Microsoft Is Shuffling the Deck—Again—While Its Stock Tumbles 20%

The Headlines

Microsoft just made three big moves in three weeks, and the market's response? Sell first, ask questions later.

In late February, the company's gaming chief Phil Spencer retired after 38 years. Then, on March 12, Rajesh Jha—the executive in charge of experiences and devices—announced he'd retire in July after more than 35 years at Microsoft. And just five days later, Microsoft overhauled its Copilot leadership structure, promoting Jacob Andreou (a former Snap executive) to oversee the AI assistant across commercial and consumer products.

Meanwhile, MSFT's stock is trading near $390, well below its 52-week high of $555. That's a gut-punch for anyone holding shares.

The Backstory

To understand why Microsoft is reshuffling leadership right now, you need to see the pressure the company is under.

In January, Microsoft beat earnings expectations—reporting $81.27 billion in revenue against a forecast of $80.27 billion. But the stock fell about 7% in after-hours trading anyway. Why? Investors are spooked. Azure and cloud services revenue grew 39%—impressive on paper. But the narrative in the room was different: growth is slowing.

Here's the real squeeze: Microsoft is pouring capital into AI infrastructure at an unprecedented rate. The company spent $37.5 billion on capital expenditures in Q2—mostly on GPUs and CPUs—and is projecting CapEx could hit as high as $93.7 billion for all of FY2026. That's money going out the door to build data centers and buy chips from NVIDIA. But does it translate to revenue that justifies the investment? The market isn't convinced yet.

The leadership changes look like Microsoft's way of saying: we need to fix this faster. Jha's departure opens the door for four executives—Perry Clarke, Charles Lamanna, Pavan Davuluri, and LinkedIn CEO Ryan Roslansky—to report directly to CEO Satya Nadella, flattening the hierarchy. And Andreou's promotion signals that Copilot and AI agents are now front-and-center in the company's org chart.

The Takes

So what's the bull case? Microsoft is saying it's doubling down on AI with a new enterprise product: the Microsoft 365 E7 bundle, launching May 1 at $99 per user per month. This bundles Microsoft 365 E5, the $30 Copilot add-on, $12 in Entra identity tools, and a new $15 Agent 365 product for managing AI agents. It's the first new enterprise tier since E5 launched in 2015.

For context: Microsoft 365 Copilot has about 15 million paying users, roughly 3% of the overall Microsoft 365 user base. That's small but growing. And the standalone Copilot app had 6 million daily active users in February—nowhere near OpenAI's ChatGPT at 440 million or Google's Gemini at 82 million, but the bundle strategy suggests Microsoft believes it can monetize AI differently—through enterprise lock-in, not volume.

Microsoft also announced Copilot Cowork, built with AI developer Anthropic, designed to handle multi-step tasks like scheduling emails and meeting prep. That's the kind of product that sticks with users if it works.

But here's the bear case: the math doesn't add up yet. Analyst concerns center on CapEx growing faster than expected while Azure growth is decelerating slightly, raising questions about return on AI infrastructure investment. In other words: Microsoft is spending enormous sums to build the infrastructure, but the revenue story isn't keeping pace. The Xbox business alone saw revenue decline nearly 10% in the December quarter, steeper than management expected—a sign that even within Microsoft's portfolio, growth is uneven.

And there's more friction: Commercial pricing for Microsoft 365 increases on July 1: E3 rises 8% to $39 per user per month and E5 rises 5% to $60 per user per month. Customers who balked at E7's $99 price tag might just stick with cheaper tiers and skip Copilot altogether.

Real Talk

Put all the pieces together and a pattern emerges: Microsoft is trying to solve a confidence problem, not a competence problem.

The company has built world-class infrastructure. At NVIDIA's GTC conference on March 16, Azure became the first hyperscale cloud provider to power NVIDIA's new Vera Rubin NVL72 systems, and Microsoft had already deployed hundreds of thousands of liquid-cooled NVIDIA Grace Blackwell GPUs across its data centers in under a year. That's not a company struggling with execution.

But the market doesn't care about infrastructure for its own sake—it cares about ROI. Microsoft Cloud revenue surpassed $50 billion for the first time, reaching $51.5 billion and growing 26%. That's real money. The problem is the gap between the CapEx ($93.7 billion projected) and the revenue growth trajectory (37–38% for Azure). If you're spending nearly $94 billion to grow $51 billion in revenue by 37%, when does the payoff arrive?

The leadership shuffle makes sense in that context. Nadella is centralizing control, flattening the org, and betting everything on AI monetization through enterprise bundles and agents. He's essentially saying: the infrastructure is built, now we need to sell it. That's a sales and product problem, not an engineering one.

The Bottom Line

Microsoft's stock has fallen roughly 20% from late November 2025 to early March 2026, with shares trading near $390 against a 52-week high of $555. The core question investors are asking isn't whether Microsoft can build AI systems—it clearly can. The question is whether $93.7 billion in annual CapEx will eventually generate returns that justify the spend.

The company is banking on enterprise adoption of Copilot and AI agents, with new products like the $99/month Microsoft 365 E7 bundle and Foundry Agent Service now generally available. But Copilot's current penetration—15 million paying users out of a Microsoft 365 base of roughly 500 million—suggests the runway to monetization is still long.

The next earnings report is due April 28. That's when the market will get fresh data on whether the leadership changes are actually moving the needle. Until then, the stock is caught between two narratives: a company with unmatched AI infrastructure and reach, or a company spending more than it's earning from that advantage. You get to decide which story you believe.