年同期每股收益成長 24%。

I will now turn to cash flow and balance sheet metrics, all of which were very strong. We generated a record $282 million in cash flow from operations in Q3. CapEx was $9 million. DSO for the quarter was 42 days. Cash and investments totaled approximately $1.44 billion at quarter end.

現在我將談談現金流和資產負債表指標，它們都非常強勁。我們在第三季創造了創紀錄的 2.82 億美元營運現金流。資本支出為 900 萬美元。 DSO本季的休假天數為 42 天。季度末現金和投資總額約為 14.4 億美元。

Deferred revenue was $1.96 billion, up 10% from the year ago period. In Q3, we repurchased $125 million worth of F5 shares at an average price of $256 per share. As of the end of Q3, we had $1 billion remaining on our authorized stock repurchase program.

遞延營收為 19.6 億美元，較去年同期成長 10%。第三季度，我們以平均每股 256 美元的價格回購了價值 1.25 億美元的 F5 股票。截至第三季末，我們授權的股票回購計畫還剩餘 10 億美元。

Year-to-date, we've repurchased shares equivalent to 52% of our annual free cash flow. Finally, we ended the quarter with approximately 6,540 employees.

年初至今，我們已回購了相當於年度自由現金流 52% 的股票。最後，本季結束時我們擁有約 6,540 名員工。

I will now speak to our Q4 outlook. With the exception of revenue, my guidance comments reference non-GAAP metrics. As Francois noted, visibility into our Q4 pipeline and customer demand remains strong.

我現在將談談我們第四季的展望。除收入外，我的指導意見參考非 GAAP 指標。正如弗朗索瓦所指出的，我們第四季的銷售通路和客戶需求的可見度依然強勁。

We expect Q4 revenue in the range of $780 million to $800 million, implying 6% growth at the midpoint. We expect growth driven by tech refresh demand, data center modernization and adoption across our ADSP platform in support of customers hybrid multi-cloud architectures.

我們預計第四季度營收將在 7.8 億美元至 8 億美元之間，這意味著中間值成長率為 6%。我們預計，技術更新需求、資料中心現代化以及整個 ADSP 平台對客戶混合多雲架構的支援將推動成長。

We expect non-GAAP gross margin in the range of 84% to 84.5%. We estimate Q4 non-GAAP operating expenses of $376 million to $388 million. We expect Q4 share-based compensation expense of approximately $57 million to $59 million.

我們預計非公認會計準則毛利率在 84% 至 84.5% 之間。我們預計第四季非公認會計準則營運費用為 3.76 億美元至 3.88 億美元。我們預計第四季股權激勵費用約為 5,700 萬至 5,900 萬美元。

We anticipate Q4 non-GAAP EPS in the range of $3.87 to $3.99 per share. As we look ahead to Q4 and beyond, we see customers increasingly leveraging both the breadth of our solutions across our portfolio and the flexibility only F5 offers to deploy in any form factor.

我們預計第四季非 GAAP 每股盈餘將在 3.87 美元至 3.99 美元之間。展望第四季度及以後，我們看到客戶越來越多地利用我們產品組合中解決方案的廣度以及只有 F5 才能提供的以任何形式部署的靈活性。

As evident in our strong systems performance this year, customers holistic hybrid multi-cloud approach spans hardware, software and SaaS deployments, and they are increasingly investing in high-performance hardware to address factors, including resiliency, data sovereignty and AI readiness. With that backdrop, I will recap the implications of our Q4 guidance on our fiscal year 2025 outlook.

正如我們今年強大的系統性能所證明的那樣，客戶的整體混合多雲方法涵蓋硬體、軟體和 SaaS 部署，並且他們越來越多地投資高效能硬體來解決包括彈性、數據主權和 AI 準備在內的因素。在此背景下，我將回顧我們第四季指引對 2025 財年展望的影響。

Our Q4 revenue guidance implies revenue growth of approximately 9% for FY25. This is up from our prior guidance for growth of 6.5% to 7.5% and reflects robust systems revenue growth and software growth at or around 10% for the year.

我們對第四季營收的預期意味著 2025 財年營收將成長約 9%。這高於我們先前預測的 6.5% 至 7.5% 的成長，反映出今年系統收入強勁成長，軟體成長率達到 10% 左右。

As implied by our Q4 guidance, we continue to expect FY25 non-GAAP gross margin in the range of 83% to 84% and non-GAAP operating margin at or around 35%. We are adjusting our expected FY25 non-GAAP effective tax rate to a range of 18.5% to 19.5% from the prior range of 20% to 22%. This change reflects the onetime benefit recorded in Q3.

正如我們的第四季度指引所暗示的那樣，我們繼續預計 25 財年非 GAAP 毛利率將在 83% 至 84% 之間，非 GAAP 營業利潤率將在 35% 左右。我們將預期的 25 財年非 GAAP 有效稅率從先前的 20% 至 22% 調整至 18.5% 至 19.5% 之間。這項變更反映了第三季記錄的一次性收益。

Our Q4 guidance implies FY25 EPS in a range of 14% to 15% growth, up from our prior guidance of 8% to 10% growth. The increase is driven by the improvement in our revenue outlook and the benefit from the tax rate change. Finally, we expect our share repurchases for the full year to be at or above 50% of our free cash flow.

我們的第四季指引意味著 25 財年每股收益將成長 14% 至 15%，高於我們先前預測的 8% 至 10% 的成長。這一增長是由我們收入前景的改善和稅率變化帶來的好處所推動的。最後，我們預計全年股票回購量將達到或超過我們自由現金流的 50%。

With that, I'll turn the call over to Francois.

說完這些，我將把電話轉給弗朗索瓦。

Thank you, Cooper. F5 delivers and secures every app and API anywhere, on-prem, cloud, edge or hybrid environment, we lead in solving today's toughest business challenges. Our F5 application delivery and security platform is the first in the industry, offering consistent policies, full visibility and AI-driven insights.

謝謝你，庫柏。F5 可在任何地方交付和保護每個應用程式和 API，無論是本地、雲端、邊緣還是混合環境，我們在解決當今最棘手的業務挑戰方面處於領先地位。我們的 F5 應用程式交付和安全平台是業界首創，提供一致的策略、全面的可視性和 AI 驅動的洞察。

We are enabling customers to modernize data centers embrace hybrid multi-cloud and scale for rising performance and security demand in an AI-driven world. Operator, please open the call to questions.

我們正在幫助客戶實現資料中心的現代化，採用混合多雲，並擴大規模以滿足人工智慧驅動的世界中不斷增長的效能和安全需求。接線員，請打開電話詢問。

(Operator Instructions)

（操作員指示）

Tim Long, Barclays.

