Microsoft's Earnings Tightrope: Azure Deceleration, a Historic Severance Offer, and OpenAI's New Rules

Microsoft reports fiscal Q3 results on April 29, with Azure growth rate critical after stock fell 20% year-to-date despite earnings beats.

When Microsoft reports its fiscal third-quarter results after the closing bell on April 29, the numbers will almost certainly beat expectations. The problem is that hasn't mattered lately. In January, the tech giant delivered a solid earnings beat only to see its stock plunge nearly 10% the next day. The culprit then was a single percentage point: Azure growth had slipped to 39%, down from 40% in the prior quarter. Investors punished the stock mercilessly.

That pattern has defined Microsoft's year. The shares have lost more than 20% since that January selloff, trailing other megacap tech names. At roughly €366, the stock sits nearly 22% below its 52-week high from last summer and remains well under its 200-day moving average. The market is no longer rewarding Microsoft for simply showing up — it wants proof that the AI infrastructure spending spree is translating into accelerating cloud revenue.

Azure Is the Only Number That Matters

Analysts are looking for earnings per share of $4.06 on revenue of $81.3 billion, a sharp jump from the year-ago period's $3.46 EPS and $70.1 billion in sales. Microsoft itself guided for revenue between $80.65 billion and $81.75 billion. But the headline figures are almost secondary. The single data point that will determine the stock's direction is Azure's growth rate.

Management has guided for Azure expansion of 37% to 38% in constant currency — the third consecutive sequential decline from 40% to 39% to now 37-38%. The critical question is whether this deceleration reflects structural demand weakness or simply a GPU supply bottleneck. On the Q2 call, executives argued that without graphics processor shortages, Azure could have grown above 40%. Bank of America analyst Tal Liani backs that interpretation, saying capacity constraints rather than flagging demand are holding back growth. Evercore's Kirk Materne takes a more pragmatic view: growth at the high end of guidance, roughly 38%, would be "good enough."

Morgan Stanley sets the bar higher. For the investment bank, 39% is the threshold that would actually move the stock. The logic is straightforward: Azure's growth is currently supply-constrained, not demand-constrained. More capacity means more revenue directly. A positive signal came on April 16, when CEO Satya Nadella announced that the Fairwater data center in Wisconsin — part of a $7 billion-plus investment in the state — went live ahead of schedule, connecting hundreds of thousands of GB200 chips into a single cluster.

A Historic Severance Program and a Restructured OpenAI Deal

Two developments have made this earnings report unusually consequential. For the first time in its 51-year history, Microsoft is offering voluntary severance to a portion of its U.S. workforce. Roughly 8,500 employees — about 7% of the U.S. headcount — are eligible, provided their age plus years of service equals at least 70. The offer extends to employees up to the senior director level, though those on sales incentive plans are excluded. Microsoft hasn't disclosed the full financial terms, but health insurance is included and participants face no non-compete restrictions. Eligible employees will receive details on May 7.

The backdrop is obvious. Microsoft is planning capital expenditures of $110 billion to $120 billion this fiscal year, up sharply from roughly $80 billion last year. That AI data center spending binge is forcing discipline elsewhere on the balance sheet.

Just two days before earnings, Microsoft and OpenAI fundamentally restructured their partnership. The licensing deal for OpenAI's intellectual property now runs through 2032 — but it's no longer exclusive. Microsoft will no longer pay OpenAI a revenue share for Copilot products, while OpenAI's cut of Microsoft's revenue continues through 2030 with an agreed cap. The immediate effect is a margin improvement for Copilot, since the revenue leakage to OpenAI disappears. The trade-off: OpenAI can now license its models to Amazon's cloud services or other competitors.

Copilot's Penetration Problem and the Capex Paradox

Microsoft's AI assistant Copilot has reached 15 million paying users — a respectable number but a tiny fraction of the roughly 450 million commercial Microsoft 365 base. At $30 per user per month, the revenue potential is enormous, but adoption remains early stage. Any new usage metrics disclosed tonight will be scrutinized closely.

The financial picture is complicated by the infrastructure buildout. Microsoft is spending more than $100 billion on capital expenditures this fiscal year, much of it on GPUs and AI hardware. That's squeezing free cash flow margins from roughly 25% last year to an estimated 21.5% in fiscal 2026. The paradox: on the income statement, operating leverage is visible — operating income is growing faster than revenue. On the cash flow statement, the story is different.

A Crowded Earnings Afternoon

Adding to the pressure, Amazon, Google, and Meta all report on the same afternoon. Investors will compare Azure, AWS, and Google Cloud growth rates in real time, making any deviation instantly visible. Of 49 analyst recommendations, 41 are "Strong Buy." The options market is pricing in a 7% swing in either direction. The bar is set high, and the market has already shown it won't hesitate to sell first and ask questions later.