微软 CEO Satya Nadella 2026 年度致股东信：搭建 “前沿生态”，以 “Token 成本对应业务结果” 作为新的软件经济学

进入 “超级智能时代”，微软的核心命题不是机会有多大，而是怎么把技术转化成实际的经济成果。它的打法是建设一个 “前沿生态”：让别人在微软平台上创造的价值，多于微软自己留下的价值。同时围绕一个新指标来优化：“cost-to-outcome”，也就是每一份 Token 开销能换来多少业务结果。推理成本让边际成本重新成了软件经济的关键变量。 https://linkedin.com/pulse/my-annual-letter-building-frontier-ecosystem-satya-nadella-iujec/…

财务成绩（2026 财年） 营收 3318 亿美元，同比增长 18%；微软云 2144 亿美元，增长 27%；Azure 增长 41%，首次突破 1000 亿美元；营业利润 1552 亿美元，增长 21%。M365 Copilot 付费席位超过 3000 万，GitHub Copilot 用户达 5000 万。

战略要点：模型优先、Agent 优先，财报改为两大板块

Agents and Infra（Agent 与基础设施） · 交互方式从同步聊天变成异步 Agent，用他的话说是“宏观委派、微观引导“：人定目标，系统去执行，需要时再回来请示。M365 Copilot 已经转向能完成长时间任务的异步 Agent；新的 Copilot 应用把聊天、协作、写代码和自动驾驶式任务合并成一个入口。所有应用合起来，被定位为“新的工作操作系统“。 · Harness 与模型分离：上下文、记忆、可执行的动作都放在 Harness 里，模型可以随时替换，按任务选合适的模型，也增强了业务连续性。Foundry 提供 1.1 万多个模型（OpenAI、Anthropic、Mistral、SpaceXSI 和自研的 MAI 系列），客户数超过 10 万。 · 客户掌握自己的“学习闭环“：他认为，可以换掉“通才“模型，却不丢失企业里沉淀下来的“老员工“式经验，这是这个时代检验控制权和主权的标准。Microsoft IQ 作为上下文引擎，其中 Work IQ 覆盖超过 17 EB 的邮件、文档、会议数据；Agent 365 作为 Agent 的管控平面，像管理员工一样管理 Agent，并附带 FinOps 成本管理。 · 基础设施：本财年新增 88 个数据中心，建成 Fairwater 超级工厂；自研芯片 Maia 300 每瓦产出的 Token 比主流 GPU 系统多 40%，Cobalt 200 性能提升超过 50%。未来自研芯片在机群中的占比会继续提高。

Devices and Consumer（设备与消费者） 推理会在云端和终端之间自动切换。随着本地模型越来越强，“智能有望不再按量计费”。月活 16 亿台的 Windows 设备是部署基础，会带动更强的 PC 和“Agent 优先“的新设备。LinkedIn 会员 13 亿，Xbox 年玩家 10 亿。

责任与文化 · 责任：升级了负责任 AI 标准，和美国、英国的 AI 安全机构合作；防钓鱼的多因素认证覆盖率达 99.97%，修复了 55 万多个高危漏洞；用电 100% 由可再生能源匹配；以数据中心所在的华盛顿州 Quincy 为例，讲社区共赢。 · 文化：从“成长型思维“进一步走向“重新成为学习者“，同时做三件事：运营好关键系统、持续演进、发明下一代产品。组织上按工作本身而不是组织架构来运作，成立 Microsoft Frontier Co.，派出 6000 名专家驻场和客户共创。信的结尾引用贺拉斯的诗句：“敢于求知，现在就开始。”

My annual letter: Building the frontier ecosystem

Source: https://www.linkedin.com/pulse/my-annual-letter-building-frontier-ecosystem-satya-nadella-iujec/ Below is my annual letter, published today in ourAnnual Report 2026:

Dear shareholders, colleagues, customers, and partners:

For Microsoft, the central question as we enter this Super Intelligence era is not just how big the opportunity becomes, but how we deliver the real-world outcomes that every country, community, and company demands and deserves. That means aligning everything we do to convert technological progress into economic benefit for everyone, everywhere.

Microsoft is uniquely positioned to do this. We have a broad tech stack, diverse businesses and business models, and a platform- and partner-first approach in our DNA. That’s why we are building afrontier ecosystemthat enables others to build more value on top of and through the ecosystem than is captured inside. This is how we create broad economic surplus and deliver durable returns for investors.

