INVESTOR NOTICE: The presentation may contain forward-looking statements, which are any predictions, projections or other statements about future events based on current expectations and assumptions. Actual results may differ materially from these forward-looking statements because of a variety of risks and uncertainties about our business, which are discussed today or described in our filing with the Securities and Exchange Commission, including our form 10-K and 10-Q. We do not undertake any duty to update forward-looking statements. Refer to our 2025 form 10-K for reconciliations of non-GAAP measures we may use when discussing our financial results. Portions of this presentation are prerecorded.

(Video segment.)

SANDRA E. PETERSON: Hello, I’m Sandra Peterson, the Lead Independent Director of Microsoft’s board of directors. We appreciate this chance to connect with our shareholders around the world. In that spirit, we use Microsoft Teams to provide a broadcast of this meeting with real-time language translation. Whether you are near or far, we thank you for joining us. After my opening remarks, we’ll hear from several leaders from across Microsoft.

Keith Dolliver, our Corporate Secretary, will conduct the business and procedural portions of the meeting. Amy Hood, our Executive Vice President and Chief Financial Officer, will review Microsoft’s financial results and business highlights. And Satya Nadella, our Chairman and CEO, will talk about how we’re leading in artificial intelligence, building technology responsibly and unlocking new opportunities. We’ll wrap up with a Q&A session with Satya, Amy, and Microsoft Vice Chair and President Brad Smith.

Last fiscal year was another year of record-breaking financial results for Microsoft. I’m thrilled that momentum has continued into this fiscal year with a tremendous first quarter. Microsoft’s leadership team and the board together are committed to sustaining that growth and advancing Microsoft’s mission. In my discussions with investors and in the lead up to this meeting, we received questions from shareholders that relate to the composition, size, and role of the Board of directors, so let me speak to those for a moment.

The board works closely with management to provide oversight and counsel related to long-term strategy, and to risks and opportunities facing the company. The board’s four standing committees provide strategic oversight for their respective areas of responsibility. Currently, three priorities receive increased focus from both Microsoft’s leadership and the board: Security, quality, and continued innovation and leadership in artificial intelligence. The board also oversees business affairs, integrity, risk management, CEO succession planning, and performs an annual evaluation of our CEO.

As lead independent director, I work to facilitate our board’s independent oversight of management, promote communication between management and our board, and support our board’s consideration of key governance matters. I also facilitate board engagement with shareholders, and every year I find great value from the conversations I have with Microsoft’s investors.

The board nominates a slate of directors it believes is the right size to operate efficiently and effectively as a group and in committees. We seek a mix of directors with complementary qualifications, expertise, and perspective to fulfill our oversight responsibilities and help guide Microsoft’s long-term success in a rapidly changing environment.

This year, Carlos Rodriguez decided not to seek reelection and will end his board service in December. During his four-year tenure, Carlos served on the audit committee and chaired the compensation committee. We thank him for his many contributions to Microsoft and the board during his service. I’m pleased to announce the nomination of John David Rainey for election to the board at this annual meeting.

John David is Executive Vice President and Chief Financial Officer of Walmart, and has experience leading companies in complex, highly competitive industries. The 2025 proxy statement provides additional information about the board and individual directors. I am now calling the 2025 Annual Shareholder Meeting to order. I’ll be serving as the chair of the meeting, and Keith will be serving as the secretary.

As chair of the meeting, I’ve adopted an agenda that will govern the order of business and the rules of conduct for the meeting. Copies of the agenda and the rules are available on the virtual meeting site, as well as the annual meeting page on the Investor Relations website. The rules of conduct also govern the Q&A session. Keith will now report on the notice of the meeting. The proxies received and present the matters to be voted on. Keith.

KEITH DOLLIVER: Thank you, Sandi. I’ll walk us through the items for shareholder consideration at this meeting. The polls are open and will close following the presentation of our business matters. If you already voted, there’s nothing else you need to do unless you want to change your vote now. We’ve collected shareholder questions since October 21st through the proxy vote website for the Q&A session later in the meeting. Shareholders can also submit questions through the virtual meeting portal. The company has appointed Broadridge to serve as the Inspector of Election, and a Broadridge representative is participating remotely.

Beginning on October 22nd, 2025, Broadridge mailed the notice of the meeting and internet availability of proxy materials to all shareholders of record as of September 30th, 2025. As a result, the meeting is being held pursuant to proper notice. Broadridge has reported that a quorum of shareholders is present to conduct this meeting based on the shareholder proxy received. Representatives of our independent auditor, Deloitte & Touche, and our board nominees are joining the meeting remotely.

Today, we have four management proposals and six shareholder proposals for you to consider. They were all described in the proxy statement for this year’s meeting. The first item is the election of directors. The following 12 individuals have been properly nominated by the board: Reid Hoffman, Hugh Johnston, Terry List, Katherine MacGregor, Mark Mason, Satya Nadella, Sandra Peterson, Penny Pritzker, John David Rainey, Charles Scharf, John Stanton, and Emma Walmsley. The board recommends a vote in favor of each nominee.

The board of directors also recommends a vote in favor of each of the following three proposals: Proposal two, providing an advisory say on pay vote to approve executive compensation as disclosed in the company’s proxy statement. Proposal three to ratify the selection of Deloitte & Touche as the company’s independent auditor for its fiscal year 2026. And proposal four, to approve the Microsoft Corporation 2026 stock plan, which allows us to continue to award equity compensation to our employees and non-employee directors to foster alignment between them and shareholders.

The next six items are shareholder proposals. These are set forth in the proxy along with the proponents supporting statement. The proxy also provides Microsoft’s response to each of these proposals, explaining why the company and the board of directors recommend that shareholders vote against each of the proposals. We have pre-recorded presentations introducing each of the shareholder proposals, which we will now present in the following order.

From the National Center for Public Policy Research, calling for a European security Program censorship risk audit. From the American Conservative Values ETF, calling for a report on risks of censorship in generative AI. From the National Legal Policy Center, calling for a report on AI data usage oversight. From an individual investor, calling for a report on data operations and human rights hot spots. From the Religious of the Sacred Heart of Mary, calling for a report on human rights due diligence. And from As You Sow, calling for a report on AI and machine learning tools for oil and gas development and production. Please play these presentations.

STEFAN PADFIELD: My name is Stefan Padfield, and I am the Executive Director of the Free Enterprise Project, which is part of the National Center for Public Policy Research. The National Center is the proponent of proposal five, which asks Microsoft to publish a report evaluating whether its European security program or ESP could expose the company to censorship risk.

This proposal is not about opposing cybersecurity. We recognize that protecting citizens from cyber-attacks is a vital public good. Rather, we are asking for transparency and risk oversight, core principles of sound governance. The European Security program partners Microsoft with European governments and NGOs that explicitly target "hate speech" and "harmful content."