巴克萊銀行的提姆朗。

Hi, thanks for the time here. Two, if I could. First, on the hardware side. Could you talk a little bit about, obviously, a lot of strength here, a lot of things going on with maybe tariffs and pull-ins and end of life and good competitive environment.

你好，感謝你抽出時間在這裡。如果可以的話，兩個。首先，在硬體方面。您能否談談這裡顯然存在的許多優勢，以及正在發生的許多事情，例如關稅、拉動、壽命終止和良好的競爭環境。

So you kind of break down that you talked about the applications and the region there, but some of the other factors that may be contributed and how you think the sustainability of that will be going into next year? And then second, on the software side, could you just touch on kind of pipeline and how you're thinking about renewals and new deal activity heading into the end of the fiscal year.

那麼，您稍微分析一下您談到的應用程式和地區，但還有一些其他可能起作用的因素，以及您認為這些因素在明年的可持續性如何？其次，在軟體方面，您能否談談通路類型以及您對財政年度結束時的續約和新交易活動的看法。

Tim, why don't I start with your question on hardware and then Cooper will take the second question. On hardware team, as you saw, the results were very strong, 39% growth year-on-year on hardware. And if I separate it, we have two dynamics.

提姆，我先從你關於硬體的問題開始，然後庫柏再回答第二個問題。硬體團隊，如您所見，業績非常強勁，硬體年增 39%。如果我將其分開，我們就有兩種動態。

One, of course, is tech refresh, which is a dynamic in the business. And we expect at this point, giving some platforms that are going end of life in the next 12 to 24 months, and that's having an effect.

其一當然是技術更新，這是業務的動態。我們預計，一些平台將在未來 12 到 24 個月內停止使用，這會產生影響。

So we're seeing strong take refresh this year. But there are other dynamics at place. And when we look at, when we kind of separate our hardware revenue between what's coming from tech refresh and what is not associated with tech refresh.

因此，我們看到今年的業績強勁成長。但還存在其他動態。當我們將硬體收入分為來自技術更新的收入和與技術更新無關的收入。

The part of the revenue that's non-tech refresh is actually growing faster than any other part of the hardware sales. So one of the dynamics that are driving that, it's really secular trends that we're seeing and specifically around hybrid cloud and data centers and around AI.

非技術更新部分的收入實際上比硬體銷售的任何其他部分的成長速度都要快。因此，推動這一趨勢的動力之一是我們所看到的長期趨勢，特別是圍繞混合雲和資料中心以及人工智慧的趨勢。

So in hybrid and multi-cloud, particularly what we're seeing a lot of customers that are embracing hybrid multi-cloud architectures they are modernizing their data centers, investing in data center capacity to get ready for future growth in applications and future architectures. That's a trend that's providing tailwinds to the hardware business.

因此，在混合雲和多雲領域，特別是我們看到許多採用混合多雲架構的客戶，他們正在對其資料中心進行現代化改造，投資資料中心容量，為未來應用程式和未來架構的成長做好準備。這是一個為硬體業務帶來順風的趨勢。

We're also seeing customers respond to, especially in financial services, more regulation requiring customers to have better resiliency, drives them to be able to fail over from public clouds to on-premises environment and ideally vice versa. So that drives customers to have stronger infrastructure on-premises in hardware that they can control.

我們也看到客戶做出反應，特別是在金融服務領域，更多的法規要求客戶具有更好的彈性，促使他們能夠從公有雲故障轉移到內部環境，理想情況下反之亦然。因此，這促使客戶在他們可以控制的硬體中擁有更強大的內部基礎設施。

That is the second secular trend. And of course, the third one is AI. And in AI, we're seeing customers invest both in AI use cases, which we are seeing already today. And this revolves around moving large amounts of data between data stores and AI applications or AI models, and we have won several dozens of customers now that are deploying AI models today and using F5 to be able to move their data to those models.

這是第二個長期趨勢。當然，第三個是人工智慧。在人工智慧領域，我們看到客戶對人工智慧用例進行投資，我們今天已經看到了這一點。這涉及在數據儲存和 AI 應用程式或 AI 模型之間移動大量數據，我們現在已經贏得了數十個客戶，他們正在部署 AI 模型並使用 F5 將他們的數據移動到這些模型。

But we also see a number of customers investing in what we would call AI readiness, meaning they're not yet in full swing of having deployed AI models, but they want to have their data in a good place they want to have enough capacity for what they expect are going to be large closer data that needs to be moved with AI, and they're investing in that capacity to be ready to deploy AI. So those are some of the trends that we're seeing in hardware that I think are pretty durable.

但我們也看到許多客戶正在投資我們所說的人工智慧準備，這意味著他們還沒有完全部署人工智慧模型，但他們希望將他們的數據放在一個好的地方，他們希望有足夠的容量來儲存他們預期的需要透過人工智慧移動的大量數據，他們正在投資這種能力以準備部署人工智慧。這些是我們在硬體領域看到的一些趨勢，我認為它們相當持久。

Now you asked about how does that affect next year? We've seen very strong growth next year. Our current expectation would be that hardware next year would probably be up year-on-year, but obviously, we would expect more modest growth than what we have seen this year.

現在您問這對明年有何影響？我們預計明年將出現非常強勁的成長。我們目前的預期是，明年的硬體可能會年增，但顯然，我們預計成長幅度會比今年更溫和。

To your last point, we have not seen any evidence of pull-in to date in hardware, and we have not seen any effect of any of the tariff discussions on our business, whether it be hardware or software. And with that, Cooper?

關於你的最後一點，到目前為止，我們還沒有看到硬體方面有任何拉動跡象，而且我們也沒有看到任何關稅討論對我們的業務產生任何影響，無論是硬體還是軟體。那麼，庫柏？

Sure. Yeah. Thanks. So I'll speak a little bit to the software dynamics that we're seeing. So we talked about. We had a large renewal base that was coming up in the second half of the year, and that gave us really good visibility.

當然。是的。謝謝。因此我將稍微談談我們所看到的軟體動態。所以我們討論了。我們下半年將迎來一個龐大的續約基礎，這給了我們非常好的可見性。

And what we saw again in this quarter was very healthy expansion against that renewal base. And so that's really coming in two forms. One, it's increased consumption that customers have been driving over the course of their prior subscription term.

我們在本季再次看到，相對於更新基礎而言，擴張非常健康。這實際上有兩種形式。一是客戶在先前的訂閱期間一直在增加消費。

And then at that time of renewal, that consumption then comes through at a higher contract value at the renewal as well as additional expansion across the portfolio. So we're seeing new use cases built in at that time of renewal.