After all, Super Intelligence is changing every organization’s inputs (how they work) as well as their outputs (the products and services they sell). At the same time, inference is making marginal cost an important factor in software economics as usage scales. As a result, customers will increasingly measure the cost of tokens against the outcomes they produce. We call this the “cost-to-outcome” frontier, and we are focused on optimizing it.

This shift expands Microsoft’s opportunity across every layer of the stack, from apps and agents that compress and complete tasks, workflows, and processes; to models that power agent reasoning and actions; to infrastructure that trains and serves those models and agent loops in the most efficient way. Any high-value activity or process that can be made faster, less expensive, or more effective through intelligence becomes part of that opportunity.

Based on the demand we see and the efficiencies we are driving, I am confident in the long-term returns on our investments. That confidence starts with being disciplined about what we build and who we build it for. We are investing to serve a diverse set of customers, while also creating the infrastructure to power our own differentiated and high-value first-party products like Copilot. We are building against a growing book of business tied to this secular shift.

We are bringing capacity into service faster, increasing utilization, and continuously improving performance and cost at every layer of the stack and across the stack as a whole. At the same time, we are evolving our business models to reflect the value and usage of our services. Together, these efforts help our customers apply technology to innovate and grow, while strengthening our ability to deliver long-term value to shareholders.

This is already translating intostrong financial performance. Fiscal year 2026 was another record year for the company. Revenue was $331.8 billion, up 18 percent year-over-year. Microsoft Cloud revenue was $214.4 billion, up 27 percent year-over-year. Azure grew 41 percent year-over-year and surpassed $100 billion for the first time, while Microsoft 365 Copilot reached over 30 million paid seats. And operating income was $155.2 billion, up 21 percent year-over-year.

As we look ahead, we remain grounded in three things:thewhy of our mission, the what of our strategic priorities, and the how of our culture.

OUR WHY

Our success depends on our customers’ success. We want every organization to own its own continuous learning system so it retains its core IP. And we want to empower every person, amplifying their agency and ambition.

This year, we made meaningful progress in defining this vision and building the foundation to realize it, and have seen many successful examples of intelligence diffusing across the world:

• In North America,Kraft Heinzput an intelligence platform on the factory floor across its North American sites, helping drive a 40 percent decrease in supply-chain waste, a 20 percent increase in sales forecast accuracy, and a 6 percent improvement in product yield.

• In Central America, Costa Rican dairy cooperativeDos Pinosis using a fleet of agents to decrease packaging-review errors by more than half, cutting time to market by roughly 10 days while enabling employees to focus on higher-value work.

• Across Europe, WWF divers are hunting abandoned fishing nets withGhostNetZero.ai, built with our AI for Good Lab, to read sonar with 94 percent accuracy and turn hours of review into minutes.

• In Asia,Air Indiahas resolved more than 13 million customer conversations with an agent at a 97 percent success rate.

• Engineers atBecain New Zealand are using Azure, Foundry, and their BEYON platform to make the National Geotechnical Database more accessible, so today’s buildings can be designed with the lessons learned from the 2011 Christchurch earthquake.

• And in Africa, route planners atCoca-Cola Beverages Africaare getting up to 90 minutes back every day as route-planning agents run across 15 markets and 10,000 users.

OUR WHAT

Our opportunity

Realizing this opportunity at scale requires us to build differently across every layer of the tech stack. That starts with making strategic choices about where we invest, what we build, and how each part of our portfolio comes together.

The tech stack we are building is model-forward and agent-first.Models will continue to become more capable, while agents will take increasingly complex workflows through to completion. As they do, our ambition is to deliver control, capability, choice, and cost advantages that compound over time for our customers and for us. And through our support for business models that span subscriptions, consumption, advertising, and hardware, we can help make Super Intelligence affordable and accessible so more people and organizations around the world can use it in the way that’s best for them.

Let me walk through how we are approaching this across our two new reporting segments: Agents and Infra and Devices and Consumer.

Agents and Infra

Computing has historically been interactive. A user moved from app to app or tool to tool to get something done. Today, it is becoming more adaptive and proactive. With agents, we can now macro delegate and micro steer. In other words, we define a goal and the system works on our behalf, returning for direction and judgment when needed.