While those phrases sound benign, they have been used under laws like Germany’s Network Enforcement Act or the UK’s Online Safety Act to suppress legitimate political and religious expression. When Microsoft puts its AI at the center of that effort, shareholders have a right to understand the potential reputational and legal exposure if those tools are used to silence lawful speech.

In Microsoft’s opposition statement, the board claims that it fails to see how these partnerships could chill expression, but investors have learned that the line between security and speech regulation is not always neat. Around the world, governments justify censorship under the banner of safety. Accordingly, we asked the board not to turn a blind eye to the red flag we are waving but rather exercise appropriate oversight over the predictable ways its technology can be abused.

Microsoft points to its membership in the Global Network Initiative as evidence of accountability, yet GNI’s own framework has been interpreted to include contested categories. In practice, this could mean, for example, that speech challenging the belief that a person can be born in the wrong body could be construed as discriminatory under GNI’s guiding framework.

Combined with funding from highly ideological foundations, this raises legitimate questions about whether GNI oversight is ideologically neutral. Membership in such an initiative should not substitute for the board’s independent assessment of censorship risk. History shows what happens when tech firms underestimate these issues, from the NSA surveillance controversies to the Cambridge Analytica scandal that erased $134 billion in value at Facebook.

Microsoft’s brand and market capitalization are built on trust. Transparency is the best protection of that trust. For these reasons, we urge shareholders to vote for to proposal five to ensure that Microsoft’s powerful technologies advance cybersecurity without undermining the free expression that sustains democracy and long-term shareholder value.

WILLIAM FLAIG: Good morning, shareholders, and board members. My name is William Flaig, founder and CEO of Ridgeline Research, the investment advisor to the American Conservative Values ETF, ticker symbol ACVF. On behalf of its shareholders, I move for shareholder proposal number six, requesting a report on the risks of censorship in generative artificial intelligence.

Shareholders request the board of directors to assess and issue a report evaluating how it oversees reputational, operational, legal, and other risks related to generative artificial intelligence, and its potential biases specifically against religion or political views.

George Orwell in his dystopian novel "1984" depicts a world of constant surveillance where the government watches every move and punishes thought crime. The Ministry of Truth rewrites history and society with propaganda. He states, "Who controls the past, controls the future. Who controls the present, controls the past." I asked Microsoft Copilot: "Can AI create a future like 1984?" It replied, "The risk lies not in AI itself, but in how it’s deployed."

I totally agree. Specific concerns Copilot highlighted were: AI reflects the value of its creators. It can reinforce systemic biases and suppress alternative viewpoints. Deepfakes and synthetic media can fabricate events and speeches. Algorithmic curation on social media platforms can create echo chambers and distort public perception.

Generative AI can produce persuasive misinformation at scale, challenging our ability to discern truth. Behavioral prediction and nudging can subtly influence decisions, undermining free will. And finally, AIs widely adopted without safeguards could lead to homogenized thought or creativity, and dissent are replaced by algorithmic conformity.

Microsoft’s statement of opposition states exactly existing safeguards, transparency efforts, and principled approach to responsible AI make the proposal report unnecessary. Rules are all well and good. We have seen similar rules fail at large tech companies and financial firms in the recent past. What is necessary is vigilance and a commitment to build a strong corporate culture, which protects free speech and minority viewpoints.

Such a commitment was lacking in Microsoft’s statement of opposition. This proposal is an opportunity for our company to take a leadership role on the issue of censorship and artificial intelligence, going beyond blind faith and rules and ignoring human nature. I encourage all shareholders to vote for shareholder proposal six, requesting a report on the risks of censorship in generative artificial intelligence. Thank you and have a great day.

LUKE PERLOT: Good morning. Artificial intelligence is revealing itself as one of the most transformative innovations in modern economic history. AI’s potential to improve everything from healthcare to financial services is undeniable, as are its risks. As we know, AI thrives on data, which is vast, intricate, and often sensitive.

Unfortunately, this hunger for data drives developers to seek out large quantities of information from the internet or other digital sources, some of which may not be obtained ethically or legally. Microsoft is a leading player in the AI space, thanks in part to its partnership with OpenAI. However, this partnership has also raised ethical concerns.

The two companies have faced legal challenges, notably from The New York Times, for allegedly incorporating proprietary information into its AI models without consent. Microsoft’s LinkedIn has also been sued for using customer information for AI training. Further, former NSA director Paul Nakasone sits on OpenAI’s board, intensifying worries about government surveillance and the erosion of individual privacy.

Additionally, after heavy criticism, Microsoft moved forward with its Recall feature, which takes screenshots of user activity. And in addition to this partnership with OpenAI, Microsoft has also increased emphasis on its own AI models, which accordingly increases the need for comprehensive disclosure.

Now, National Legal and Policy Center filed the same shareholder proposal last year, and it received 36% support from shareholders, which was strong indication that Microsoft shareholders were concerned that the company’s efforts to protect user privacy and its AI development might not go far enough, and since then the company has made minimal changes to its AI reporting. And as I’ve already alluded to, the company has actually doubled down on some of shareholders biggest privacy concerns that were identified in last year’s proposal. Meanwhile, the stakes have continued to increase.

After some debate, European Union announced it will roll out its stringent AI rules on schedule just as planned, and in the U.S. politicians from both parties have pushed for similar laws at both the state and federal level. And more importantly, AI equity valuations have continued to grow, yet none of the major players have established themselves as the go-to name for privacy focused AI.

This is a tremendous opportunity for Microsoft to potentially cement its status as one of the most valuable companies in the world for years to come or comprehensive AI privacy reporting would be a good first step towards that goal. For these reasons, we encourage our fellow shareholders to vote for proposal seven.

ALI AL-AHMED: Greetings. My name is Ali Al-Ahmed. I am a Saudi Arabian national living in the United States in exile.

I was 14 when the Saudi government began its campaign to silence me, putting me in prison and the entirety of my family, my brothers, nephews and many members of my family over the years. My youngest brother, Kamel, has been in Saudi prison since 2012. He has spent over 20 years of his life, most of his life, in fact, in Saudi prison for what? For speaking, for protesting.

Many friends of mine were killed, tortured and exiled and tracked by the Saudi government. The last one was Dr. Abdullah Shumari, a renowned academic, a diplomat and a political analyst and TV host, who was executed in February last year for contacting me. That was treason in the Saudi government, and he was executed for that charge.

Saudi Arabia lives currently in its worst human right conditions since its foundation in 1952. Since MBS take over in 2016, the government of MBs testified its crackdown against dissidents and killing hundreds protesters and human right activists.