然後在續約時，該消費將以更高的合約價值以及整個投資組合的額外擴展來實現。因此，我們看到在更新時內建了新的用例。

And so that's really where we're seeing the growth in the software. On the new, when we referenced new, we're talking about either net new customers or net new software projects where the customer wasn't previously consuming in software, that business is up year-to-date, but it was down a little bit in Q3.

這就是我們真正看到軟體成長的地方。關於新業務，當我們提到新業務時，我們談論的是淨新客戶或淨新軟體項目，這些項目是客戶以前沒有使用過的軟體，該業務今年迄今一直在成長，但第三季略有下降。

And some of that we kind of goes back to what Francois was referencing around an appetite from customers to support their applications in an environment where regulatory concerns are more prevalent resiliency is more kind of at the forefront and just generally a bias for more performance. And so at the margins, there were a couple of opportunities that actually closed in hardware, where previously the customer might have chose we go with software.

我們在某種程度上可以追溯到 Francois 提到的客戶對在監管問題更加普遍的環境中支持其應用程式的渴望，彈性更為重要，並且通常偏向於更高的性能。因此，在邊緣領域，實際上在硬體方面存在一些機會，而以前客戶可能會選擇我們採用軟體。

And so that's really kind of what's behind the business coming from new projects coming through more in hardware than software. And so we feel good about the year. We have, again, a strong base in Q4. And just based on the expansion trends that we've seen and continue to see, we feel pretty good about our (inaudible) Q4.

這就是新專案業務背後的真正原因，新專案更依賴硬體而非軟體。因此我們對今年感到十分滿意。我們在第四季再次擁有了強勁的基礎。僅基於我們已經看到並將繼續看到的擴張趨勢，我們對（聽不清楚）第四季感到相當滿意。

Meta Marshall, Morgan Stanley.

摩根士丹利的 Meta Marshall。

This is Mary on for Meta. I had a question on gross. Are there any reasons for why gross margins came in at the lower end of the range despite the upside on revenue.

我是 Meta 的 Mary。我有一個關於總額的問題。儘管收入增加，但毛利率卻處於較低水平，這其中有什麼原因嗎？

Yeah. So it's pretty slightly at the low end of the range. That was really driven more by some of the high-performance use cases on the systems business. We have some deals that had FIPS compliance demands, which tend to have a slightly lower gross margin profile.

是的。因此，它處於該範圍的低端。這實際上更多是由系統業務中的一些高效能用例推動的。我們有一些符合 FIPS 合規要求的交易，這些交易的毛利率往往會略低一些。

And then just broadly, the strength in the systems business in general, of course, hardware has a still a very high gross margin profile, but it's not quite as high as software. And so that was the other dynamic that really was behind our gross margins. But as you see from our guidance for Q4, we expect gross margins to improve in the current quarter.

整體而言，系統業務的優勢當然是硬體仍然具有很高的毛利率，但不如軟體那麼高。這就是我們毛利率背後的另一個動力。但正如您從我們對第四季度的預測中看到的那樣，我們預計本季毛利率將會提高。

Michael Ng, Goldman Sachs.

高盛的 Michael Ng。

Hi, good afternoon, thanks for the question. I just have two. First, on the hardware piece. It's encouraging to hear that you expect systems revenue to be up next year. In the past, you've talked a little bit about how much of the installed base was on the legacy iSeries and VIPRION.

嗨，下午好，感謝您的提問。我只有兩個。首先，關於硬體部分。聽到您預計明年系統收入將會上升，真是令人鼓舞。過去，您曾談到有多少安裝基礎是基於舊版 iSeries 和 VIPRION 的。

I was just wondering if you could talk about where you are today and how much of that systems or fresh we should expect over the next couple of years? And then second, just on software kind of relatedly, could you see software revenue growth in fiscal '26 just given the very strong term renewals that we've seen this year.

我只是想知道您是否可以談談您目前的狀況以及未來幾年我們可以期待多少這樣的系統或新產品？其次，就軟體而言，考慮到今年定期續約的強勁表現，您是否認為 26 財年的軟體收入會成長？

Michael, thank you for the question. And we'll do the same thing. I'll start with hardware and ask Cooper will take the second part. On hardware, Michael, we have indeed some end of software support dates that are coming up in '26 and early '27.

邁克爾，謝謝你的提問。我們也會做同樣的事情。我將從硬體開始，然後請 Cooper 負責第二部分。關於硬件，邁克爾，我們確實有一些軟體支援結束日期，即將在 26 年和 27 年初到來。

We expect the refresh to continue to be strong, certainly over throughout 2026 and beyond because customers continue to refresh even after this end of software support dates well after that, they continue to refresh. So not the entire installed base is not refreshed at those dates.

我們預計更新將持續強勁，肯定會持續到 2026 年及以後，因為即使在軟體支援結束之後，客戶仍會繼續更新，他們也會繼續更新。因此，並非所有安裝基礎都會在這些日期進行刷新。

So we continue to expect to see that to be strong over the next 18 months. Now that said, Michael, we're seeing a part of our hardware business that is driven not by tech refreshed motions by other trends that I've just articulated around hybrid multi-cloud architectures, data center modernization and increasingly, customers investing in AI readiness and AI use cases.

因此，我們繼續預計未來 18 個月內這一趨勢將保持強勁。話雖如此，邁克爾，我們看到硬體業務的一部分不是由技術更新的動向驅動的，而是由我剛才闡述的其他趨勢驅動的，這些趨勢包括混合多雲架構、資料中心現代化以及越來越多的客戶對人工智慧準備和人工智慧用例的投資。

We think these trends are less cyclical and more durable. And so we're excited to see these developments, and we'll see how this plays out over time. But for the time being, we're really encouraged by what we're seeing outside of the tech refresh motion in our hardware.

我們認為這些趨勢的周期性較小，而且更加持久。因此，我們很高興看到這些發展，並且我們將觀察其隨著時間的推移將如何發展。但就目前而言，我們對硬體技術更新以外的動態感到非常鼓舞。

And then on the software side, it's a little bit early. We wouldn't guide software for next year, but I'll kind of give you just a few dynamics for consideration is kind of what's a reasonable estimate for where we sit today.

而在軟體方面，現在還為時過早。我們不會指導明年的軟體，但我會給你一些動態供你考慮，這對我們目前所處的位置來說是一個合理的估計。

So I'll start with just a way to think about FY26 in terms of growth rates, and then I'll get into some of the dynamics that we're looking at. I think it's reasonable to assume software would grow in the mid-single digits for next year. And then it could reaccelerate in the following year.