It’s why we evolved Microsoft 365 Copilot this year, from synchronous chat assistants to asynchronous agents that can complete complex, long-running tasks. We surpassed 30 million paid seats and saw strong early adoption of ournew E7 suite.

In software development, GitHub Copilot now has 50 million users, and we worked to scale the platform to meet unprecedented demand.

In security, we introducedProject Perception, a multimodel agentic security system that uses agents to simulate attacks, identify threats, and implement fixes.

And in healthcare, we are on pace this calendar year to document more than 100 million patient encounters withDragon Copilot, helping clinicians automate workflows so they can spend more time with patients, not paperwork.

We increasingly think of all our appstogetheras a new operating system for work that transcends traditional software categories to understand user intent and act on it. As our apps move toward agent-first capabilities that can be called and orchestrated in the flow of work, the system becomes more valuable and more deeply embedded. Just last month, we announced the first version of this, anew Copilot appthat brings chat, cowork, code, and autopilots into one unified experience, bringing this intelligence into the flow of human ambition at home and at work.

Underlying all this is a new multimodel system where the harness—including context, memory, and action space—is separate from the model. This architecture optimizes the cost-to-outcome frontier by allowing us to use the right model for the right job. And it strengthens business continuity and resilience because models become interchangeable.

That’s why we offer the broadest model catalog in the cloud, with more than 11,000 models from OpenAI, Anthropic, Mistral, SpaceXSI and others, as well asour own MAI family, which we expanded with more than a dozen new models this year.

Microsoft Foundry brings these models together with knowledge, tools, and enterprise controls so organizations can build their own learning loops that know their business context and take action safely at scale. This year Foundry surpassed 100,000 customers.

The world’s most valuable data resides inside customers’ systems. Every customer should be able to use that data to build agentic systems that improve continuously whileretaining control of their IP. And they should be able to switch out a “generalist” model without losing the “company veteran” expertise accumulated in their learning system. This is the defining test of control and sovereignty of this era. Customers should demand that the value created by their proprietary learning loop accrues to them, not to an external provider. That is the core value proposition of Microsoft’s approach.

For that learning system to work, agents need to understand how an organization actually operates: how decisions are made, how work gets done, and which people and information matter to each task. OurMicrosoft IQcontext engine provides that grounding across the full stack, connecting systems of record and operational and analytical data with the knowledge in everyday emails, documents, meetings, and chats. So much of the context needed to complete a business task lives beyond any single application or system of record.

The breadth and depth of this context gives us and our customers a fundamental advantage. Work IQ alone now spans more than 17 exabytes of the most valuable data, including billions of emails, documents, and chats; hundreds of millions of Teams meetings; and millions of SharePoint sites added each day. And as Super Intelligence use grows, this context layer becomes richer, creating a flywheel that continuously improves the grounding, relevance, and effectiveness of every agent customers use and build.

Agent 365, which we introduced this year, completes this platform by providing the control plane for agents. It extends a company’s existing governance, identity, security, and management frameworks to the agents they build, so they can manage agents the same way they manage people. It also provides the critical FinOps component needed to deploy agentic systems at enterprise scale.

Taken together, these tools enable customers to optimize their token spend for a given outcome while maintaining control and building their own learning loop.

Finally, in infrastructure, we are focused on maximizing the efficiency of the entire system, delivering the highest performance and utilization per dollar and per watt. We optimize across every layer—from datacenter design, to hardware and silicon, to systems software, to models—all with the goal of continuously lowering costs and increasing performance.

We expanded Azure’s footprint with new regions and added 88 new datacenters this year, including our first-of-its-kind superfactory spanning ourFairwater datacentersin Atlanta and Wisconsin. We believe long-term innovation requires long-term trust, which is why we build datacenters with our community-first commitments to affordable electricity, water stewardship, job creation, local tax contributions, skills, and nonprofit investment.

We continue to take advantage of Moore’s law and refresh our fleet, as shown by our support of the latest silicon from NVIDIA, AMD, and Intel, as well as our own innovation. We made significant progress with our own first-party silicon this year, introducingMaia 300, which delivers 40 percent more tokens per watt compared to leading GPU-based systems, as well asCobalt 200, which delivers over 50 percent higher performance compared to our first custom-built processor for cloud-native workloads.