Authorities now are tracking people overseas, including in the United States. They have infiltrated big tech companies, including Twitter, using that to spy on my account and hack it and get information leading to the killing of many people.

Giving this grim reality, it’s shocking that Microsoft plans to do the same by building a data center in Saudi Arabia, because it will kill people. I’m telling you, Microsoft, you will kill people if you do that in Saudi Arabia, and I’m willing to testify in court for that.

Microsoft says its project is consistent with the commitment to protecting human rights, but the anti-crime laws in Saudi Arabia says otherwise, that that state has a complete authority to get the information from these data centers. The last expose that the Israeli government was doing that with Microsoft is enough proof.

Despite Microsoft claiming to have done due diligence, the Saudi data center will have none of that. It will be open to the Saudi government. I remember many people who were tracked, Nourah al-Qahtani, a woman that I knew, 45 years old with a special needs child, who has been arrested and sentenced to 45 years in prison.

Please help us to passing resolution 8. I have lost enough friends. Thank you.

SR. MARIA TIMONEY: Good morning, members of the board and shareholders. My name is Sr. Maria Timoney, and I am representing the religious of the Sacred Heart of Mary, Eastern American area and 58 co-filers.

I come before you today as a long-term investor and person of faith, asking for your support of shareholder proposal number 9. It calls on Microsoft to assess the effectiveness of its human rights due diligence processes in preventing, identifying and addressing customer misuse of Microsoft’s AI and cloud products or services that violate human rights or international humanitarian law.

As part of a community of women religious, our lives are grounded in our mission that all may have life. Our diverse ministries work to promote fullness of life and dignity for all whom we serve. We are long-term investors in Microsoft who believe technology should be harnessed for the common good of all humanity.

When we made a decision to invest in Microsoft years ago, we believed in the company’s vision of developing technologies and services that enrich individuals and respect human dignity. Never did we imagine that Microsoft’s technologies would be instrumental in enabling a genocide in Gaza.

Under the UN guiding principles on business and human rights, companies have obligations to respect human rights. One of the most effective ways to do this is through ongoing human rights due diligence. Companies are expected to track the effectiveness of their human rights due diligence and to publicly communicate on this. This is essential for companies who want to prevent, identify and address customer misuse.

What is happening in Gaza is not an isolated case. Microsoft’s technologies have also been implicated in human rights violations in other regions, from the Uyghur region in China to Saudi Arabia and the United States. This points not to a single failure, but rather to a broader pattern of ineffective human rights due diligence.

Historically, Microsoft has been a leader in promoting and protecting human rights, even when holding to this commitment meant severing lucrative contracts. In 1986, Microsoft announced it would end its business in apartheid South Africa, citing the untenable nature of apartheid and the need to object in the most direct manner, demonstrating itself as a leader for human dignity.

As a first step, the shareholder proposal urges Microsoft to conduct an evaluation of the effectiveness of its human rights due diligence process. This simple request could help ensure Microsoft is never again complicit in genocide or apartheid.

HOLLY ALPINE: My name is Holly Alpine, and I spent nearly a decade leading global sustainability programs at Microsoft. I’m speaking in support of Proposal 10, requesting a report on the climate-related and financial risks of Microsoft’s AI and advanced technology contracts for oil and gas expansion.

Microsoft is hailed as a global climate leader, pledging to be carbon negative by 2030 and promoting AI as central to a just energy transition, saying we must establish AI guardrails that steer the world toward net zero emissions. And it must ensure all the technology it creates benefits the planet itself.

These claims underpin Microsoft’s brand, customer trust and investor confidence, including major ESG investments from firms like Blackrock and Vanguard. Yet the reality is stark. Microsoft has no such guardrails. In fact, it creates and promotes advanced technology explicitly for fossil fuel expansion across the value chain, marketed as efficiency, yet designed to extend asset lifetimes, cut costs and boost productivity, leading to faster, cheaper fossil fuel production. In Chevron’s words, this is all around just producing more oil or gas.

Microsoft has been reported to be the world’s leading oil and gas cloud provider, a multi-billion dollar business for the company, serving some of its largest customers. While Microsoft claims climate leadership on the global stage and that it’s advancing sustainability with AI, its technology is also enabling additional emissions on a scale, potentially several times larger than the company’s entire operational and value chain emissions, or the reductions it advertises. This is locking in fossil expansion, undoing global climate progress and exposing investors to growing financial, legal and transition risks.

In response, the company points to its responsible AI and energy principles, yet neither addresses this issue, which leadership acknowledged internally. Microsoft also argues it already complies with existing disclosure standards, which don’t include either side. Yet it freely promotes avoided emissions on its reports and market facing communications, and for a climate leader, defining the frontier of AI.

Compliance with frameworks that don’t capture how its technology is used is not transparency or leadership. Employees have raised these concerns for years, and some have left the company because of the lack of meaningful change.

None of the company’s rebuttals address the central issue, the climate and financial risks of selling technology that dramatically expands global fossil fuel production.

We have spoken with major ESG and pension fund investors who share our growing concerns. This proposal simply asks Microsoft to disclose its risks. A company cannot claim to lead on climate while its technology drives the very emissions it vows to eliminate, or claim transparency while concealing those risks from shareholders.

As a former sustainability employee, I know Microsoft has the talent, integrity and innovation to lead the world towards true climate alignment. I urge my fellow shareholders to vote for Proposal 10.

KEITH DOLLIVER: Thank you. Maintaining the trust of our employees, customers, partners, shareholders and the public is fundamental to Microsoft’s mission and our long-term business success. As we shared in our responses in the proxy, Microsoft takes each of these matters seriously. Indeed, through our proxy responses and extensive public reporting, Microsoft has worked to explain how we’re addressing each of the topics raised in these proposals. A few examples:

Microsoft shares how it sources data for training generative AI models in ways that are consistent with global laws and that respect privacy, safety and content.

We’ve detailed our existing human rights due diligence practices and our plans to put in place recommendations we’ve received from human rights experts to further enhance those practices.

We’ve shared our commitment to freedom of expression and to uphold our customers’ ability to create, publish and search for information through our platforms, products and services.

And we’ve published our principles on working with the energy industry.

We think our record of engagement on these topics with shareholders and other stakeholders is strong, and we hope you found Microsoft’s responses in the proxy statement helpful in making your voting decisions.

The discussion of matters for shareholder consideration and the polls are now closed. With the proposals presented and with the vast majority of shareholder votes cast in advance of the meeting, let me now share with you the preliminary results.

First, all 12 director nominees on the ballot are elected. They are therefore authorized to serve until the next annual shareholder meetings and until their successors are elected and qualified.

Shareholders also approved the remaining management proposals, affirming executive compensation, ratifying Deloitte & Touche as the company’s independent auditor, and approving the Microsoft 2026 stock plan.