因此，我將首先從成長率的角度來思考 26 財年，然後我將介紹我們正在關注的一些動態。我認為，可以合理地假設明年軟體的成長率將達到中等個位數。然後它可能會在來年再次加速。

And just to kind of walk through some of the things to consider. So first, we're still seeing really strong consumption in that renewal motion. So as I said earlier, this increased performance and consumption along with new use cases that are part of that renewal motion.

只是想介紹一些需要考慮的事情。首先，我們仍然看到更新運動中的消費非常強勁。正如我之前所說，這提高了效能和消耗，同時也帶來了新的用例，這些都是更新動議的一部分。

And just as a point of emphasis, the growth comes from over time is coming from new use cases and the increased consumption that's embedded in the renew and expand motion. So it's not simply repeat business.

需要強調的是，隨著時間的推移，成長來自於新的用例以及更新和擴展運動中嵌入的增加的消費。所以這不僅僅是重複業務。

Then a second dynamic, and we talked about this a little bit on the last call in April is that the subscription base that comes up for renewal in FY26, that base largely comes from our software revenue from FY23 because of that three year renewal cycle. And so our FY23 software sales were roughly flat over FY22.

然後是第二個動態，我們在 4 月的上次電話會議上討論了這一點，即 26 財年需要續訂的訂閱基礎，由於三年的續訂週期，該基礎主要來自 23 財年的軟體收入。因此，我們的 23 財年軟體銷售額與 22 財年基本持平。

So that represents a bit of a mass headwind on that subscription base where we do that renew and expand motion. And then that same dynamic becomes a tailwind into FY27, and it's why we would expect the growth rate to reinflect from there.

因此，這代表著我們在更新和擴展訂閱基礎方面遇到了一些巨大的阻力。然後，同樣的動態將成為 2027 財年的順風，這就是我們預計成長率將從那時重新反映的原因。

And then the last dynamic we touched on it a little bit is just some of the evolving customer preferences around deployment models and this is really one of the big strengths around our ADSP platform that we do give customers choice in how they want to deploy. And so when we talk about hardware and software Something to keep in mind is that those are not products, they're delivery models.

然後，我們稍微談到的最後一個動態只是客戶對部署模型不斷變化的一些偏好，這實際上是我們 ADSP 平台的一大優勢，我們確實為客戶提供瞭如何部署的選擇。因此，當我們談論硬體和軟體時，要記住的是，它們不是產品，而是交付模型。

BIG-IP is a product. And some customers are choosing to deploy big IT in that hardware form factors because of the evolving needs for more performance. And so those are all the things that we consider as we look ahead to next year, and then we'll see how that plays out.

BIG-IP 是一種產品。由於對更高效能的需求不斷增長，一些客戶選擇以這種硬體形式部署大型 IT。這些都是我們在展望明年時會考慮的事情，然後我們再看看結果如何。

Samik Chatterjee, JPMorgan.

摩根大通的 Samik Chatterjee。

This is Priyanka tap on for Samik. Great job this quarter. I got a couple of questions. First of all, on the concept of software next year, what is the new business pipeline for software look like next year?

這是 Priyanka 為 Samik 所做的改變。本季做得很好。我有幾個問題。首先，關於明年軟體的概念，明年軟體的新業務管道是什麼樣的？

And is that kind of incremental to your expectations of mid-single digits software growth? And I have a follow-up.

這是否符合您對中位數個位數軟體成長的預期？我還有一個後續問題。

Sure. Thank you, Priyanka. So yes, our pipeline right now is healthy. So it's early, again, I mean pipeline gives you really good visibility into the current quarter and then decent visibility into the next quarter. And so as you get beyond that, that's going to be a business that we are servicing now is we're engaging with our customers.

當然。謝謝你，Priyanka。是的，我們現在的管道是健康的。所以現在還為時過早，我的意思是，管道可以讓你很好地了解當前季度，然後可以很好地了解下一季。因此，當你超越這一點時，這將成為我們現在服務的業務，我們正在與客戶互動。

But generally, I would say that as we do our planning for next year, we feel good about new opportunities for software. And again, these are either net new customers or customers that are consuming for the first time in software.

但總的來說，我想說，當我們制定明年的計劃時，我們對軟體的新機會感到樂觀。再說一次，這些人要不是淨新增客戶，就是首次購買軟體的客戶。

So that's kind of factored into some of our thinking around the growth rate for next year. But of course, the majority of our software revenue now comes through that renew and expand motion, which is great for visibility, and that's where we would expect the majority of the growth to come from next year.

所以這在某種程度上影響了我們對明年成長率的思考。但當然，我們現在的大部分軟體收入都來自於更新和擴展，這對於提高知名度非常有益，我們預計明年的大部分成長將來自於此。

All right. And on to my second question, you anticipate hardware to grow strongly in 2026. How much of that strength is this newfound shift where people are using systems instead of software that you would otherwise expect for them to use software, like you saw in this particular quarter? Was this unexpected? And is this a trend that you think might continue?

好的。關於我的第二個問題，您預計硬體將在 2026 年強勁成長。這種新轉變有多大力量在於人們正在使用系統而不是軟體，而您原本期望他們使用軟體，就像您在本季看到的那樣？這是否出乎意料？您認為這種趨勢可能會持續下去嗎？

Thanks. Yeah. So we would expect hardware to grow, albeit it will be at a more modest growth rates than what we're seeing this current year because clearly, we're well into the 20% growth rate year-to-date for the current year.

謝謝。是的。因此，我們預計硬體將會成長，儘管其成長率將比今年的成長率更為溫和，因為顯然，今年迄今的成長率已達到 20%。

But we expect continued growth next year. I would say that there's a portion of it that is coming from customer preferences to moving into hardware model.

但我們預計明年將繼續成長。我想說，其中一部分來自於客戶偏好轉向硬體模型。

I don't think that's the main driver. It's really both tech refresh and some of the dynamics where customers really need more performance, and they're trying to scale out their data center capacity to support those performance needs. And then at the margin, there are cases where customers may choose hardware deployment model in lieu of what previously they may have been thinking software for the deployment model.

我認為這不是主要驅動因素。這實際上既是技術更新，也是客戶真正需要更高效能的一些動態，他們正在嘗試擴展資料中心容量以支援這些效能需求。然後在邊際上，有些情況下客戶可能會選擇硬體部署模型來取代他們之前可能考慮的軟體部署模型。

George Notter, Wolfe Research.

喬治·諾特（George Notter），沃爾夫研究公司。

I was just curious on kind of an update on some of the newer products. I was thinking about the NVIDIA DPU product. I think you mentioned it earlier in the monologue. Can you just remind me, is that GA now? How have you guys priced that?

