Microsoft conference call: In the past 12 months, over 2 gigawatts of data center capacity have been added, the number of users for the Copilot product line has exceeded 100 million, and cloud migration business has become a growth engine

Microsoft demonstrated strong growth momentum at the close of fiscal year 2025, with AI and cloud business continuing to surge, and Azure cloud service revenue increasing by 39% year-on-year. The expansion of AI infrastructure accelerated, leading in data center construction compared to competitors, with the number of users for the Copilot product line surpassing 100 million, and cloud migration business becoming a growth engine. On July 30, Microsoft announced its fiscal year 2025 fourth-quarter financial report, showing revenue of $76.4 billion, a year-on-year increase of 17%, with Azure cloud service revenue growing by 39% year-on-year, and annual revenue exceeding $75 billion for the first time. Microsoft's cloud business revenue surpassed $168 billion for the first time, a year-on-year increase of 23%. According to a previous article from Jianwen, Microsoft's total capital expenditure of $24.2 billion in the fourth quarter of fiscal year 2025 set a record for a single quarter, increasing by $2.8 billion from the first quarter, with a growth rate of nearly 13.1%, and exceeding analysts' expected expenditure of $23.17 billion by 4.4%. The expenditure data for the second quarter did not show any signs of a slowdown in data center investment. After the financial report was released, Microsoft CEO Satya Nadella, CFO Amy Hood, and Vice President of Investor Relations Jonathan Neilson attended a conference call to elaborate on the company's quarterly financial performance, AI products, cloud business growth, and answered analysts' questions regarding the commercialization path in AI transformation and capital expenditure. CEO Satya Nadella stated, "The company ended a record fiscal year with a very strong performance," noting that the rapid innovation and widespread application of AI technology are driving breakthrough growth in multiple business areas. The user adoption rate of AI products such as Microsoft 365 Copilot and GitHub Copilot has reached "unprecedented" levels. Key factors driving Azure's rapid growth include active VMware and SAP migration activities, as well as the demand for migration of Microsoft's server products. Nadella pointed out, "It turns out we are far from the finish line—at best, we may still be in the middle of the game," implying that there is still significant growth potential in the migration business. The company is also continuously exploring new models for monetizing AI products. In response to analysts' concerns about the continued growth of capital expenditure, CFO Amy Hood emphasized that the company has $368 billion in backlog orders, and capital expenditure is highly aligned with actual demand. She stated that the supply-demand tightness is expected to continue until December this year, and the company is more focused on gaining market share rather than the inflection point of capital expenditure. Microsoft maintains a leading position in the AI infrastructure race, adding over 2 gigawatts of data center capacity in the past 12 months, operating over 400 data centers in 70 regions worldwide, surpassing any other cloud service provider. The company's AI capabilities have over 800 million monthly active users, and the monthly active users of Copilot applications for business and consumer have surpassed 100 million. Microsoft Chief Financial Officer Amy Hood revealed that commercial bookings exceeded $100 billion for the first time, a year-on-year increase of 37%, and commercial remaining performance obligations rose to $368 billion. It is expected that fiscal year 2026 will continue to see double-digit revenue and operating income growth, with Azure revenue expected to grow by 37% year-on-year in the first fiscal quarter. The following is a summary of key points from Microsoft's conference call: 1. Microsoft Cloud revenue hits a record high: "Microsoft Cloud's annual revenue exceeded $168 billion, a year-on-year increase of 23%," "Azure's annual revenue exceeded $75 billion, a year-on-year increase of 34%." 2. Large-scale expansion of AI infrastructure: "In just the past 12 months, we have added over 2,000 megawatts of new capacity," "We continue to expand our data center capacity faster than any competitor." 3. Acceleration of Microsoft 365 Copilot adoption: "Our Co-Pilot application family has surpassed 100 million monthly active users in both commercial and consumer sectors," "We set a new record for the number of new seats added this quarter, achieving the highest number of customer seat repurchases since launch." 4. Financial performance exceeds expectations: "This quarter, we achieved revenue of $76.4 billion, a year-on-year increase of 18%," "Our commercial bookings exceeded $100 billion for the first time, a year-on-year increase of 37%." 5. Rapid growth of Azure AI Foundry platform: "80% of Fortune 500 companies are using Foundry. Based solely on the number of tokens provided by the Foundry API, we processed over 500 trillion tokens this year, a year-on-year increase of over 7 times." 6. GitHub Copilot users exceed 20 million: "We currently have 20 million GitHub Copilot users," "The number of GitHub Copilot enterprise customers increased by 75% year-on-year." 7. Optimistic outlook for fiscal year 2026: "We expect to achieve double-digit revenue and operating income growth in fiscal year 2026," "We anticipate capital expenditures in the first quarter will exceed $30 billion, reflecting the strong demand signals we are seeing." 8. Security business continues to lead: "We now have nearly 1.5 million security customers and continue to gain market share across all major categories of our services," "Purview is used by three-quarters of Microsoft 365 Copilot customers to protect their data." Large-scale expansion of AI infrastructure drives growth Microsoft has added over 2,000 megawatts of data center capacity in the past 12 months, currently operating over 400 data centers in 70 regions across six continents, surpassing any other cloud service provider All Azure regions of the company have achieved "AI-first" deployment, supporting liquid cooling technology to enhance equipment versatility and flexibility. The Azure business performed particularly well, with revenue growth of 39%, far exceeding expectations, mainly due to the accelerated growth of the core infrastructure business among large customers. Despite the launch of more data center capacity this quarter, demand still exceeds supply. The Microsoft Sovereign Cloud launched by the company provides customers with the industry's most comprehensive public and private cloud deployment solutions. In the data and analytics field, Microsoft Fabric has become a complete data analytics platform for the AI era, with revenue growing by 55% year-on-year and the number of customers exceeding 25,000, making it the fastest-growing database product in Microsoft's history. The Azure AI Foundry platform helps customers design and manage large-scale AI applications, with the number of tokens processed this year exceeding 500 trillion, a year-on-year increase of over 7 times. Cloud Migration Business Becomes Growth Engine The rapid growth of Azure cloud services is largely attributed to the continued activity of enterprise migration. Nadella specifically pointed out that there is strong demand for customers migrating SAP and VMware instances as well as Microsoft server products to the cloud. " It turns out we are still far from the finish line—at