Shareholders did not approve any of the six shareholder proposals.

We expect to post the details of the final voting results on these matters within four business days on our Investor Relations website and in a Form 8k that will be filed with the Securities and Exchange Commission.

And with that, we have completed the formal portion of the meeting, and the meeting is now adjourned.

Let me now hand it over to our Executive Vice President and Chief Financial Officer, Amy Hood, to discuss our business results.

AMY HOOD: Thank you, Keith. Hello, everyone, and thank you for joining us today.

As Sandi mentioned, we delivered another year of record results in FY25. Revenue grew 15% to over $281 billion. Operating income grew 17% and earnings per share grew 16%. We continue to innovate in areas that deliver increasing value for our customers, empowering them to achieve more and drive meaningful impact in their organizations.

And we are committed to growing our leadership in AI as we invest in capacity needs across Azure, first-party apps and AI solutions, our own R& D efforts and end-of-life server replacements. We also continue to invest in organic growth and maintain our commitment to capital return, which included a total cash return of $37.7 billion in FY25, up 10% from last fiscal year. In September, we announced a 10% increase in our quarterly dividend, and we continue to execute against our current $60 billion buyback authorization.

Now, I’ll share a few highlights from the past fiscal year.

Our Microsoft Cloud business surpassed $168 billion in revenue, growing 23% year over year, with growth across key services. Azure grew 34% to over $75 billion. Microsoft 365 Commercial Cloud grew 15%, and Dynamics 365 grew 19%. We delivered Microsoft Cloud gross margin percentage of 69% through the sustained efforts of our engineering, sales and marketing teams, even with the impact of scaling our AI infrastructure to meet growing demand.

Our server products and cloud services business exceeded $98 billion, up 23% year over year, with strong demand for our trusted, differentiated hybrid offerings. We continue to deliver high-value solutions that help customers accelerate their digital transformations and realize the benefits of AI.

Now, let’s turn to progress across Microsoft 365.

Our Microsoft 365 business across both commercial and consumer exceeded $95 billion, up 14% year over year, as we helped even more users around the world be more productive, collaborative and secure. We rolled out our biggest update to M365 Copilot to date, and continue to see customers adopt Copilot faster than any other new Microsoft 365 suite. Additionally, our M365 consumer subscription base grew to 89 million.

In Windows, we’re seeing momentum as customers prioritize enhanced security and best-in-class AI capabilities. Commercial deployments of Windows 11 grew more than 75% year over year, and we’re now bringing Copilot to every Windows 11 PC, natively built in as part of the Windows experience.

We continue to grow our Search business, gaining share across both Bing and Edge. With Copilot Mode now available in Edge, we’re transforming our browser into an AI experience that goes beyond traditional browsing.

In LinkedIn, revenue surpassed $17 billion and membership grew to 1.2 billion professionals, marking four consecutive years of double-digit member growth. And in Gaming, revenue exceeded $23 billion with Game Pass revenue reaching nearly $5 billion for the first time. We finished the year as the top publisher on both Xbox and PlayStation, driven by strong performance from our first-party portfolio.

Now, a few comments on our current fiscal year.

We had a solid start to Fiscal Year ‘26, as 1st quarter revenue growth of 18% was driven by Microsoft Cloud revenue, which was $49 million, growing 26% year over year. Looking forward, we believe that the highest shareholder value is created by investing for the future and created differentiated, high-value solutions for every person and organization to achieve more. We will continue to invest in security, quality and AI innovation aligned to the demand signals we see.

We are excited about FY26, as we lead in this next frontier of innovation, driving impact at scale for every customer, every partner and every community we serve around the world.

With that, please join me in welcoming our Chairman and Chief Executive Officer, Satya Nadella,

SATYA NADELLA: Thank you, Amy, and thank you to everyone joining us today for your continued commitment and investment in Microsoft.

Fifty years after our founding, Microsoft is once again at the center of a generational moment in technology, the AI platform shift. More than any transformation before it, this era of AI is changing every layer of the tech stack, and we are changing with it. We’re delivering our current platforms at scale while building the next generation, always striving to create more value for our customers, our partners and the world.

Financially, as Amy shared, it was a year of record performance. We delivered over $281 billion in revenue and over $128 billion in operating income. Looking ahead, we remain focused on three core business priorities, security, quality and AI innovation.

Security and quality are non-negotiable. Our infrastructure and services are mission critical for the world. This year, we made significant progress across both our Secure Future Initiative as well as our Quality Excellence Initiative, but we recognize that our work here is never done and that we must continually raise the bar.

And at the same time, we made major advances in AI innovation, including across two foundational area, our AI platform and our family of Copilots. When it comes to our AI platform, we are building a planet-scale cloud and AI factory, maximizing tokens per dollar, per watt, while supporting the sovereignty needs of customers and countries.

We opened new data centers across six continents this year and now operate over 400+ data centers in 70 regions, more than any other cloud provider. We’re building a fungible fleet that is continuously being optimized and spans all stages of the AI lifecycle. Just last month, we announced our new Fairwater data center in Atlanta, which connected with our first Fairwater site in Wisconsin to create the world’s first AI super factory. And we also continue to invest in sovereign cloud offerings to meet the unique data residency needs of governments and industries across the globe.

On top of this infrastructure, we introduced Microsoft Foundry this year to help customers design, customize and run their own AI applications and agents. Foundry includes access to more than 11,000 models from partners like OpenAI, Cohere, DeepSeek, Meta, Mistral, xAI, and as of last month, Anthropic. And this fall, we introduced our first in-house MAI models for text, voice and image generation, which debuted among the top in the industry leader boards.

And now, let us turn to the apps and agents we ourselves are building on this platform.

Our family of Copilots is helping people thrive at home, at school, at work, across all the high-value domains of information work, coding, security, science, health and consumer. All up, we have surpassed 150 million active Copilot users.

When it comes to information work, we have integrated chat and agentic workflows into everyday tools like Outlook and Word and Excel and PowerPoint and Teams. We introduced Agent Mode, which returns single prompts into expert quality documents, spreadsheets and presentations. And Work IQ is our new intelligence layer that enables Microsoft 365 Copilot to know you, your job, your company, inside and out.

Copilot Studio enables customers to extend the Microsoft 365 Copilot to build their own custom agents using the no-code, low-code tools. And just last month, we introduced agent 365, which extends the systems customers already count on for governance, identity, security, and management to protect every AI agent they use or create.

And in coding, GitHub Copilot is the most popular AI pair programmer now with 26 million users. And with Agent HQ, GitHub is becoming the organizing layer for all coding agents beyond just GitHub Copilot. In security. We were first in the industry to introduce agents that help defenders autonomously handle high-volume security and IT tasks, and we now have over three dozen agents in Copilot integrated across Entra, Defender, Purview, and Intune.