Over the coming years, we expect our first-party silicon to power a greater share of our fleet, extending these full-stack efficiency gains across the system while ensuring our fleet always has the latest innovations from across the ecosystem. This is how we deliver the best outcomes for customers and the highest revenue throughput and margins for Microsoft.

This combination of widely used first-party agents, a multimodel system grounded in rich enterprise context, and hyperscale infrastructure is our differentiator. Our platform is designed for the realities of diffusion and adoption. That is why our stack is open, interoperable, model-diverse, and built for heterogeneity at every layer. It allows us and our customers to continuously optimize as we build this frontier ecosystem together.

Devices and Consumer

Now, let me turn to our Devices and Consumerreportingsegment, where we bring ubiquitous intelligence to billions of people every day, with expansive total addressable markets across search advertising, gaming, and Windows OEM and devices.

Two key things are changing how we see this opportunity going forward. First, inference will increasingly move between cloud and client, driving more demand for compute at the edge. Second, it will fundamentally change consumer behaviors, expectations, and economics across all end-user categories. That means we will need new business models that bring together subscriptions, consumption, advertising, commerce, and local compute to drive broad adoption and value.

Just imagine how much intelligence can effectively become unmetered as more and more capable models run locally. We are uniquely positioned to deliver on this. Workloads will run in the cloud with seamless handoff to the edge with auto-routing, leveraging new architectural techniques to run models efficiently on-device. And with more than 1.6 billion monthly active Windows devices, we have an installed base to deploy this at massive scale, driving demand for more capable PCs and an entirely new class of agent-first devices.

On LinkedIn, we continue to see strong engagement across the platform, with over 1.3 billion members and five consecutive years of double-digit member growth. Our core vision remains: create economic opportunity for every member of the global workforce. To realize it, we are focused on helping our members access the knowledge and opportunities they need to navigate rapidly changing professional and labor markets.

And in XBOX, we have the best IP in the industry and talented studios around the world. Over 1 billion players choose to play XBOX each year across console, PC, streaming, and mobile. We are excited to bring these strengths together to delight our fans and drive new growth.

Across all our consumer franchises, we are grounded in the fundamentals and core use cases while increasing their relevance in this new era.

Our responsibility

As Super Intelligence moves from experimentation to broad adoption, we are focused on ensuring its benefits and economic value are shared and translated into lasting impact for people, organizations, and communities.

At Microsoft, we take this responsibility seriously. We have long believed that we do well when the world around us does well, and that our permission to operate must be earned every day.

Earning trust is not a finite effort. It is an ongoing practice built on the choices we make, the standards we uphold, and the trustworthiness of our systems. In a period of massive geopolitical and technological change, that includes respecting customer control, strengthening collective cybersecurity and resilience, and rapidly evolving our governance as intelligence systems become more capable.

In the past few months, we have strengthened ourResponsible AI Standardfor increasingly capable and agentic systems. We are working with theCenter for AI Standards (US) and Innovation and the AI Security Institute (UK)to advance approaches to safety and evaluation. These efforts aim to build more consistent, transparent, and rigorous methods for assessing these systems as they become more powerful and widely used.

Keeping our customers safe remains foundational. Through ourSecure Future Initiative, we are embedding security across our culture, governance, and engineering practices so we can adapt as quickly as attackers do. This year, phishing-resistant multifactor authenticationcoverage reached 99.97 percent,and we remedied over 550,000critical, high-risk vulnerabilities, improving the resilience of the systems our customers rely on every day.

As a platform company, Microsoft’s responsibility extends beyond our own systems and customers. Building a trustworthy ecosystem means collaborating across sectors and borders. This year, we joined technology leaders from Africa, Asia, Europe, and North America to launch theTrusted Tech Allianceand further verifiable standards for transparency, security, data protection, and resilience. And the Microsoft Digital Crimes Unit continued working with international law enforcement agencies and other organizations toidentify and disrupt cybercrime networksaround the world.

We believe this transformation should happenwithpeople, nottothem; that it should grow their potential rather than replace it. What matters more than the technology is how people use the technology to strengthen their abilities, solve problems, and create opportunity. That means helping workers, educators, communities, and nonprofit leaders build the skills they need to participate, while protecting and amplifying the judgment, expertise, and creativity that only people can bring.