best, we may still be in the mid-game stage," he stated. In addition to migration business, the ongoing expansion of new AI workflows and cloud-native applications has also driven Azure's growth. Some Microsoft cloud application customers are new users of Azure, bringing new growth momentum to the platform. In the enterprise market, Microsoft performed impressively. Microsoft CFO Amy Hood revealed that commercial bookings have surpassed $100 billion for the first time, a year-on-year increase of 30%. Commercial remaining performance obligations (CRPO) increased to $368 billion, a year-on-year growth of 35%, with about 35% expected to convert to revenue in the next 12 months. Copilot Product Line User Count Exceeds 100 Million The Copilot product line performed exceptionally well. Monthly active users for commercial and consumer users exceeded 100 million, and the monthly active users of AI features across all Microsoft products surpassed 800 million. The adoption rate of Copilot has outpaced any other new product in the Microsoft 365 suite, achieving the largest seat growth since its launch this quarter. Among them, Barclays will deploy Copilot among 100,000 employees globally, and UBS plans to expand to all employees after piloting with 55,000 employees. Companies such as Adobe, KPMG, Pfizer, and Wells Fargo have all purchased over 25,000 seats this quarter. GitHub Copilot continues to gain strong momentum among the developer community, currently boasting 20 million users, with the number of enterprise customers increasing by 75% year-on-year, and 90% of Fortune 100 companies are using the service. The number of AI projects on GitHub has doubled in the past year, with the platform executing millions of code reviews each month In the healthcare sector, Dragon Copilot has performed outstandingly, with clients recording over 13 million doctor-patient consultations using AI environmental solutions this quarter, a year-on-year increase of nearly 7 times. Mercy Health System has over 1,000 doctors using the service, saving more than 100,000 hours. Technology Stack Maturity Accelerates Application Development In describing this quarter's performance, Nadella particularly emphasized the improvement in the maturity of the AI application development technology stack. He pointed out that the platform has moved beyond the simple phase of "here's a model, here's an API, go call it," to a very stateful application model that requires significant rethinking of the entire application stack. The complexity of the storage layer demands a substantial increase, including how many indexes to create, how many indexes to build through preprocessing, so that prompt engineering and context engineering can run better and at higher quality. Frameworks around products like Azure Search, Fabric, and Cosmos DB have become robust enough to build truly serious applications. Nadella expressed excitement about the learning curve of the technology stack, the speed of diffusion, and the speed of application building. He compared this to the emergence of relational databases, noting that based on the current maturity being formed, the industry is rapidly building quite complex applications. The full AI translation of Microsoft's conference call is as follows: Microsoft FY2025 Q4 Conference Call Date: July 30, 2025 Company Name: Microsoft Event Description: FY2025 Q4 Earnings Call Source: Microsoft Host: Hello everyone, welcome to Microsoft's FY2025 Q4 earnings call. Currently, all participants are in listen-only mode. After the formal remarks, there will be a Q&A session. (Host instruction) A reminder that this meeting is being recorded. Now, I am honored to introduce Jonathan Neilson, Vice President of Investor Relations at Microsoft. Jonathan Neilson, Vice President of Investor Relations: Good afternoon, thank you for joining our call today. Joining me today are: Chairman and CEO Satya Nadella, Chief Financial Officer Amy Hood, Chief Accounting Officer Alice Jolla, and Corporate Secretary and Deputy General Counsel Keith Dolliver. On the Microsoft Investor Relations website, you can find our earnings press release and financial summary slides, which are intended to supplement the remarks prepared for today's conference call and provide reconciliation information between GAAP and non-GAAP financial metrics More detailed outlook slides will be released simultaneously on the Microsoft Investor Relations website when we provide forward-looking commentary today. In this conference call, we will discuss certain non-GAAP items. The non-GAAP financial metrics provided should not be viewed as a substitute for or superior to the financial performance metrics prepared in accordance with GAAP. These non-GAAP metrics are provided as additional explanatory content, aimed at helping investors further understand the company's performance in the fourth quarter, as well as the impact of related items and events on financial results. All growth comparisons made during today's conference call, unless otherwise noted, are relative to the same period last year. Where possible, we will also provide growth rates calculated at constant currency as a reference framework for assessing our underlying business performance, excluding the impact of foreign exchange rate fluctuations; when the growth rates are the same under constant currency, we will only mention the growth rate itself. After the conference call, we will immediately publish the prepared remarks on our website until the complete transcript is released. Today's conference call is being live-streamed and recorded. If you ask a question, your question will be included in our live transmission, meeting transcript, and any recordings that may be used in the future. You can replay this conference call and view the transcript on the Microsoft Investor Relations website. In this meeting, we will make forward-looking statements, which are predictions, outlooks, or other statements regarding future events. These statements are based on current expectations and assumptions and are subject to various risks and uncertainties. Actual results may differ significantly from expectations due to factors discussed in today's earnings press release, conference call remarks, and reports and documents filed with the U.S. Securities and Exchange Commission, including Form 10-K, Form 10-Q, and others. We do not undertake any obligation to update any forward-looking statements. Next, I will hand the call over to Satya. Satya Nadella, Chairman and CEO: Thank you, Jonathan. We ended a record fiscal year with very strong performance. Overall, Microsoft Cloud annual revenue exceeded $168 billion, growing 23% year-over-year. The speed of innovation and the breadth of diffusion are unprecedented. Because of this, we are building the most comprehensive AI product and technology stack at scale. To clarify this further, I want to start from the bottom of the technology stack, beginning with Azure. Azure annual revenue exceeded $75 billion, growing 34% year-over-year, driven by comprehensive growth across various workloads We continue to lead the AI infrastructure wave and have achieved market share growth in every quarter of this fiscal year. We have built data centers on six continents, currently operating over 400 data centers in 70 regions, more than any other cloud service provider. There is currently much discussion in the industry about building gigawatt-level (gigawatt) and even multi-gigawatt data centers, and in just the past 12 months, we have added over 2 gigawatts of new capacity. We continue to expand our data center capacity faster than any competitor. Now, every Azure region has