In healthcare, Copilot is being used to document millions of clinical encounters. And in consumer front, we’re also excited about our momentum. Copilot is now integrated across Bing, Edge, GroupMe, MSN, Windows, and Xbox. And among the many updates we made recently is Groups, which turns Copilot for the first time into a shared experience.

While we have made great progress, we also recognize that we must earn our permission to operate every day in every country and every community, with every customer interaction. That’s why we remain grounded in our mission to empower every person and every organization on the planet to achieve more. We have a real opportunity to ensure broad diffusion of this technology with real choice, real control, so that benefits are felt by everyone, everywhere. And I’m confident than ever before that we’ll meet that massive opportunity.

Next, I’m excited to share a few examples of our innovation from the past year. Thank you all very, very much.

CASSIDY WILLIAMS: Thanks, Satya. GitHub is the home for developers. It powers collaboration for open source and enterprise developers across the world. In fact, this year someone new joined GitHub every second. The fastest growth we’ve ever seen. In 2021, Microsoft created a brand-new category of product for developers when we launched GitHub Copilot, using AI to accelerate the software development process.

GitHub Copilot started as a pair programmer and has become an entire fleet of coding agents, and now it’s the single largest contributor to GitHub’s codebase. But we’re not stopping there. Developers want choice, enterprises want control, and everyone wants fewer subscriptions to manage. That’s why at GitHub Universe, we announced Agent HQ, a single place to pick agents from GitHub, OpenAI, Anthropic, Google, Cognition, xAI, and more, all with consistent governance and observability as a part of a Copilot subscription.

Let me show you how it works. Let’s say I’m a developer at an online retail store. We use GitHub for our entire software development life cycle. While I’d normally be brewing some tea right about now, I’m going to start burning down issues in our backlog. Here is a new one from Lydia, improving how the site works on mobile.

If I’m being honest, this is important, but it’s not my favorite thing to do, so I’m going to ask Copilot to start work on this. And because this is a UI task, I’m going to use the custom coding agent our team built, focused specifically on front-end changes based on our frameworks and styles. Let’s also ask Copilot to include screenshots so I can easily see the difference without running the app.

And just like that, Copilot gets to work right away, planning and writing code tests, and opening a pull request for my team to review and merge later. I can manage this task and all of my agent sessions in one place: Agent HQ. We announced Agent HQ just a few weeks ago at GitHub Universe. It’s a single place to manage all of my agent sessions as I scale my work.

Okay. Copilot has that task in hand. So let’s look at some of our security issues. Our security team created a campaign. They chose which vulnerabilities to focus on and a deadline to get it done so we have an achievable target. I am going to select all of these vulnerabilities and assign them to Copilot. Copilot will generate these fixes all in one single pull request for our team to review.

Now, I also spent a bunch of time building in VS Code, and naturally I use coding agents to help with my work there, too. Check it out here I can see all of the agent sessions related to the code I’m working on: Local chats, CLI sessions, and even sessions with third party agents like OpenAI Codex. I can even start new tasks right from here and run them in the cloud.

I just realized I haven’t written any tests for this code, so before my team calls me out in the pull request, let’s add some tests to improve our security posture even more. Our teams have developers using Copilot and a lot of different ways, and now all enterprises can see how Copilot is being used within their organization with Copilot Metrics. I can see usage statistics, trends, what models my team is using or not using, and even the programing languages that Copilot is assisting with most.

Okay. Let’s go back to where we started, Agent HQ. We kicked off a bunch of coding tasks, and we can see all of that work here. I want to take a look at the mobile task I created earlier to resolve Lydia’s issue. I want to give it some more context, and I can do that while it’s running and clarify that I only care about phone sizes right now, not tablets.

So in just a few minutes, GitHub Copilot has worked with me to remediate security vulnerabilities, start making our online retail store mobile-friendly, and improve our testing so I can stay off the naughty list this year, just in time. All of this while keeping me in control.

GitHub Copilot meets developers where they work in the CLI, on mobile, on web and inside editors like VS Code, Visual Studio, and IntelliJ and through custom agents built from the tools you already trust: Jfrog, New Relic, and more. So now developers can collaborate with your agents across the development stack. Okay. That’s what we’ve been working on here at GitHub. Now, I’ll hand it over to our M365 Copilot team to show you what’s new over there. Thank you.

CALLIE AUGUST: Thanks, Cassidy. Productivity is core to Microsoft’s mission. With Microsoft 365 Copilot, we’re enabling professionals around the world to be more productive with AI. Today, I want to show you some examples to bring this to life. Let’s start in the Microsoft 365 Copilot app with one of my favorite new functionalities: voice. Hey, Copilot, what’s on my agenda today?

COPILOT DEMO: You’ve got a busy day ahead. This morning you had a daily stand-up for Zava Corp. Up next, there’s a Contoso sales call at 1030, then a kickoff call for the M365 Copilot--

CALLIE AUGUST: Great. Do I have any emails from my boss?

COPILOT DEMO: Yes. You got quite a few emails from Aaron Friedman, your boss. The most recent one is about prioritizing finalizing the monthly business review today. She wants to ensure you are fully prepared for tomorrow’s leadership meeting and asks you to let her know if there are any issues getting it done by end of day.

CALLIE AUGUST: Okay, great. I better get started on that right away. Pretty incredible. You can see that Copilot has access to all of my emails and meetings, which is just what made those responses so thorough. This makes Copilot really feel like a collaborative thought process.

Now, what made this all so incredible is the fact that Copilot doesn’t just access your work data, it actually understands it. We call this Work IQ, and it’s the intelligence layer behind Microsoft 365 Copilot and agents that helps Copilot understand you, your job, and your company. This enables Copilot to give you more accurate, relevant, and personalized responses. I’m going to continue that workstream in chat.

Now, the first thing I’m going to do is I’m going to get a quick understanding of where the team is on this MBR. I’m going to have Copilot pull together updates from my MBRB team, and because of Work IQ, Copilot actually knows who my MBRB team is, and it’s going to instantly generate a list of core emails giving me immediate visibility into the work the team has already done.

Now, I know my team has already pulled everything together, so I need to get to work on this report. For that, I’m going to go over to Word right here, and I’m going to open up my October MBR template that’s ready to go. I have all that information that I need from my emails that were just sent, so I’m actually going to use Agent Mode Word to help me update this document.

Now, Agent Mode and Word lets you draft, edit, and format documents using natural language directly in word. When I open up Copilot, I’m actually able to turn Agent Mode on right here. Now, I’m going to give it a bit of direction. Here I’m going to ask it to update this October MBR template with the November data. I want to make sure the entire report gets updated, and it removes any October updates with recent updates from my chats and emails.