我只是對一些新產品的更新感到好奇。我正在考慮 NVIDIA DPU 產品。我認為您之前在獨白中提到過這一點。可以提醒我一下嗎，現在是 GA 嗎？你們對此的定價是多少？

How significant can that be in the context of your model? Just anything you can say on progress there? And then a similar question on the AI Gateway platform. I know you're just getting into the marketplace right now. I'm just curious on initial views, feedback on the product, what perspective can you share.

在您的模型中這有多重要？能透露一下那裡的進展嗎？然後是關於 AI Gateway 平台的一個類似問題。我知道你現在剛進入市場。我只是好奇對產品的初步看法、回饋，您可以分享什麼觀點。

Thank you, George. Let me start with the partnership with NVIDIA. Our solution is GA, and as a reminder, it's really taking the BIG-IP software that we have refactored to work on our architectures and specifically integrate with NVIDIA's BlueField-3 DPUs.

謝謝你，喬治。讓我先從與 NVIDIA 的合作開始。我們的解決方案是 GA，提醒一下，它實際上採用了我們重構的 BIG-IP 軟體來在我們的架構上運行，並與 NVIDIA 的 BlueField-3 DPU 特別整合。

And the solution is GA, and we're now engaging customers in proof of concept to validate the benefit in their production and architectural environment and the early feedback from customers around these proof of concepts and tests are generally very positive I mentioned a couple in my prepared remarks earlier.

解決方案是 GA，我們現在正在讓客戶參與概念驗證，以驗證其在生產和架構環境中的優勢，客戶對這些概念驗證和測試的早期反饋總體上非常積極，我在之前準備好的發言中提到了幾點。

But everything in AI, as you know, is about tokens and time to first toekn, the cost per token, the GPU utilization and the number of tokens one can generate GPU infrastructure. And we're finding opportunities to basically increase all these metrics and increase the efficiency of the factories.

但正如你所知，人工智慧中的一切都與代幣和首次代幣發行的時間、每個代幣的成本、GPU 使用率以及可在 GPU 基礎設施上產生的代幣數量有關。我們正在尋找機會來提高所有這些指標並提高工廠的效率。

Now it's very early days because whilst we're doing this proof of concept, we think the real strength of the value proposition comes when customers go into inferencing for their AI workloads. And most customers are not there yet.

現在還處於早期階段，因為雖然我們正在進行這個概念驗證，但我們認為價值主張的真正力量來自於客戶對他們的人工智慧工作負載進行推理。但大多數顧客還沒有到達那裡。

So the target customers for the solution are either to as a service providers or selling or renting GPUs to customers, and we'll want to make that infrastructure as efficient as possible or enterprises that will build self-hosted AI factories, and we also want to make their own investment in AI factories as efficient as possible.

因此，該解決方案的目標客戶要么是服務提供商，要么是向客戶銷售或出租 GPU，我們希望使該基礎設施盡可能高效，要么是將構建自託管 AI 工廠的企業，我們也希望使他們自己對 AI 工廠的投資盡可能高效。

So we think as enterprises, start building more of the factories and really moving to inferencing in these factories, then our opportunity to move more to revenue will be more concrete. So the technical validation has to perceive that with customers, and that's what we're doing, and it's pretty promising.

因此我們認為，作為企業，開始建造更多的工廠並真正轉向這些工廠的推理，那麼我們獲得更多收入的機會將更加具體。因此，技術驗證必須讓客戶感受到這一點，這就是我們正在做的事情，而且前景非常光明。

As it relates to the AI gateway, that solution is also early in the market. We have I think more than a dozen proof of concept in place with customers. This really is about routing AI traffic and processing AI traffic at Layer 7 to really understand what is the cost per token that customers have, the cost of this model versus that model, providing security to this AI traffic.

由於它與 AI 網關相關，該解決方案在市場上也處於早期階段。我認為我們已經與客戶進行了十多次概念驗證。這實際上是關於路由 AI 流量和在第 7 層處理 AI 流量，以真正了解客戶擁有的每個代幣的成本、此模型與該模型的成本，為該 AI 流量提供安全性。

And we think this is going to be the whole area of AI delivery and security, we think it's going to be a pretty important area in AI because if you step back from it, web infrastructure has really been about largely processing packets. And we think that which is largely a Layer 3 activity.

我們認為這將成為人工智慧交付和安全的整個領域，我們認為這將成為人工智慧中一個非常重要的領域，因為如果你退一步來看，網路基礎設施實際上主要是處理資料包。我們認為這主要是第 3 層活動。

And we think that AI infrastructure is really going to be about delivering and securing token. And that is the Layer 7 activity. And we think F5 with our Layer 7 expertise built over a decade is going to be really well positioned for that. So more to come on the AI gateway over time and what we're going to do in AI security, but it's an area that we are going to invest in and take a position.

我們認為人工智慧基礎設施的真正目的是傳遞和保護令牌。這就是第 7 層活動。我們認為，憑藉十多年來累積的第 7 層專業知識，F5 將在這方面佔據有利地位。因此，隨著時間的推移，我們將在人工智慧網關方面做出更多努力，並在人工智慧安全方面採取更多行動，但這是我們將進行投資並佔據一席之地的領域。

Matthew Hedberg, RBC Capital Markets.

加拿大皇家銀行資本市場 (RBC Capital Markets) 的馬修‧赫德伯格 (Matthew Hedberg)。

Mike Richards on for Matt. Two quick ones for me. Just first, anything to call out on Fed. I know you guys were expecting there might be some pull-in in Q3 and anything to call out as sort of one month into the Fed fiscal year-end? And then secondly, any dog rails you put around free cash flow looking into next year as we think about the R&D credits?

麥克·理查茲替換馬特。對我來說，有兩個簡單的問題。首先，有什麼可以向聯準會呼籲的嗎？我知道你們預計第三季可能會出現一些拉動，聯準會財政年度結束一個月後會有什麼消息嗎？其次，當我們考慮研發信用時，您是否對明年的自由現金流設定了任何限制？

Yeah. So I'd say Fed has been right at kind of our plan for the year. Q3 was a little bit softer. I think there were a couple of projects that pushed out or were downsized kind of related to some of the government efficiency initiatives.

是的。所以我想說聯準會今年的計畫是正確的。Q3 稍微軟一點。我認為有幾個項目被推遲或縮減，這與政府的一些效率措施有關。

But broadly, the pipeline is still healthy for federal, so nothing else that we would really call out. And then can you remind me the question on free cash flow?

但總體而言，對於聯邦政府來說，管道仍然健康，因此我們沒有什麼可說的。然後您能提醒我關於自由現金流的問題嗎？

Yeah. Just as we're thinking about the tax changes around the R&D credits, just anything around free cash flow as we look to next year?