Through Microsoft Elevate, we bring together technology, skills, research, and philanthropic support to help communities, workers, and institutions participate in and shape the new economy. New programs such asElevate for Changemakersand ourpartnerships with labor and local organizationsput that commitment into practice by making no-cost tech literacy courses and credentials available to millions of skilled craft professionals and educators. And our first-of-a-kind agreement with the American Federation of Teachers and the United Federation of Teachers sets a new standard forsafe and responsible Super Intelligence use in schools.

As adoption accelerates, we also need to understand where adoption barriers remain. Access to opportunity should not depend on where someone lives, the language they speak, or the challenges they face. That’s why the Microsoft AI Economy Institute, together with top academic organizations, studies how intelligence is changing work, productivity, and economic opportunity. Its Global AI Diffusion Report highlights persistent disparities between regions. And through initiatives such asLINGUA Africa, accessibility innovations, and support for open innovation ecosystems and open-weight models, we are working to ensure more people can benefit from and shape the future.

But access alone is not enough. Technology must help people address real-world challenges and create measurable value in communities. Efforts such asProject SPARROWdemonstrate how intelligence can help expand scientific understanding of remote ecosystems and support more informed stewardship of natural resources.

Sustainably building the future involves managing the demand for energy, water, land, and other resources.This year, we matched 100 percent of our annual electricity consumption with renewable energy. We are making progress, but addressing the challenges ahead requires continued operational rigor, deeper integration, and sustained partnership.

We believe technology should strengthen, not strain, the communities and systems on which it depends. The choices we make will determine whether we earn public trust and support. And this work has to happen locally. That belief underpins ourCommunity-First AI Infrastructurecommitments. Over two decades of building datacenters in and alongsideQuincy, Washington, we have seen what success looks like. Quincy’s poverty rate was cut in half, property tax rates for residents have dropped, and a stronger tax base has helped fund community needs like a new hospital, a new library, and new public safety stations.

Like other platform shifts before it, Super Intelligence is not about tech for tech’s sake, but about what people and organizations achieve with it. At Microsoft, we will continue to measure progress not just by what we build, but by how responsibly we build it and how widely its benefits are shared.

OUR HOW

Amid this sea change, our culture matters more than ever. Growth mindset helped Microsoft move from a fixed view of itself to a company built on curiosity and reinvention. This moment asks us to apply that same mindset under more complex and ambiguous conditions. It requires unlearning the habits, assumptions, and ways of working that made us successful in the past.

We must become learners again at a time when many expected to feel like experts. And we have to do three things at once: maintain the mission-critical systems and businesses that operate at enormous scale; continuously evolve them, because nothing stays static; and invent what comes next.

Doing all three requires us to operate differently. We are becoming faster and more agile, giving our people the authority to make decisions and removing the friction that stops them from doing so. This year, we introduced new operating frameworks designed to follow the work rather than the org chart, including cross-functional cohorts and squads; faster feedback loops between engineering, the field, and customers; and new operating forums.

And to help customers accelerate their own transformation, we launched Microsoft Frontier Co., which will bring together more than 6,000 industry and engineering experts to work alongside customers—co-designing, co-innovating, and continuously improving their own systems at scale.

***

In July, as I prepared for our annual all-company kickoff meeting, I found myself reflecting on a poem written 2,000 years ago by the Roman poet Horace. He wrote about a person standing at a riverbank, waiting for the water to stop before he crossed. The poem encourages, “He who has begun is half done. Dare to know. Begin.”

We are in a time unlike any other. History is rarely shaped by those who wait for certainty. It is shaped by those willing to jump in, learn, and adapt. Each of us must have the courage to begin before every answer is known.

The next chapter of technology will not be defined by any single model, product, or company. It will be defined by the strength of the ecosystem that emerges around them. It will be defined by what each of us does with it to amplify our own ambition and agency.

By expanding access to intelligence while preserving choice, control, safety, and trust, we can ensure the benefits of Super Intelligence reach every person and every organization. That is how we convert innovation into opportunity. And together, it is the future we will dare to know.

Satya Nadella

Satya Nadella (@satyanadella): In my annual letter, I share how we’re building a frontier ecosystem to turn technological progress into real-world outcomes and economic opportunity for people, organizations, and communities everywhere.