You can see Agent Mode in Word immediately gets to work, and it’s going to explain some of the action that’s taking as it goes. Now, typically, updating a document like this would take me hours. I’d be sifting through my emails, copy pasting answers, editing, but the agent can actually take that off of my plate. You can see it’s already started those updates here. It’s completely rewritten the executive summary for me, it’s done key highlights, and now it’s starting to pull in all of the performance metrics from my emails and chats.

This part is just so incredible. I can actually see it co-writing with me as it fills out this template. The best part is I always have full control. I can go ahead and make edits myself, or I could even ask Copilot to take another pass once it finishes this template. Only one more to go here. It’s just filling out the last two boxes, and this is looking pretty good to me, so I’m going to go ahead and keep this.

Now, I’m going to move over to Outlook, and I’m going to let my boss know that this is completed. Aaron asked about this earlier, so I’m just going to have Copilot draft a quick response. I’ll add the file. And it’s ready to send. In minutes I was able to get up to speed, pull together my update, and ship it off to the team. Next, we’ll kick it over to Allan to show us the latest innovations in Copilot that you can use in your day-to-day life. Thank you.

ALLAN CARRANZA: Thanks, Callie. Copilot is your AI companion built to serve you. It’s here to help you think, plan, and dream, and it does that always on your terms. This is AI that listens, that learns, that earns your trust. Most importantly, it looks out for your interests so you can focus on what matters most. Let me show you what I mean, starting with how Copilot helps us collaborate.

I come from a big family, and we’re scattered all over the country. Planning our family trips used to mean endless text threads and zero decisions. Now, with Copilot Groups, we all jump into the same chat and make all of our lives easier. Let me show you. I’ll start by giving Copilot some context about where we’re going and dates that we’re thinking of and knowing how much coordination this takes across the family, I’ll invite the rest of my family to join me in the chat.

Dad follows up. Of course, he wants to know how he’s supposed to get there from our hometown. My wife wants to know what the kids are going to do while we’re there. And as mom always does, she’s worried about what the family will eat during our vacation. Knowing how much coordination this is going to take, I asked Copilot to generate a day-by-day itinerary and divvy up tasks amongst the team.

Copilot helps us make decisions faster so we can focus on what really matters: Having fun together. Of course, collaboration is just one part of the magic. Copilot isn’t only about what it can do, it’s also about how you experience it. Speaking to Copilot is my favorite since it’s just the most natural way I connect, but it’s always nice to see what’s on the other side of your conversation, and that’s exactly where Mico comes in. Hey, Mico, what’s the weather typically like in San Sebastian, Spain, during the summer?

COPILOT DEMO: San Sebastian summer weather is pretty mild compared to the rest of Spain. It’s usually around 20°C to 24°C. That’s roughly mid 70s in Fahrenheit.

ALLAN CARRANZA: Oh, awesome. So it sounds like we’ll be able to do a lot of outdoors during the month of June.

COPILOT DEMO: Absolutely. June’s a fantastic time for all those outdoor adventures. You get that mix of pleasant temps and maybe a fresh breeze off the Bay of Biscay. Just perfect for exploring.

ALLAN CARRANZA: Mico gives Copilot a visual identity that reacts with animations in real time, matching the tone and flow of our conversation. And because you can change the appearance and voice of Mico, it becomes a companion that feels distinctly yours. Being a companion also means being there in moments that really matter, like when it comes to your health.

Health is something that’s always important to me and my family. My toddler woke up in the middle of the night with a fever, and I needed help fast. That’s the beauty of Copilot; it’s always there for me. I go ahead and ask Copilot, "My toddler doesn’t feel well. What can I do to help?"

Copilot searches across its credible sources, giving me a parent support I need in the middle of the night. Knowing I need to visit the doctor in the morning, I can ask for guidance from trusted sources and when I’m ready to take action, like finding care nearby, Copilot has got me covered. It’s health support I can trust, grounded in credible health sources and available exactly when I need it most. Now let’s talk about how Copilot shows up right where you’re already working. On Windows, Copilot is readily available anytime I need it, I can just say: Hey, Copilot.

COPILOT DEMO: Hey there Allan, what’s cooking on this fine afternoon?

ALLAN CARRANZA: So we’re going on a family camping trip. I think I’ve got my packing list ready, but my wife is convinced I’m missing something. Can you look at the Excel doc on screen and tell me that one thing that might not be on the list?

COPILOT DEMO: Yeah. Sure thing. Look like you’ve got a pretty solid list here for your camping trip. One thing that might be handy though I didn’t notice any insect nets or mosquito netting. Those can be a lifesaver if you’re camping in a boggy area.

ALLAN CARRANZA: Great suggestion. I am, Copilot. Thank you. Copilot Vision on Windows can see what’s on my screen and help me make sure I’m not forgetting anything. It’s your everyday task, like planning trips, managing schedules, and reviewing files made simpler. And when you’re online, Copilot is there to help you make sense of the web. Browsers used to mean doing all the work yourself. Typing, clicking, juggling tabs, Copilot mode in Edge changes all of that.

I’ve got what feels like 100 tabs open right now to find the best game for my new Xbox Ally X. Rather than doing all the searching myself, I’ll ask Copilot to do the search for me and recommend the best option. Copilot can search across all of my open tabs, bringing me exactly what I need and helping me make my best decision. It’s browsing that feels less like work and more like having a partner by your side. I’m so excited about all these new features that you can try today by downloading the Copilot app to your mobile device or launching it on your desktop or browser. With that, I’ll hand it over to Jonathan for Q&A.

JONATHAN NEILSON: Welcome to the Q&A portion of the meeting. I’m Jonathan Nielsen, head of investor relations at Microsoft. I’m joined by Satya, Amy, and Brad Smith, Microsoft Vice Chair and President. We want to thank you for all the questions you submitted in advance. We’ll try to get to as many as we can.

As we reviewed submissions, there were several questions on similar topics, so we’ve selected a representative question from those submitted to answer today. Let’s go ahead and get started. We received quite a few questions on the topic of artificial intelligence and the strategy around it. To start us off, Satya, what do you see as Microsoft’s role in AI?

SATYA NADELLA: Well, that’s the entire meeting. But I’d say the approach we have taken is similar to the approach we’ve taken with other big platform shifts. As a core platform company, you look at the entire portfolio of Microsoft’s products and technologies and really take a principled approach on how is AI impacting each layer.

So, for example, our infrastructure business, one of the great opportunities we now get is to build out our cloud infrastructure so that it’s AI ready for what is going to be the token factories or the intelligence factories. So that’s our number one priority. We’re obviously at scale and building out rapidly.

The layer above is that you want to use these token factories to create AI agents, so there’s a new app server being born, and that’s where things like Foundry and GitHub Copilot all come together, so that we can then enable our developers and our partners to build great agents. And then the layer above that is where the high-value domains in which we, you know, have always participated.