achieved "AI first" deployment. All our regions now support liquid cooling technology, enhancing the versatility and flexibility of our equipment. We are driving and riding the composite "S-curve" powered by silicon, systems, and models, continuously improving customer efficiency and performance. For example, our GPT-4.0 series models have the highest inference token processing capacity. Through software optimization alone, we have achieved a 90% increase in the number of tokens processed per GPU compared to a year ago. In addition to the AI fleet, we are also continuously building our Commercial Cloud to meet specific customer needs regarding data localization and sovereignty. This quarter, we launched the Microsoft Sovereign Cloud, the most comprehensive solution in the industry, covering both public and private cloud deployments. All these innovations are driving our strong performance. This quarter, we again saw accelerated growth from migration business. For example, Nestlé migrated over 200 SAP instances, more than 10,000 servers, and 1.2PB of data to Azure with almost no business interruption. This makes it one of the largest and most successful migrations in business history. The next important accelerator in cloud computing will be quantum computing, and I am very excited about our progress. In fact, just earlier this month, we announced a partnership with Atom Computing to deploy the world's first operational version of a Level 2 quantum computer. This is exactly how we formulate our investment thinking: planning strategies on a ten-year basis while making tangible progress every quarter The next layer is Data, which is the foundation of every AI application. Microsoft Fabric is becoming a comprehensive data and analytics platform for the AI era, covering everything from SQL to NoSQL to analytical workloads. It continues to maintain growth momentum, with revenue increasing by 55% year-on-year, and has over 25,000 customers. This is our fastest-growing database product in history. Fabric 1.0 can scale and run across all databases and cloud environments, including semantic models from Power BI, making it the ideal source of knowledge and semantics for AI applications and context engineering. Databricks and Snowflake on Azure are also accelerating their growth. Cosmos DB and Azure PostgreSQL are providing large-scale support for mission-critical workloads. For example, OpenAI uses Cosmos DB in the hot path of every ChatGPT interaction to store chat logs, user profiles, conversation states, etc., while Azure PostgreSQL stores the metadata necessary for ChatGPT's operation and important information relied upon by the OpenAI developer API. This year, we launched Azure AI Foundry to help customers design, customize, and manage scalable AI applications and agents. Foundry provides top-tier tools, management capabilities, observability, and built-in trustworthy AI controls, as customers increasingly wish to use multiple AI models to meet specific performance, cost, and use case requirements. With Foundry, they can configure inference throughput at once and apply it to more models than any other hyperscaler, including those from OpenAI, DeepSeek, Meta, XAI's Grok, and upcoming models from Black Forest Labs and Mistral AI. Just this year, we simultaneously deployed 15 models from OpenAI on Foundry (sim-shipped), providing customers with state-of-the-art models that are deeply integrated with our infrastructure and tools, available the same day. We are also seeing an acceleration in the adoption of the newly launched Foundry Agent Service, with 14,000 customers currently using the service to build agents capable of executing complex tasks automatically For example, Nasdaq is using Foundry to build smart agents that help clients prepare for board meetings, reducing preparation time by up to 25%. Overall, 80% of Fortune 500 companies are already using Foundry. Just from the number of tokens provided by the Foundry API, we have processed over 500 trillion tokens this year, a year-on-year increase of more than 7 times. This is a good indicator of the platform's true diffusion (rather than just a few popular applications and services). Speaking of the application layer, these applications are gradually being embedded into our daily work and life. Our Co-Pilot application family has surpassed 100 million monthly active users in both commercial and consumer sectors. In terms of the broader use of AI features in our products, our monthly active users have exceeded 800 million. Microsoft 365 Co-Pilot is becoming a new way for organizations to work, manage workflows, and produce outputs. This quarter, we launched the largest update to Microsoft 365 Co-Pilot to date, integrating chat, search, notebook creation, and smart agents into an intuitive structure. With this innovation and continuous improvement of the product, we are seeing real growth momentum. Customers continue to adopt Co-Pilot at an unprecedented pace, as reflected in our weekly retention rates. We set a new record for the number of new seats added this quarter, achieving the highest number of customer repurchased seats in a single quarter since launch. For example: Barclays will deploy Microsoft 365 Co-Pilot to 100,000 employees globally, having successfully completed an initial deployment to 15,000 employees; UBS will expand its initial pilot deployment to 55,000 employees to all employees; Adobe, KPMG, Pfizer, and Wells Fargo have all purchased over 25,000 seats this quarter. Tens of thousands of organizations have used our researchers and analysts' deep reasoning agents within the first few weeks of launch, and we have also introduced group-level agents such as "Facilitator" and "Interpreter" in Teams, which can generate real-time translation notes during meetings. Hundreds of partners, such as Adobe, SAP, ServiceNow, and Workday, have built third-party agents integrated with Copilot and Teams. We are also seeing more and more customers using Copilot Studio to extend Microsoft 365 Copilot and build their own agents. This year, customers created 3 million agents using SharePoint and Copilot Studio. Through Copilot Tuning, they can easily create agents fine-tuned based on their company data, workflows, and styles, which can reflect their unique tone, language style, and expertise. We are also seeing particularly noticeable adoption trends in certain roles and functional areas, with developers being one of them. GitHub Copilot continues to maintain strong momentum in integrated development environments (IDEs), with its agent mode and new forms such as "coding agent" explicitly executing developer tasks. We currently have 20 million GitHub Copilot users. The number of GitHub Copilot enterprise customers has increased by 75% year-on-year, as companies begin to customize Copilot to match their own codebases, with 90% of Fortune 100 companies already using GitHub Copilot. More broadly, usage and the number of code repositories on GitHub have exploded due to AI. In the past year, the number of AI projects on GitHub has doubled. AI coding projects and AI programming agents (including Cloud Code, CodeX, Cursor, GitHub Copilot, etc.) are driving more pull requests and repository creations. Our code review agents are heavily used on the platform, performing millions of code reviews each month. In the healthcare sector, Dragon Copilot has performed particularly well. Customers recorded over 13 million doctor-patient consultation records using our AI environment solutions this quarter, a nearly 7-fold increase year-on-year. For example: At Mercy Health System, over 1,000 doctors are using Dragon Copilot to reduce paperwork