是的。正如我們正在考慮有關研發抵免的稅收變化一樣，我們對明年的自由現金流有什麼展望嗎？

No, there wouldn't be a big impact from what we see in our business right now. There are certain elections that you have the discretion to take. So that's something we're evaluating right now. But no material change from where we sit today.

不，從我們目前的業務情況來看，這不會造成太大影響。您有權自行決定是否參加某些選舉。所以這是我們現在正在評估的事情。但與我們今天的情況相比，沒有實質的變化。

James Fish, Piper Sandler.

詹姆斯·菲什、派珀·桑德勒。

Francois, you keep talking about the non-refresh piece on systems growing faster, but let's be frank, most of the systems is refresh. So really, the crux of my question though is, can you just help us with how big essentially the non-refresh pieces and really like trying to understand how much of an AI benefit you're seeing today?

弗朗索瓦，你一直在談論非刷新系統部分增長更快，但坦白說，大多數系統都是刷新的。所以實際上，我的問題的關鍵是，您能否幫助我們了解非刷新部分本質上有多大，並真正想了解您今天看到的 AI 好處有多大？

So let me just resize this for you. About two-third of what we're seeing in hardware deployment models is what we would attribute to tech refresh and about one-third is what we would attribute to non-tech refresh. Just to give you a little bit of the weighting of what we are seeing this year.

因此，讓我為您調整一下它的大小。我們在硬體部署模型中看到的大約三分之二是技術更新的結果，大約三分之一是非技術更新的結果。只是想讓你稍微了解一下我們今年所看到的情況。

Now to your second question about AI there really are a couple of answers to that. What we are seeing as direct AI use cases where our sales teams working with customers can actually attribute a project or a certain order specifically to a net new AI initiative.

現在，對於您關於人工智慧的第二個問題，確實有幾個答案。我們所看到的直接人工智慧用例是，我們的銷售團隊與客戶合作，實際上可以將某個項目或某個訂單具體歸因於全新的人工智慧計畫。

And today, I would say that this is still small in single-digit millions of dollars on a quarterly basis. Several dozens of wins over the last several quarters. There is, however, another portion of AI, which is harder to estimate exactly, but it's kind of indirect AI spend coming to us.

今天，我想說的是，按季度計算，這個數字仍然很小，只有幾千萬美元。過去幾個季度取得了數十場勝利。然而，還有另一部分人工智慧，很難準確估計，但它是我們獲得的間接人工智慧支出。

And this is where customers are investing in having higher capacity for data connections that may be related to an AI use cases, but the folks who are interacting with F5 may not be involved directly in this use case and don't know about the alternate use of the technology.

這就是客戶投資於更高容量的資料連接的地方，這些資料連接可能與 AI 用例相關，但與 F5 互動的人可能並未直接參與此用例，也不知道該技術的替代用途。

And we're calling that internally, kind of shadow AI use cases. We think there are more and more of these growing, but it's difficult to give you an exact sizing of how much that is impacting the business.

我們在內部稱之為影子人工智慧用例。我們認為這些現象會越來越多，但很難確切地估計它們對業務的影響有多大。

Understood. And maybe if I could follow up there. appreciate that added color. How are you thinking about maybe the magnitude of customers that on tech refresh could actually become more virtual or even given the nearly 40% growth here, and we talked about this in the piece we did earlier this week. But what inning of refresh do you think we're actually in?

明白了。也許如果我能跟進的話，我會很感激所添加的色彩。您如何看待技術更新後客戶的數量實際上可能變得更加虛擬化，或者甚至考慮到這裡近 40% 的增長，我們在本週早些時候的文章中討論了這一點。但您認為我們實際上處於哪個更新階段？

Yeah. So Jim, we're pretty early in terms of the refresh opportunity for the eye care and VIPRION product families. Well over half of the base is still on those families. And as Francois said, that every customer is different in terms of their time lines. We would expect that base to get refreshed really over the next two plus years.

是的。吉姆，就眼部護理和 VIPRION 產品系列的更新機會而言，我們還處於相當早期的階段。超過一半的人口基礎仍然依賴這些家庭。正如弗朗索瓦所說，每個顧客的時間線都是不同的。我們預計該基礎將在未來兩年多的時間內真正更新。

So we think that it's a good growth opportunity for FY26 and into '27. And then you would expect that to start to tail off in FY28. And we're not seeing a lot of migration from those legacy systems into software form factors. I think that the dynamic that we're actually seeing right now is where customers are really getting more consideration to their performance needs as they evaluate between hardware and software.

因此我們認為這對 2026 財年和 2027 財年來說是一個很好的成長機會。然後，您會預期這一數字將在 28 財年開始下降。我們並沒有看到大量從傳統系統遷移到軟體形式因素的。我認為我們現在實際看到的動態是，客戶在評估硬體和軟體時真正更多地考慮他們的效能需求。

Simon Leopold, Raymond James.

西蒙李奧波德、雷蒙詹姆斯。

I guess one of the things I'm curious about is you've cautioned us in the past about your service provider or telco vertical, having a tendency towards lumpiness. And this was a particularly strong quarter. It looks like it was up more than 40% sequentially and nice year-over-year growth.

我想我很好奇的一件事是，您過去曾警告過我們，您的服務供應商或電信垂直行業有不均衡的趨勢。這是一個特別強勁的季度。看起來它比上一季增長了 40% 以上，而且同比增長也不錯。

How should we be thinking about that should we take this as more of a one-off? Or is there some new trending that we should think about? And then I've got a quick follow-up.

我們該如何看待此事，是否應該將其視為一次性事件？或者是否存在一些我們應該考慮的新趨勢？然後我有一個快速的跟進。

Thank you, Simon. The service provider business continues to be lumpy, and it is tied to specific projects at some of our large carrier customers. And so I wouldn't read too much into a single quarter trend in service provider, whether it's a big quarter with things going north very quickly or a down quarter.

謝謝你，西蒙。服務提供者業務持續不穩定，並且與我們一些大型營運商客戶的特定專案相關。因此，我不會過度解讀服務提供者單一季度的趨勢，無論這是一個快速上升的季度，還是一個下滑的季度。

Generally, what we're seeing with service providers I mean the promise of 5G hasn't fully materialized. I would say, for them or for equipment providers to a large extent, in part because there hasn't been the take-off that we expected.

總體而言，我們看到服務提供者對 5G 的承諾尚未完全實現。我想說，對於他們或設備供應商來說，很大程度上，部分原因是沒有出現我們預期的起飛。

We continue to see service providers invest in 4G to 5G infrastructure over the last couple of years, they have tended to sweat their assets significantly. We have seen that relaxed a little bit over the last couple of quarters. But I don't think the dynamics are fundamentally changing in that segment.