Information work, that’s where Microsoft 365 Copilot comes in. Software development with GitHub Copilot or Security Copilot, or what we’re doing in business applications, and of course, in the consumer space. So we’re taking a pretty full-stack approach, but the real goal is not only for us to reimagine the current categories and also build new extensions.

A great example is what we’re doing within healthcare, both with DAX Copilot is a new category of product for us. And so, those are the kinds of things that we are trying to do across the entire length and breadth of the AI industry.

JONATHAN NEILSON: That’s great. We also received a question about how much we’re spending on AI infrastructure and our level of confidence that we’ll see a return on our investment. Amy, maybe that’s one for you.

AMY HOOD: Of course. To maybe build on where Satya started, which is that if you think about the opportunity and the addressable market of AI, there’s the expansive layer at the platform, there’s the expansive app server layer where developers will build applications that benefit every industry, and maybe even new industries that get built, and then there’s the application layer and first-party needs that we have to run our Copilot across M365, GitHub, healthcare solutions as Satya talked about, and for consumers and commercial customers of all sizes.

And so, if you think about you’ve got this expanding opportunity at every layer, what it really means is we’ve seen a very large surge in demand, and we’ve been talking about, and I think it’s important to ground that this is a demand-driven spending. We often do spend to ahead of opportunity. That’s what innovation teaches us to do. But you also spend when you see demand, and it’s resonating and growing.

This is best thought of as the combination of those two. We’re investing for innovation and pushing the frontier. We’re also investing to make sure we can meet the surging demand that we’ve seen.

And I think in some ways, we’ve talked about this over the past couple quarters. Every time we think we’re getting close to meeting demand, demand increases again. And so, what you’re seeing is this capital that we’re spending is being spent to both meet and hopefully, get ahead of the demand growth we’re seeing.

And then the return we expect to earn on that, I feel quite confident on. We’ve been through these transitions before. And in fact, if you think about what makes this AI transition, in many ways, similar to the cloud transition from, I guess it’s almost been a decade and a half, two decades ago at this point, is that you’re seeing margins actually be better at this point in the AI cycle than they were at this point in the cloud transition cycle that we went through.

We’re also seeing demand happen much faster, and you’re seeing broad based demand. It’s not just from one customer, it’s from a long tail of customers and adoption. And when you see those type of signals, and you see differentiation in our products and services, whether that’s in Azure or M365, that’s where you start to see returns.

And so, I’m encouraged by the early progress. I’m encouraged by the improvements in gross margins we’re already seeing. I’m encouraged by customer demand. We have about $400 billion, as we talked about at the quarter close, of committed contracts that we needed to deliver against. And so, being able to spend these dollars, get it revenue ready, get customers the value they’re seeking from this platform, I feel very good about the return profile.

JONATHAN NEILSON: That’s great. Now our next question comes from a shareholder who asks how Microsoft is ensuring that artificial intelligence continues to enhance human capability rather than replace it, while maintaining the transparency and the trust in its use of products across the portfolio.

Maybe this one’s for both Satya and Brad.

SATYA NADELLA: Yeah, maybe I’ll start, Brad. I think we are approaching this with our mission at the center of it. What’s the mission? The mission is to empower people and organizations through technology, and in this case, it happens to be AI.

And so, the principles we have had is to put the person and the human at the center, and then have this technology be something that they can use, they can delegate to, they can steer, they can control. That’s, I think, at the end of the day, the core.

And we have not just talked about this as principles in the abstract, but we’re even reducing it to what I’ll call as engineering practices around fairness, around transparency, around security, around privacy, and building it into the tool chain.

For example, when you go use something like Azure Foundry, you have these principles encoded, even, as a set of services that not only we use, but also any third party building on our platform can use, because at the end of the day, there needs to be trust in technology, trust in AI for its broad diffusion. And so, we are taking a principled approach, but even reducing it into what I’ll call everyday engineering practice and product building, which is aligned with our mission.

BRAD SMITH: And I would just supplement that. I think in some ways, this has become, or is becoming, almost a three dimensional exercise, the first dimension being the technical guardrails, which you’ve just described. We have 400 people at Microsoft working on responsible AI, half of them full time. But the other two dimensions really involve skilling and the broader societal decision making that will take place.

On the skilling side, one of the keys to enabling people to use it to augment themselves is to help them learn how to use it. We invest in a broad based skilling, but mostly, I would say this is something for everyone. And the best way to skill yourself is to just start using the products.

This morning, I was in a meeting at eight o’clock, and there was a topic that came up. And I said, “We should really learn from this experience we had with a similar issue about seven or eight years ago. Let’s see if we can remember that.” And then I said, “Well, I’m going to ask the Researcher Agent in Copilot to produce a report.”

Fifteen minutes later, I actually had a 25-page report with 100 citations to information on the Web, graphs, put it in a Word document, sent it around. And the other people in the meeting said, “Hey, can you share the prompt,” because we’re all learning, even inside Microsoft every day, how to use AI more effectively. There’s just no substitute for that.

But then the other piece is the societal decision making. One of the questions that’s emerging is, what’s the right age for kids in schools to be using AI? Tech companies won’t make those decisions. School boards, states will make those decisions. In fact, there’s a broad conversation today about when smartphones should be available in schools.

Well, this is almost 20 years after the smartphone was really popularized by the iPhone. I think, quite rightly, people have learned from that experience and saying, let’s have these conversations now. That’s what we’re trying to do as a company, engage in them, not just think about what’s good for our revenue, but what is going to be good for society, because I think that’s how we’ll earn trust on an ongoing basis.

JONATHAN NEILSON: That’s great. Thank you. Now, we’ve received some questions about how the company decides where to locate data centers, and how we think about the potential privacy issues that may arise.

Brad, I think this is another one for you.

BRAD SMITH: Well, there’s a number of factors, as we all know, that go into where we put data centers. In the first instance, we’re building them around the world because we do need to have data centers, especially those that provide cloud services, in reasonable proximity to our customers, so that we don’t have too much latency.

But beyond that, which goes back to Amy’s point about demand, we have to have access to low-cost energy. We have to have access to connectivity. We obviously have to have access to land and zoning and permitting.

But beyond that, the privacy issue is also important. We build data centers with an eye towards the laws that will govern, in those countries, their operation. We do human rights impact assessments to ensure that we can protect not only privacy, but other aspects of human rights, and we update that on an ongoing basis. That’s enabled us to build data centers now in more than 40 countries, but we’re not everywhere, although we can serve the world everywhere there are people out of the data centers we have.

JONATHAN NEILSON: That’s great. Now, a question more about the internal effects of AI. Are we seeing AI have any effect on the company workforce? Are we seeing wide adoption?