burdens to focus on providing better care services; They have saved over 100,000 hours so far and plan to roll it out to 5,000 healthcare personnel; As one doctor said, "This is the best thing that has happened in my 10 years of practicing medicine." In the security field, we are the industry's first agent service provider to help defense personnel automatically respond to a large number of security and IT tasks. More broadly, AI is driving a fundamental transformation in the business application market, with customers shifting from traditional systems to agent-driven business applications. Dynamics 365 has gained market share this fiscal year, attracting customers from various industries: Verizon uses it for sales; Domino’s Pizza Group uses it as an ERP system; 1-800-Flowers uses it as a contact center. In consumer applications, we are continuously driving multi-platform innovation. This Monday, we introduced Copilot mode in the Edge browser. The revitalization of the browser space is very exciting. Copilot mode integrates Copilot Composer, chat, discovery, search, and operations, creating the next-generation browser for the AI era. Our Copilot consumer applications continue to show strong usage growth and high-quality interactive sessions. We are bringing Copilot to every Windows 11 PC. With Copilot Vision, you can share your screen with Copilot for real-time insights and assistance. As Windows 10 stops support in October, we are ready, thanks to Windows 11 and Copilot Plus PC, we provide customers with compelling security assurances and AI functionality value. Speaking of security, it not only supports our cloud and AI infrastructure, but also supports our Copilots and Agents. In the past year, we have launched over 100 new features. Just last week, we added a modern data lake to our SIEM (Security Information and Event Management) product Microsoft Sentinel, aggregating customer data from our first-party tools and over 350 third-party ecosystem connectors. We are also expanding customers' existing governance, identity, security, and management systems to protect every AI agent. Entra now extends identity permission policies and access control to AI agents. Defender provides security protection for nearly 2 million general AI applications. Purview is used by three-quarters of Microsoft 365 Copilot customers to protect their data Overall, we now have nearly 1.5 million secure customers and continue to gain market share across all major categories we serve. Before I conclude, I want to talk about our two consumer businesses with a large end-user base: LinkedIn and Xbox. LinkedIn has 1.2 billion members, achieving double-digit growth in membership for four consecutive years. Overall, the number of comments on LinkedIn has increased by over 30%, and video uploads have grown by over 20%. We continue to integrate AI into every aspect of LinkedIn, including the introduction of intelligent agents in recruitment and sales. In gaming, we have 500 million monthly active users across multiple platforms and devices. This quarter, we became the largest publisher on the Xbox and PlayStation platforms, successfully launching "Forza Horizon 5" and the remastered "Oblivion." The "Call of Duty" series has never been stronger, with 50 million players having played "Black Ops 6," totaling over 2 billion hours of gameplay. "Minecraft" set a record for monthly active users and revenue this quarter, largely due to the success of the "Minecraft" movie. We currently have nearly 40 games in development, so there is a lot of content coming soon. This year, game time through cloud streaming exceeded 500 million hours, and Game Pass's annual revenue approached $5 billion for the first time. In summary, we are experiencing a generational technology shift driven by AI, and I am unprecedentedly confident in Microsoft's ability to seize this opportunity, drive long-term growth, and define the future. Next, I will hand it over to Amy to introduce our financial performance and outlook. Amy Hood, Executive Vice President and Chief Financial Officer: Thank you, Satya, and good afternoon, everyone. This year, we achieved over $281 billion in revenue, a year-over-year increase of 15%, reflecting the broad strength of our products and services. Operating profit exceeded $128 billion, a year-over-year increase of 17%, while we are also investing in the vast opportunities ahead. In our largest quarter of the year, we significantly exceeded expectations, thanks to the strong execution of our sales and partner teams. As Satya mentioned, we are innovating at an unprecedented pace, bringing new value to our customers In this quarter, we achieved revenue of $76.4 billion, a year-on-year increase of 18%, and a growth of 17% at constant exchange rates. Gross profit increased by 16%, with a growth of 15% at constant exchange rates, and operating profit grew by 23%, with a growth of 22% at constant exchange rates. Earnings per share were $3.65, a year-on-year increase of 24%, and a growth of 22% at constant exchange rates. Our commercial bookings exceeded $100 billion for the first time, with a year-on-year increase of 37%, and a growth of 30% at constant exchange rates, reflecting strong performance on last year's robust base. The strong execution of our core annuity sales model, including good renewal rates, as well as the increase in the number of contracts worth tens of millions and hundreds of millions of dollars in Azure and Microsoft 365, drove this result. Commercial remaining performance obligations (RPO) grew to $368 billion, a year-on-year increase of 37%, and a growth of 35% at constant exchange rates. About 35% will be recognized as revenue in the next 12 months, a year-on-year increase of 21%. The remainder will be recognized after the next 12 months, with a year-on-year increase of 49%. In this quarter, our annuity revenue still accounted for 98%. The impact of foreign exchange (FX) on the company's total revenue, segment revenue, cost of goods sold (COGS), and operating expense growth was generally in line with expectations. Microsoft Cloud revenue was $46.7 billion, exceeding expectations, with a growth of 27%, and a growth of 25% at constant exchange rates. Microsoft Cloud's gross margin was slightly better than expected at 68%, down two percentage points year-on-year, due to the impact of scaling our AI infrastructure, partially offset by the continued efficiency improvements in Azure and M365 commercial cloud. The overall gross margin for the company was 69%, down one percentage point year-on-year, primarily due to the shift in sales structure towards Azure, as well as the lower gross margin of Microsoft Cloud mentioned earlier. Operating expenses grew by 6%, with a growth of 5% at constant exchange rates, and the operating margin increased by two percentage points year-on-year to 45%. Higher-than-expected revenue growth, combined with our focus on operational efficiency, drove margin expansion. Overall, as of the end of June, our total number of employees remained roughly the same as last year. Next is the performance of each business segment: Revenue from the Productivity and Business Processes segment was $33.1 billion, a year-on-year increase of 16%, and a growth of 14% at constant exchange rates, exceeding expectations This is mainly attributed to M365 commercial products and cloud services as well as M365 consumer products and cloud services. M365 commercial cloud revenue exceeded expectations, growing 18% year-over-year and 16% at constant currency, with two percentage points of growth coming from revenue recognition in the current period. Excluding end-of-period revenue