過去幾年，我們不斷看到服務提供者對 4G 到 5G 基礎設施進行投資，他們往往投入大量資產。我們看到，過去幾個季度，這種情況有所緩解。但我並不認為該領域的動態會發生根本性的變化。

And then I want to see if you could talk a little bit more about how to think about the longer-term trends or fiscal '26 expectations for services in that I would have thought we'd see some correlation to the stronger hardware business, but maybe there's a lag effect. And I'm assuming there's very little correlation to your software stand-alone business, so how should we think about services trending?

然後，我想看看您是否可以再多談談如何看待服務的長期趨勢或 26 財年預期，我原以為我們會看到與更強勁的硬體業務之間的某種關聯，但也許存在滯後效應。我認為這與您的軟體獨立業務關聯很小，那麼我們應該如何看待服務趨勢？

Yeah. So you're actually right. There is a lag effect on product. And so we would expect to see the growth rate kind of come up from what we just reported with the 1%. I think there's a kind of an outlier dynamic that's behind the deceleration in the last couple of quarters, which is really around the kind of the last of the refresh from the prior product family, those laggard customers that have finally retired some of those end of technical support units.

是的。所以你是對的。產品存在滯後效應。因此，我們預計成長率將比我們剛剛報告的 1% 上升。我認為過去幾季的成長減速背後存在著一種異常動態，這實際上與先前產品系列的最後一次更新有關，那些落後的客戶最終淘汰了一些技術支援部門。

So those came out of the maintenance base that drives our services revenue. That's now behind us. And so what we do expect is to see the services revenue start to grow now from that live effect on the strength of the product revenue.

所以這些都是來自於維護基礎，推動了我們的服務收入。現在，這一切已經過去了。因此，我們確實期望看到服務收入從產品收入強勁的現場效應中開始成長。

And you can see that in the deferred revenue, which is up 10% year-over-year, the short-term deferred revenue, which is really kind of a proxy for the maintenance revenue growth. That's up 5% year-over-year. And so we do expect to see the services growth rate to be a little bit better into FY26.

您可以看到遞延收入年增 10%，短期遞延收入實際上是維護收入成長的替代指標。這比去年同期增長了 5%。因此，我們確實預期 2026 財年的服務業成長率會略有改善。

Tal Liani, Bank of America.

塔爾·利亞尼（Tal Liani），美國銀行。

It's Tomer Zilberman on for Tal. Two questions for you. Maybe going back to hardware but asking it a different way. I believe last quarter, you mentioned you expected hardware systems to moderate. So what changed in the last 90 days for that to accelerate?

托默·齊爾伯曼 (Tomer Zilberman) 代替塔爾 (Tal)。問你兩個問題。也許可以回到硬件，但以不同的方式詢問。我相信上個季度您曾提到過，您預計硬體系統將會放緩。那麼，過去 90 天發生了哪些變化，導致這項進程加速？

Was it better tech refresh than you initially expected? I know you called out last quarter that you were seeing competitive displacements in hardware. Are you still seeing that?

這次技術更新是否比您最初預期的更好？我知道您在上個季度就提到過，您看到了硬體方面的競爭性取代。你還看到這個嗎？

Yes. So yes, for the last point, yes, we still continue to see competitive displacements. Our value proposition is very strong relative to competitors, both with our current offering and our road maps. And so that continues.

是的。是的，對於最後一點，是的，我們仍然繼續看到競爭性取代。相對於競爭對手而言，我們的價值主張非常強大，無論是我們目前的產品還是我們的路線圖。事情就這樣繼續下去。

I think what's changed over the last 90 days is, we saw more of these secular trends gathering pace on customers investing in more capacity in their data centers in chart because they're addressing hybrid multi-cloud in part because they are sometimes getting more resiliency in place, sometimes in response to regulation in certain sectors and in part because of getting ready for AI. We see customers repatriating some apps.

我認為過去 90 天發生的變化是，我們看到更多長期趨勢在加速發展，客戶在圖表中投資資料中心的更多容量，因為他們正在解決混合多雲問題，部分原因是他們有時獲得了更多的彈性，有時是為了響應某些行業的監管，部分原因是為人工智慧做好準備。我們看到客戶正在遣返一些應用程式。

We see customers be patenting data in readiness for AI and doing so because the cost of having this data in public clouds and accessing this data are quite prohibitive. So we see more customers also doing data repatriation and needed to securely connect these data stores that are on-prem with their applications.

我們看到客戶正在為人工智慧準備資料併申請專利，這樣做是因為在公有雲中儲存和存取這些資料的成本非常高。因此，我們看到越來越多的客戶也在進行資料遣返，並且需要將這些內部部署的資料儲存與他們的應用程式安全地連接起來。

So some of these trends when you aggregate them, created a positive surprise on our hardware deployment model. That said, we've talked a lot on this call about hardware versus software and really want to make sure that what we convey to you is what we're seeing is, the power of F5's platform, our application delivery and security platform, which really brings all of these deployment models together, hardware, software and Software as a Service in a single platform.

因此，當你將這些趨勢匯總在一起時，會對我們的硬體部署模型產生積極的驚喜。也就是說，我們在這次電話會議上討論了很多關於硬體與軟體的問題，並且真的想確保我們向您傳達的是我們所看到的 F5 平台的強大功能，我們的應用程式交付和安全平台，它真正將所有這些部署模型整合在一起，將硬體、軟體和軟體即服務整合到一個平台上。

Increasingly, we're seeing is that's what's differentiating us in the marketplace. So this quarter, we are happy about the numbers, of course, but we are even happier about where we are winning. We are winning in customers that may have had a single SaaS vendor and or a single hardware vendor, and we are consolidating all that spend on because we can deliver a hardware in software and SaaS.

我們越來越多地看到，這就是我們在市場上的差異化優勢。因此，本季我們當然對業績數字感到滿意，但更讓我們高興的是我們的勝利。我們正在贏得那些可能只有單一 SaaS 供應商和/或單一硬體供應商的客戶，我們正在整合所有的支出，因為我們可以在軟體和 SaaS 中交付硬體。

We are seeing customers that have point security solutions, maybe in software or hardware and they're consolidating on F5 because we bring this platform that brings all these form factors together, all the delivery and security capabilities together and a single pane of glass to make the operation of these technologies much easier.

我們看到客戶擁有點安全解決方案，可能是軟體或硬件，他們正在 F5 上進行整合，因為我們帶來了這個平台，它將所有這些外形尺寸、所有交付和安全功能整合在一起，並提供單一玻璃面板，使這些技術的操作變得更加容易。

And increasingly, that's what you're seeing from large enterprise customers if they want their life to be simplified and the platform approach that we have taken speaks to that pain point, and we're getting significant traction as a result of that. So that's what is exciting for us in the business is really the early traction we're getting on our platform strategy and the wins across multiple form factors.

