Microsoft just delivered a fiscal year that would be the envy of any tech giant—record revenues, soaring profits, and a market cap that surpassed $4 trillion for the first time. But beneath the eye-popping numbers, a harsher reality is unfolding: over 17,000 employees have been shown the door this calendar year, even as the company plowed an unprecedented $85 billion into artificial intelligence. The contrast couldn’t be starker. On one hand, CEO Satya Nadella is betting the company’s future on AI, with spending set to explode to $30 billion in the current quarter alone. On the other, the human cost of that transformation is becoming impossible to ignore.
The financial engine driving this duality is staggering. For the quarter ended June 30, Microsoft posted net income of $27.2 billion on revenues of $76.4 billion—gains of 24% and 18%, respectively, over the same period last year. Full-year figures were even more breathtaking: $101.8 billion in profit (up 16%) on $281.7 billion in revenue (up 15%). Azure, the cloud division long shrouded in secrecy, finally opened its books, revealing $75 billion in annual revenue—a 34% leap from the estimated $56 billion the year before. It was a coming-out party for a unit that now underpins practically every AI service Microsoft offers.
Yet the earnings call was as much about what wasn’t said as what was. Nadella and his team faced no questions about the 17,000-plus layoffs executed in 2025, thanks in part to a barrage of carefully timed headline-grabbing announcements: a “Windows vision” video forecasting life in 2030, the quiet burial of Windows 11 SE, and new AI features everywhere from Xbox to GitHub. Headcount, Microsoft insisted, was “unchanged” year-over-year, a claim that stretched believability given the bloodletting. For the thousands who lost their jobs, the message was clear: the AI future has no room for redundancy.
The AI Moonshot: $85 Billion and Counting
Microsoft’s AI spending has reached levels that were unthinkable even two years ago. The $85 billion figure for fiscal 2025 covers everything from the costly custom silicon needed to train large language models to the massive data-center buildouts that keep Azure humming. And the pace is accelerating: the company expects to drop another $30 billion on AI in just the next three months. To put that in perspective, that single-quarter sum rivals the annual R&D budgets of many Fortune 500 firms.
The investment is anchored by the ever-deepening partnership with OpenAI, whose GPT models power Copilot across Microsoft 365, GitHub, and a growing array of enterprise tools. Microsoft also made waves by releasing OpenAI’s first “open-weight” reasoning models for free—a move that democratizes cutting-edge AI but has sparked debate about whether the company is giving away its competitive moat. Meanwhile, Apple and Amazon are racing to infuse their own products with generative AI, from Siri to Alexa+, setting the stage for a multi-front war where Microsoft’s enterprise credibility and integration give it a significant head start.
Copilot: A New Productivity Tier
Nadella now refers to Copilot less as a feature and more as a new tier of Microsoft 365, and the numbers back him up: over 100 million monthly active users (MAUs) now have access to Copilot within the productivity suite. While that MAU figure likely includes anyone with the feature enabled—not just power users—it still represents a massive embedded base. GitHub Copilot, with its 20 million MAUs (many on free licenses), has become indispensable to developers, even as Microsoft contemplates how to convert free tiers into sustained revenue.
The vision goes far beyond answering emails. In the Windows 2030 concept video, AI mediates almost every interaction—drafting documents, summarizing meetings, and even anticipating needs before the user utters a command. David Weston, a security architect and the unlikely face of this campaign, stressed that AI should reduce what he calls “toil work,” freeing humans for creativity and collaboration. Multimodal inputs—voice, gaze, gestures—are poised to replace keyboard-and-mouse drudgery, a shift that would fundamentally alter how we think about an operating system.
Azure: The $75 Billion Backbone
For years, observers had to guess Azure’s size based on convoluted “commercial cloud” metrics. No more. The disclosure of $75 billion in annual Azure revenue not only validated the cloud unit’s critical role but also gave Wall Street the transparency it long demanded. Even more intriguing is the math: if Azure generated $56 billion the previous year, the growth trajectory suggests it could overtake Office and Windows as Microsoft’s single biggest revenue engine within a few years. That shift would cement a reality where Microsoft is, first and foremost, an infrastructure and AI company.
Gaming’s Golden Age: 500 Million Players
The Xbox division, often treated as a sideshow in earnings calls, delivered its loudest statement yet. Microsoft Gaming now claims over 500 million monthly active users across console, PC, and cloud—a number that dwarfs many standalone platforms. Call of Duty alone contributes 50 million MAUs, underscoring the value of the Activision Blizzard acquisition. Game Pass revenues hit $5 billion for the year, with over 500 million hours played, while nearly 40 first-party titles are in the pipeline. A new Gaming Copilot, currently in beta for Xbox Insiders on Windows, hints at an AI layer that could eventually offer real-time tips, troubleshoot issues, and make social gaming more seamless.
Windows 2030: Ambitious, Ambiguous
Microsoft’s “vision video” for Windows was long on ambition but short on specifics. It showcased a world where the OS is almost invisible—an ambient layer that translates user intent across devices. Security, Weston insisted, must become appliance-level: “it just works.” That’s a direct response to years of criticism that Windows security is too complex and too brittle. The death of Windows 11 SE, the education-focused variant that never caught on, signals a willingness to cull distractions and focus resources on this grand reimagining. Yet the developer ecosystem, long a sore point, remains a glaring weakness. Native Windows apps still lag behind their web and mobile counterparts in polish and performance, and until Microsoft fixes that, the 2030 dream will struggle to materialize.
Tools and Tips for the Here and Now
While the future unfolds, power users have immediate ways to gauge and improve their systems. The Windows System Assessment Tool (WinSAT) is back in the spotlight. Running winsat formal from an elevated command prompt generates a detailed XML report stored in %systemroot%\performance\winsat\datastore. Open the newest file and look for the <WinSPR> section to see a component-level breakdown of your PC’s performance—a modern echo of the old Windows Experience Index.
For security, Proton’s new standalone authenticator app provides an alternative to platform-tied solutions, though many will still need Microsoft Authenticator for its deeper integrations. The key takeaway: no single tool covers every scenario, and a layered approach remains essential.
The Road Ahead: Profits, Pain, and Promise
Microsoft stands at a crossroads. The financials are historic, the AI bet is truly audacious, and the product vision is as bold as it’s been in decades. But the company cannot outrun the human consequences of its efficiency drive forever. Employees, developers, and customers alike will be watching whether the $85 billion AI push delivers not just revenue growth but sustainable, human-centric innovation. Windows 2030 may be the ultimate test: can a 40-year-old operating system reinvent itself as an ambient, AI-first platform without losing the trust of the billion-plus people who rely on it every day? The answer will define not just Microsoft’s next chapter, but the very future of personal computing.