recognition, the business trend this quarter remained relatively stable compared to the previous quarter, with average revenue per user (ARPU) growth again driven by E5 and M365 Copilot. Paid M365 commercial seats grew 6% year-over-year, with growth across all customer segments, particularly notable expansion in small and medium-sized enterprises and frontline worker products. M365 commercial product revenue grew 9%, with a 7% increase at constant currency, exceeding expectations, primarily due to higher-than-expected transactional purchases of Office 2024. M365 consumer cloud revenue exceeded expectations, growing 20% year-over-year, driven mainly by ARPU growth following the January price increase and an 8% increase in subscription users. LinkedIn revenue grew 9%, with an 8% increase at constant currency, with all businesses achieving growth, although Talent Solutions continued to be impacted by a weak hiring market. Dynamics 365 revenue grew 23%, with a 21% increase at constant currency, driven by strong execution of the core annuity sales model, which propelled growth across all workloads. The department's gross profit increased 16%, with a 15% increase at constant currency, and the gross margin slightly improved, benefiting from previously mentioned efficiency gains, despite delivering more AI features in products and expanding AI infrastructure. Operating expenses grew 7%, with a 6% increase at constant currency, and operating profit grew 21%, with a 19% increase at constant currency. Next is the Intelligent Cloud segment: Revenue was $29.9 billion, growing 26% year-over-year, and 25% at constant currency, exceeding expectations, primarily driven by Azure and our on-premises server business. Azure and other cloud services revenue grew 39%, far exceeding expectations, mainly due to accelerated growth in our core infrastructure business, especially among large customers. Just a reminder, new cloud and AI workloads are built and scaled using our extensive services. Revenue from Azure AI services was generally in line with expectations. Despite launching more data center capacity this quarter, demand still exceeds supply. In the local server business, revenue decreased by 2% year-on-year and by 3% at constant exchange rates, but it was better than expected, mainly driven by transaction-based purchases, which have a higher proportion of current income recognition. Revenue from enterprise and partner services grew by 7%, increasing by 6% at fixed exchange rates, with growth in enterprise support services partially offset by a decline in industry solutions. The gross profit of this segment increased by 17% in USD terms and by 16% at fixed exchange rates; the gross margin decreased by 4 percentage points year-on-year, mainly due to the expansion of our AI infrastructure, partially offset by the previously mentioned efficiency improvements in Azure. Operating expenses grew by 6%, increasing by 4% at fixed exchange rates; operating income grew by 23%. Now looking at the "More Personal Computing" segment. This segment's revenue was $13.5 billion, a year-on-year increase of 9%, mainly due to the better-than-expected performance of Windows OEM and Xbox content and services. Windows OEM and device revenue grew by 3% year-on-year, better than expected, as inventory levels remained high. Search and news advertising (excluding TAC) revenue grew by 21%, increasing by 20% at fixed exchange rates, benefiting from continued growth in search volume and revenue per search, along with about an 8 percentage point favorable impact from third-party partners, including the effect of a low base. In gaming, revenue grew by 10%. Xbox content and services revenue increased by 13%, growing by 12% at fixed exchange rates, mainly due to the better-than-expected performance of first-party content and Xbox Game Pass. The gross profit of this segment grew by 15%; the gross margin increased by 3 percentage points year-on-year, with improvements across all businesses. Operating expenses grew by 4%, increasing by 3% at fixed exchange rates; operating income grew by 34%, increasing by 33% at fixed exchange rates, reflecting our ongoing strategy to prioritize high-margin opportunities. Next, back to the company-wide level. Capital expenditures were $24.2 billion, including $6.5 billion in finance leases, where we recognize the full value upfront when the lease begins. Cash expenditures for property, plant, and equipment (PP&E) were $17.1 billion. The difference mainly comes from finance leases. More than half of the expenditures are used to support the monetization of long-term assets for the next 15 years or more, with the remaining expenditures mainly for servers (including CPUs and GPUs), driven by strong demand signals. Operating cash flow was $42.6 billion, a year-on-year increase of 15%, mainly due to strong billing and collections in the cloud business, partially offset by higher vendor payments. Free cash flow for the quarter was $25.6 billion. Other income and expenses were negative $1.7 billion, mainly due to losses from equity method investments. Our effective tax rate was approximately 17%. Finally, we returned $9.4 billion to shareholders through dividends and stock buybacks, bringing the total return for the fiscal year to over $37 billion Now let's turn to our forward-looking outlook. Unless otherwise specified, my comments on the full year and next quarter are in U.S. dollars. First, let's talk about the full-year outlook for fiscal year 2026: Foreign Exchange (FX): Assuming current exchange rates remain stable, we expect FX to drive full-year revenue growth and cost growth of about two percentage points, and to drive operating expense growth of one percentage point. Building on our strong momentum over the past year, we expect to achieve double-digit revenue and operating income growth in fiscal year 2026. We will continue to invest in capital expenditures and operating expenses to capitalize on our leading position in commercial cloud, strong signals of demand for cloud and AI, and the vast opportunities presented by our large contract backlog. As mentioned last quarter, capital expenditure growth will slow in fiscal year 2026 compared to fiscal year 2025, with a higher proportion of short-term assets. Due to the delivery of more new capacity in the first fiscal quarter, including large financing lease projects, the capital expenditure growth rate in the first half will be higher than in the second half. We remain focused on achieving revenue growth and improving operational flexibility. Therefore, we expect the full-year operating margin to be roughly flat compared to last year. Finally, we expect the effective tax rate for fiscal year 2026 to be between 19% and 20%. Now for the outlook for the first fiscal quarter: At current exchange rates, we expect FX to drive total revenue growth of two percentage points. By business segment, we expect FX to drive revenue growth of approximately 3 percentage points for "Productivity and Business Processes," and about 1 percentage point for "Intelligent Cloud" and "More Personal Computing." In terms of commercial orders, we expect FX to drive growth in COGS (Cost of Goods Sold) and operating expenses of about one percentage point. We expect commercial orders to achieve solid growth, supported by an expanding pipeline of upcoming contracts. Order growth will continue to be driven by our core annuity sales model and customers' long-term commitments to our platform. It is important to note that the timing of large long-term contracts for Azure is uncertain, leading to some volatility in quarterly order growth rates. The gross margin for Microsoft Cloud