越來越多的大型企業客戶希望簡化他們的生活，他們就會這樣做，而我們採用的平台方法正好解決了這個痛點，因此我們獲得了顯著的吸引力。因此，對我們來說，令人興奮的是我們在平台策略上取得的早期進展以及在多種形式因素上的勝利。

Got it. Maybe as a follow-up, Francois, on the AI piece. You mentioned that it's still early days as we're awaiting enterprises to really build out these AI factories, we generally view that phase maybe by the end of 2026, moving into 2027.

知道了。弗朗索瓦，也許這是關於人工智慧部分的後續。您提到現在還處於早期階段，因為我們正在等待企業真正建立這些人工智慧工廠，我們普遍認為這個階段可能要到 2026 年底，也就是 2027 年。

My question for you is, do you see that adoption time line actually accelerating as we move more into the world of genetic. And then generally speaking, how big do you view the opportunity in AI delivery versus AI security? What's going to be more beneficial to you?

我的問題是，隨著我們進一步進入基因世界，您是否認為採用時間表實際上正在加速。那麼整體來說，您認為人工智慧交付與人工智慧安全的機會有多大？什麼對你來說更有益？

Well, that is both a great question, and I'll pick up my crystal ball to give you an answer. Of course, it's very difficult to extrapolate when you're very, very early stage in market to the big predictions about the impact.

嗯，這是一個很好的問題，我會拿起我的水晶球來給你答案。當然，當你處於市場非常早期的階段時，很難推斷出對其影響的重大預測。

But I would say AI did a delivery is probably more immediate for us because it's things that F5 has always done the ability to process traffic at Layer 7 and securely connect applications to applications or users to applications. Now we have to connect data stores to applications.

但我想說，AI 的交付對我們來說可能更直接，因為 F5 一直在做的事情就是處理第 7 層的流量，並安全地將應用程式連接到應用程式或將用戶連接到應用程式。現在我們必須將資料儲存連接到應用程式。

It requires new protocols but F5 is ideally positioned to do that. And the more customers need high performance and high scale, the more they come to F5 because of our ability to do that at high speed and to support protocols like MC3 and other protocols that are specific to AI in general and increasingly, it's AI security, we think, especially in the case of AI runtime security, protecting AI workloads in production we think it's going to be a very substantial opportunity for F5.

它需要新的協議，但 F5 具有理想的條件來實現這一點。客戶對高性能和高規模的需求越多，他們就越傾向於選擇 F5，因為我們能夠高速實現這些目標，並支援 MC3 等協議以及其他特定於 AI 的協議，而且我們認為，AI 安全性日益重要，特別是在 AI 運行時安全性的情況下，保護生產中的 AI 工作負載，我們認為這對 F5 來說將是一個非常重要的機會。

Again, that AI Layer 7 security is going to be extremely important where you have to look at every token and make sure that it's going in the right way and there's no malicious front injection and other threats that are dealt with. We think that is going to be a very large opportunity.

再次強調，AI 第 7 層安全性極為重要，您必須查看每個代幣並確保其以正確的方式運行，並且沒有惡意前端注入和其他需要處理的威脅。我們認為這將是一個非常巨大的機會。

I think it will develop a little, it will take a little longer to develop that AI data delivery, which we're starting to see right now. But we're starting to see customers already be very sensitive to securing their AI workload and starting to investigate what solutions they could have to do that.

我認為它會有所發展，開發人工智慧資料傳輸需要更長的時間，我們現在開始看到這一點。但我們開始看到客戶已經對保護他們的人工智慧工作負載非常敏感，並開始研究他們可以採用哪些解決方案來做到這一點。

So we think that market is absolutely going to happen. The pace at which it happens is less clear for us.

所以我們認為這個市場絕對會發生。我們不太清楚這種情況發生的速度。

Ryan Koontz, Needham & Company.

瑞安·孔茨（Ryan Koontz），Needham & Company。

First, just a clarification. When you talk about tech refresh, I assume you're talking about F5 to F5 legacy to modern product. Second question I have is regarding your our modest price increase you've talked about phasing in kind of where are you in terms of that way can make its way through the model? And what's been the customer feedback on that relative to some of your competitors moves?

首先，只是澄清一下。當您談到技術更新時，我假設您談論的是 F5 從舊版 F5 到現代產品。我的第二個問題是，關於你們所說的適度提價，你們打算分階段實施這種提價方式嗎？與競爭對手的舉措相比，顧客對此有何回饋？

Sure. Yes. When we talk about tech refresh, we're talking about refreshing F5. We also have had good success displacing competitors. So there is a kind of a separate refresh motion around competitive installed base, but that's not part of that (inaudible) speaking specifically to tech refresh.

當然。是的。當我們談論技術更新時，我們談論的是更新 F5。我們在取代競爭對手方面也取得了成功。因此，圍繞競爭性安裝基礎存在一種單獨的更新動議，但這不是（聽不清楚）專門針對技術更新的一部分。

And then in terms of the price increases. So yes, we announced a price increase in January, that we're starting to see that through come through in the numbers. I think that customer reception has been reasonable.

然後是價格上漲方面。是的，我們在一月份宣布了價格上漲，現在我們開始從數據上看到這項變化。我認為客戶接待還是合理的。

We've seen other peers in the space that have had much more aggressive pricing practices that have frankly turned off a lot of customers, and that's driving some business our way. We want to make sure that we're delivering value commensurate with the price increases that we introduced in January. And so we feel pretty good about where we sit from that perspective.

我們發現該領域的其他同行採取了更為激進的定價策略，坦白說，這已經讓許多客戶望而卻步，而這卻為我們帶來了一些業務。我們希望確保我們提供的價值與我們在一月份推出的價格上漲相稱。因此，從這個角度來看，我們對自己的處境感到非常滿意。

We have reached the end of the question-and-answer session. I'd now like to turn the call back over to Francois Locoh-Donou, for closing comments.

問答環節已結束。現在我想將電話轉回給 Francois Locoh-Donou，請他發表最後評論。

Thank you for joining us today. We look forward to seeing many of you during the quarter and to discussing F5's growing role in the broader hybrid multi-cloud landscape. Thank you.

感謝您今天加入我們。我們期待在本季度見到你們，並討論 F5 在更廣泛的混合多雲領域中日益增長的作用。謝謝。

This concludes today's conference. You may disconnect your lines at this time, and we thank you for your participation.

今天的會議到此結束。現在您可以斷開線路，感謝您的參與。