Satya, how are we seeing AI play out internally?

SATYA NADELLA: Yeah. I mean, in fact, Brad gave the best example at some level, which is, to me, what’s most important, starting right at Microsoft, is just the broad diffusion of these tools. I always go back to the PC era and how it diffused throughout and changed work, work artifacts and workflow. Same thing is happening in AI.

When I think about the standard issue of Copilot across the company, or GitHub Copilot across the company, all these very domain-specific things like Security Copilot, they’re broadly in use. And that’s the best thing, because people are getting skilled. People are using it.

We’re also now not only having the chat interactivity with reasoning, but we now have even agents like the Researcher Agent, where you can fire and then come back 15 minutes, half hour, even an hour later, and then see the work, and then inspect it, and then, whether commit it into a repo or publish it as a document, or what have you. I think we’re all learning this new agentic workflow across the board.

And then we have some fantastic top-down cases. I know Amy has done a lot in Finance. There’s stuff in Legal. There’s stuff in HR. And customer service is a case study that Judson, if he were here, he would go on about how fantastic it’s been, and just improving the quality of our customer service.

And the thing though, what is exciting to me is to wake up in the morning and suddenly see in my inbox or on Teams channel somewhere, some case study that I never heard of, someone who’s done something in supply chain. One of my favorite is someone in our Network Operations, where we are dealing with 500+ fiber operators around the world, and completely automated the entire DevOps pipeline on being able to deal with any repair operations on fiber.

And so, those are the kinds of things that are now emerging, just because the tools are broad spread, the knowledge and the skills that Brad talked about are getting broadly spread. And so, that’s the type of transformation, I think, that’s happening at Microsoft, and quite frankly, with all of our customers.

JONATHAN NEILSON: Yeah, it’s amazing. Now, over the past few years, Microsoft has generated significant free cash flow. One shareholder wrote to ask how we decide between reinvestment, buybacks and dividends, and how the company evaluates allocating its dollars.

Amy, could you take this one, please?

AMY HOOD: Sure. I would start by saying that that is right. We have generated and continue to generate significant free cash flow. And one of the most important things, I always think, is our commitment to long-term growth, for our shareholders, for our employees, our partners, our customers. And you start with that, which I think forms the basis of how we approach, the first answer is always to look to invest in yourself, invest in talent, invest in the computing resources they need to be able to invent and grow and create.

And so, as we’ve talked about, I think even as recently as our Q1 earnings call, we talked a lot about the need to make sure we start there, especially when the opportunity is so vast. When you have a vast opportunity, and you have a really strong feeling about your competitive position in the AI platform wave, I think it’s really important that we always start with saying, let’s invest where we can generate the highest long-term shareholder return, and that is in ourselves.

Think about that primarily as being an investment in research and development, in our core engineering culture and technical acumen, to be able to build all the tools and the examples that were just discussed. It’s also an investment in the local resources that are needed, the development of data centers around the world to serve governments and customers. Being able to invest there is critical.

I put that in one big bucket, which says the first thing is long-term growth. That means invest in yourself, and that remains our number one priority.

After that, we are lucky to be able, given our growth and our position, and have been very committed to shareholder return as well. We obviously start with the dividend. Dividend is a commitment we make. And every time we make that commitment, it’s viewed as a long-term commitment, and we’ve been able to increase that dividend every year now for a while. That’s maybe the second bucket.

And then finally, because it’s a very flexible tool for us, is buyback. And I think you’ve seen that be more opportunistic, which it always will continue to be. And that’s the right place for the buyback to exist.

JONATHAN NEILSON: Makes tons of sense. Thank you. If we could shift gears a bit now, we’ve received a few questions on a topic of U.S. politics.

Brad, could you talk about how we weigh political donations or activities?

BRAD SMITH: Yeah. I think, like most companies, we have focused efforts, and we do have donations to political candidates. Most of those go through a political action committee, which, under the law, is run by a separate committee that manages those decisions, and then the company makes some donations as well.

I think regardless, we have a broad based, philosophical approach of supporting individuals, typically incumbents, but not always, who are committed to the same goals that we feel best serve the needs of our customers and the interests of our shareholders.

When you think about the needs of our customers, we have supported office holders, legislators, typically, who have done a lot of work to protect privacy, especially privacy online, who’ve protected cybersecurity. Those are two areas where we can do a lot as a company, but strong laws need to be in place to protect our customers.

Skilling is another area. We’re very supportive of people who are helping to advance education and make technical education, technology education more broadly available to people.

And then we think about the interests and needs of our shareholders. It starts with what serves our customers. Anything that serves our customers well will, by definition, serve our shareholders well. But we also look at things like intellectual property rights, so we may look at taxation or other fiscal policies that we think are going to serve the shareholders who own Microsoft stock.

And that’s basically it. It’s a very focused effort, and we always try to make decisions based on principle.

JONATHAN NEILSON: That’s great. Now, we have time for just one last question.

Satya, what’s the company’s direction on quantum computing?

SATYA NADELLA: Yeah, I mean, when it comes to these foundational technologies, much like AI, the approach Microsoft has always taken is, in some sense, a long-term approach. If you think about AI, we got started not in 2019, but we got started in 1995 when Bill first started Microsoft Research, and the first research group was, I think, speech. And so, these are long arcs where you continue to learn, invest, and then deploy in the real world, and then have that feedback cycle.

In some sense, quantum is the same, which is, we’ve had conviction that quantum is going to be the next big breakthrough in computing. And 20+ years ago is when we started, and here we are with major breakthroughs in the last year, because it’s not only the breakthroughs on the software side, but it’s really the physics breakthroughs that need to be coupled with it to build what is essentially, a warehouse scale or utility scale quantum computer that truly can do computation that’s stable and it can solve some real-world problems.

We’re making great progress on this approach we’ve had, while, at the same time, making sure our quantum software runs on even all the other types of quantum computers, whether it’s neutral atoms, whether it’s ion trap, or what have you, and then have those quantum computers all run in Azure.

And then the other exciting part is to use quantum computers to generate data. I mean, obviously, the world is quantum. And so, therefore, if you want to simulate nature, quantum can generate that simulation data, and that data can then be used to train AI models in chemistry and material science.

And so, I see even near-term application, even with logical qubits that are a lot smaller than utility scale, where it will advance some of these AI models. We feel very, very good about the progress we’re making on quantum.

JONATHAN NEILSON: Yes, it’s exciting, for sure.

Well, this concludes today’s Q&A session, and wraps up our Annual Shareholders Meeting.

Thank you for joining us today and for your participation. If you have further questions, please visit our Investor Relations website. Thank you.

SATYA NADELLA: Thank you, all.

AMY HOOD: Thank you.

END