is expected to be around 67%, a year-over-year decline primarily due to the impact of our ongoing expansion of AI infrastructure. We expect capital expenditures in the first quarter to exceed $30 billion, reflecting the strong demand signals we are seeing. It is important to note that the timing of cloud infrastructure construction and financing lease deliveries may lead to fluctuations in quarterly spending. Next, let's look at the guidance for each segment: Productivity and Business Processes: We expect revenue to be between $32.2 billion and $32.5 billion, a year-over-year growth of 14% to 15%, with FX contributing about 3 percentage points. M365 Commercial Cloud: Revenue is expected to grow by 13% to 14% at constant currency, with business trends expected to be roughly in line with the previous quarter The growth of Average Revenue Per User (ARPU) will continue to be driven by the E5 version and M365 Copilot. Revenue from M365 commercial products is expected to achieve mid to high single-digit growth. It is important to note that M365 commercial products include Windows commercial on-premises components and one-time purchases of Office within the M365 suite, and the revenue recognition for these items has cyclical fluctuations. Revenue from M365 consumer cloud is expected to grow by around low 20%, primarily driven by the price increase in January. For LinkedIn, we expect its revenue to achieve high single-digit percentage growth, while for Dynamics 365, we anticipate revenue growth to be in the high teens percentage, with all workload areas continuing to show growth. For the Intelligent Cloud business, we expect revenue to be between $30.1 billion and $30.4 billion, with a year-over-year growth of between 25% and 26%, of which about one percentage point of growth comes from the favorable exchange rates mentioned earlier. Azure will continue to be the main driver of revenue growth, with its year-over-year growth rate potentially fluctuating between quarters, depending on the timing of capacity delivery and its go-live dates, as well as the revenue recognition methods under different contract structures during the period. For Azure, we expect revenue growth in the first fiscal quarter on a constant currency basis to be around 37%, driven by strong demand for our service portfolio on a large base. Even as we continue to bring more data center capacity online, we still expect to face capacity constraints throughout the first half of the fiscal year. In our on-premises server business, we expect revenue to decline in the low to mid-single digits, due to customers continuing to migrate to cloud services and more shifts towards personal computing. We expect this segment's revenue to be between $12.4 billion and $12.9 billion. Windows OEM and device revenue is expected to decline in the mid to high single-digit percentage. We expect the high inventory levels at the end of the fourth fiscal quarter to gradually decrease throughout the quarter (in Windows OEM), although the range of potential outcomes remains wider than normal. Device revenue is expected to decline. The growth of X-Tech advertising revenue from search and news ads (excluding TAC) is expected to be in the range of 10% to 15%, slowing compared to the previous quarter, as the growth rate returns to normal, following the gradual decline of favorable factors from third-party partners. Growth will still be driven by search volume and revenue per search on Edge and Bing. Overall search and news advertising revenue growth is expected to be at a low double-digit (10%-12%) level. In the gaming business, we expect revenue to decline in the mid to high single-digit percentage Due to the strong performance in the same period last year as a comparison benchmark, we expect Xbox content and services revenue to decline in the mid-single-digit percentage range. Now back to the company's overall guidance expectations. We expect cost of goods sold (COGS) to be between $24.3 billion and $24.5 billion, an increase of 21% to 22% year-over-year. Operating expenses are expected to be between $15.7 billion and $15.8 billion, an increase of 5% to 6% year-over-year. Other income and expenses are expected to be negative $1.3 billion, primarily due to equity method investments. Please note that we will not recognize gains or losses on equity method investments based on market value changes. Finally, we expect the effective tax rate for the first fiscal quarter to be between 19% and 20%. To summarize: We achieved double-digit revenue and operating profit growth this fiscal year, exceeding our commitment made a year ago regarding FY25 operating profit margins. We will continue to focus on investing in safety, quality, and AI platform and product innovation to bring value and opportunities to our customers. We are very excited about fiscal year 2026. Alright, now let's move into the Q&A session, Jonathan. Q&A Session Host: Q&A session Jonathan Neilson, Vice President of Microsoft Investor Relations: Thank you, Amy. Now we will enter the Q&A session. Out of respect for those participating in this conference call, we ask each questioner to ask only one question. Host, please repeat the instructions for asking questions. Host: Thank you. (Operational instructions) Our first question comes from Keith Weiss of Morgan Stanley. Please go ahead. Keith Weiss: Thank you for answering my question, and congratulations on such an outstanding performance at the end of fiscal year 2025. I have been following Microsoft for many years, and I don't think I have ever seen a quarter where everything aligns perfectly like this, so congratulations to you. (Technical question) Thank you for calling Income Conferencing. The next available operator will be with you shortly. Thank you for calling Income Conferencing. The next available operator will be with you shortly. (Unidentified speaker): ——The results of these workloads are very valuable for us in building products and platforms, and then from a broader perspective, or over time, these will spread widely. In fact, one of the metrics that Amy and I track is not just the usage of core applications, but also the status of those second-tier applications that are under development. (Unidentified speaker): I can't hear clearly. The audio is currently missing. Microsoft Chairman and CEO Satya Nadella: So, this somewhat addresses your question, Keith— as long as we have leading applications to shape the platform, and then see broader diffusion, in a sense, we are seeing progress on both fronts simultaneously. Therefore, I am very optimistic about the future development prospects. Jonathan Neilson: Thank you, Keith. Host, please go to the next questioner. Host: The next question comes from Mark Moerdler of Bernstein Research. Please go ahead. Mark Moerdler: Thank you. I also want to congratulate you, it's truly amazing. I didn't know how you would surpass last quarter's results, but you did, congratulations. Thank you for answering my question. Satya and Amy, we have been in the generative AI revolution for over two years now, and we are still in the early adoption phase and accelerating. What do you think is the best way for software companies to profit from AI? Do you think more general horizontal applications like Microsoft 365 Copilot or Dynamics CRM Copilot will commercialize differently than some more targeted agent capabilities? Additionally, how do you view the future trend of SaaS AI profit margins? Thank you. Satya Nadella: Okay, I'll start answering, Amy you can chime in anytime. If we broaden our perspective beyond just the SaaS category, I think it can be likened to the transition from servers to the cloud, which is essentially an expansion of server usage, right? That is to say, we are using more servers, but in the past, deploying servers, scaling, and managing them required a lot of specialized knowledge, capital, and time costs. So the market size was limited. But with the cloud, you can purchase flexibly, scale elastically, and the required expertise has been greatly reduced. So, the scale can expand by several orders of magnitude. And what is happening now is similar. If you agree with the notion that "intelligence = log(computational power)," it means that computation will continue to grow, and you must use it with maximum efficiency to continuously generate intelligence. So, aside from infrastructure, how is this intelligence manifested? I actually mentioned some earlier when answering Keith's question, such as how infrastructure shapes, how the data layer is constructed, how application servers are built—these are all classic categories of infrastructure, and they will all scale by one to two orders of magnitude. In fact, another metric we are tracking now is that each GPU requires corresponding storage and computing power, and this ratio is growing exponentially At the application layer, SaaS applications are integrating agents and chat interfaces, injecting intelligent capabilities, and are also building autonomous agents. Agents are essentially like database applications, but they are increasingly being used for user interactions. I think GitHub Copilot is a great example. It started as a code completion feature in the IDE, then we added a chat interface, and later introduced an agent mode. Now we have launched a fully asynchronous autonomous agent. These four forms are now all part of GitHub. Moreover, all other tools that work with code are also connecting to more and more GitHub repositories. So if you ask me about GitHub's commercialization logic, we have the opportunity to profit through GitHub Enterprise, and we can also consider various forms of GitHub Copilot for further revenue generation. This logic also applies to Microsoft 365 and Dynamics 365. You must open up your data layer, business logic layer, and user interface layer for broader expansion. As long as you do this, usage will increase, which is exactly what we see in our performance results. Amy Hood: Mark, I think the various layers you mentioned are actually seeing similar commercialization tools at work in this transformation, right? There is a logic of charging per user, as well as a user tiering approach, sometimes these tiers are linked to usage, and sometimes it's purely a usage-based billing model. I think you will continue to see a mix of these models, especially as AI model capabilities continue to improve, and teams want better control over usage and to choose the most suitable model for specific tasks. I believe this fusion of pricing and business models will persist in the future. Mark Moerdler: Thank you for your answers, I appreciate it. Satya Nadella: Thank you, Mark. Host, please take the next question. Host: The next question comes from Karl Keirstead of UBS. Please go ahead. Karl Keirstead: Okay, thank you. Satya and Amy, this is the second consecutive quarter that Azure has significantly exceeded expectations, and it sounds like this is mainly due to the acceleration of on-premises to Azure migration activities. I would like you to talk about whether there are two or three specific catalysts driving these migrations in your recent conversations with customers. How durable do you think this trend is? Thank you. Microsoft Chairman and CEO Satya Nadella: Yes, I think there are three things happening here. The first is migration. A good example is Nestlé, which I mentioned in my speech; they migrated their SAP instances along with a large amount of related data and servers. This is a fairly typical example, whether it's VMware migration, SAP migration, or our own server migration, the overall situation is quite healthy. It turns out we are far from reaching the finish line; we are at most in the middle stage. The second thing happening is the scaling of cloud-native applications. This doesn't even include all the AI content; take typical cloud-native e-commerce companies as an example, their businesses are growing significantly. Some of them were not on Azure before, but now more and more are migrating over, possibly initially for AI, but now their reasons for staying on Azure go beyond just AI. For me, this is also a very important factor in the overall growth trend of Azure. Of course, there are also new AI workloads. These three things are somewhat overlapping and mutually reinforcing, and they are the main drivers of our growth. Analyst: Understood, thank you. Microsoft Vice President of Investor Relations Jonathan Neilson: Thank you, Karl. Host, please go to the next question. Host: The next question comes from Brent Thill of Jefferies. Please go ahead. Brent Thill: Satya, back to the overall strong performance this quarter, was there anything that particularly surprised you, or something you didn't anticipate that really happened this quarter? The extent of the earnings beat, I think, shocked a lot of people. Microsoft Chairman and CEO Satya Nadella: Well, Brent, I don't know if there was anything that really surprised us, but I think we noticed that when we are building these AI applications ourselves, the platform overall is no longer just at the stage of "here's a model, here's an API, go call it." In a sense, that was the state of technology about a year ago, and now you are actually seeing very stateful application patterns emerging, which require quite a bit of rethinking of the entire application stack. For example, even the storage layer has a high demand for complexity. How many indexes do you actually want to build? Or how many indexes do you want to build through preprocessing so that your prompt engineering, or what I call "context engineering," can be better and of higher quality? So I think these are all evolving. When I look at products like Azure Search, Fabric, and Cosmos DB, the frameworks around them have become robust enough to build truly serious applications What excites me is that whether internally or externally, the learning curve of the entire tech stack, the speed of diffusion, and the speed of application building are all very fast. I always think back to when relational databases emerged, it took time for everyone to build an ERP system, but now, based on what I believe is a developing maturity, we are rapidly building quite complex applications. Microsoft Vice President of Investor Relations Jonathan Neilson: Thank you, Brent. Microsoft Chairman and CEO Satya Nadella: Host, please proceed to the next question. Host: The next question comes from Raimo Lenschow of Barclays. Please go ahead. Raimo Lenschow: Very good, thank you. I also want to congratulate you. I have a question about Copilot; as a loyal user of Barclays, I am obviously very satisfied. If you think about it carefully, we all realize that Copilot is part of AI, but data is becoming increasingly important, and from there we can start thinking about agents. Satya, when you communicate with customers, have you seen them recognize that Copilot is actually just the starting point, and from there the scope will become broader? Thank you. Microsoft Chairman and CEO Satya Nadella: Yes, I think that's correct. Even within Copilot, I believe you see that, right? You now have analysts and researchers—just to name two examples, and of course all the third-party agents, so yes, things go far beyond a simple request-response model. It is actually generating applications that will perform tasks and then provide feedback, but the user interface is still very important. Even for asynchronous work, you need a UI to orchestrate these asynchronous tasks, to monitor asynchronous work, you also need a UI. This UI may be different, it may not be a chat interface, and of course you also need a way to check what the asynchronous work actually is. Take GitHub, which I mentioned earlier, as an example. Even if you are not using GitHub Copilot to create code submissions or pull requests, interestingly, we see a significant increase in the usage of GitHub Copilot's code review agents, even if you might have written the code using Cloud Code or other tools. So I think that's exactly what's happening with all these systems, so you are absolutely right, all work begins with some sort of more chat-oriented user interface, but it quickly transcends that, and you can see this trend in M365, Dynamics 365, and GitHub Analyst: Thank you. Microsoft Vice President of Investor Relations Jonathan Neilson: Thank you, Raimo. Host, please proceed to the next question. Host: The next question comes from Kash Rangan of Goldman Sachs. Please go ahead. Kash Rangan: Hello, thank you very much, Amy. I want to mention that I remember a few quarters ago, you said you had reached a stage where you could accelerate Azure's growth while slowing down capital expenditures (CapEx). You have indeed done that. So what does the outlook look like? When I see the capital expenditure guidance for the upcoming quarter, I certainly view this as a positive signal for the health of your cloud services business. However, how should we view the relationship between the capital expenditure curve and Azure's growth rate over the next few years, especially considering Satya's comments about the AI stack consuming more and more infrastructure? Are we at a stage where we must continue to do this and then magically wait for inference and application launches to create a richer gross margin structure? Thank you very much for your comments, and congratulations on your strong performance this quarter. Microsoft Executive Vice President and Chief Financial Officer Amy Hood: Thank you, Kash. Let me first recap that when you consider my comments on full-year capital expenditures and the guidance of over $30 billion for the first quarter, you need to understand that we have $368 billion in contract backlog. We need to deliver not only on Azure but also cover the breadth of the entire Microsoft Cloud. Regarding confidence in return on investment (ROI), growth rates, and their correlation, I am very confident that our spending is highly correlated with the contract business we need to deliver on our books, and we need our teams to execute capacity builds quickly and effectively. You see— we have talked about growth rates declining year-over-year, but fundamentally, our investments, particularly in short-term assets like servers, GPUs, CPUs, and network storage, are indeed highly correlated with the backlog orders and demand curve we are seeing. I mentioned in January that I thought supply and demand conditions would improve by June. Now I say I hope they will improve by December. This is not because we are slowing down capital expenditures; although spending is accelerating, we are also trying to pre-lease and get CPUs and GPUs into systems as quickly as possible, but demand is still growing. So, Kash, I am not overly focused on picking a point where capital expenditures intersect with revenue growth. I am more focused on building backlog orders, business growth, and capacity delivery, and from our current situation, the return on investment for doing so is very good. Therefore, I do not want everyone to overly focus on so-called inflection points because when you are in an expansion phase like this, picking a data point often means you will be overly conservative in gaining market share and winning competition Therefore, I tend to focus more on business growth. Microsoft Chairman and CEO Satya Nadella: Yes, I think another point, Kash, which I mentioned in previous earnings reports, is the difference between hosters and hyperscalers lies in software, and this is the same here. The example of GPT-4.0 I mentioned is entirely software, and even the optimizations over the past year are the same. So we know how to leverage software skills to multiply any hardware. That is the source of output enhancement. But as Amy said, when you really start building data centers, you don’t want to do it serially; you want to push forward on all these fronts simultaneously, and this approach will create a compound effect over time. Microsoft Executive Vice President and CFO Amy Hood: I do feel that when Satya talks about the software layer, he is linking it to the compound S-curve. I want to remind everyone that this is exactly what we saw in the previous cloud transformation process. We operated in this way, and the same skills and logic, just at a faster pace, will apply to this transformation. Kash Rangan: Sounds very encouraging, thank you very much. Microsoft Vice President of Investor Relations Jonathan Neilson: Thank you, Kash. Host, please proceed to the next question. Host: The next question comes from Michael Turin of Wells Fargo. Please go ahead. Michael Turin: Hello, thank you very much for answering questions, and congratulations on coordinating your metrics so well. Amy, regarding margins, I was impressed to hear that you expect operating margins to remain stable next year, even as you accelerate investments in Azure and more AI-focused products. Can you elaborate on how you are managing these trade-offs and offsetting some of the additional expenses? I’m particularly interested in any productivity gains you’ve achieved internally using AI, or any other factors you’d like to mention that support your full-year expectations. Thank you. Microsoft Executive Vice President and CFO Amy Hood: Thank you, Michael. I think the focus should be on margins. Sometimes people get very focused on cost control, thinking that it’s the driver of margins. Another driver is ensuring that you deliver an excellent, competitive, and innovative product that can win market share, as this drives revenue growth. As you may know better than I do, revenue itself and revenue growth are the lasting ways to achieve margin enhancement, and it is self-reinforcing Secondly, what I want to mention is, as I previously told Kash, both Satya and I have mentioned applying all our skills to enhance efficiency. Regardless of which layer of the stack, the compound effect of the S-curve, we are doing this work and focusing on it. At the same time, we are also continuously building, so you will see improvements in efficiency, even as we continue to invest. Of course, we also need to continuously attract top talent and focus on the largest and most promising products and opportunities in the market. When these three points are in place, with abundant energy and focused attention, I am confident in achieving profitability. But please do not misunderstand, it all starts and ends with the product, which is our true core focus and the key to our delivery to customers. Analyst: Sounds great, thank you very much. Microsoft Vice President of Investor Relations Jonathan Neilson: Thank you, Michael. The Q&A session for today concludes here. Thank you all for participating, and we look forward to communicating with you again soon. Microsoft Chairman and CEO Satya Nadella: Thank you, everyone. Microsoft Executive Vice President and CFO Amy Hood: Thank you. Host: Today's conference call is now over. You may disconnect your phone. Thank you for your participation