
- Microsoft Earnings: Strong Numbers, But Not A Clean Celebration
- Why Microsoft Stock Fell: AI Growth Is Not Cheap
- What Investors Should Not Miss in Microsoft Earnings
- What Analysts Are Saying About MSFT Stock
- What The Earnings Call Was Really Hinting At
- Microsoft Copilot Is The Real Test
- What Should Investors Do Now?
Microsoft reported the kind of quarter most companies would happily frame on a wall. Revenue rose 18% year-on-year, EPS jumped 23%, Azure grew 40%, and Microsoft Cloud revenue touched $54.5 billion.
Still, Microsoft stock slipped around 5% after earnings. Why? Because investors are worried about how expensive that growth is becoming.
Let’s break down what really spooked MSFT stock investors, what analysts are saying, and what investors should watch before taking a view on Microsoft stock.
Microsoft Earnings: Strong Numbers, But Not A Clean Celebration
The headline Microsoft earnings numbers were solid. Revenue, profit and cloud growth all came in strong. Microsoft Cloud grew 29%, while commercial remaining performance obligation, or contracted future revenue, nearly doubled to $627 billion. That means enterprise demand is still very real.
| Metric | Q3 FY26 | Investor Read |
| Revenue | $82.9 billion | Up 18% YoY |
| EPS | $4.27 | Up 23% YoY |
| Microsoft Cloud revenue | $54.5 billion | Up 29% YoY |
| Azure growth | 40% | Strong, but expected |
| Commercial RPO | $627 billion | Demand visibility is high |
| Capex | $31.9 billion | The main concern |
So, the problem was not the quarter. The problem was the bill attached to the quarter.
Why Microsoft Stock Fell: AI Growth Is Not Cheap
The biggest pressure point for MSFT stock was capital expenditure. Microsoft spent $31.9 billion on capex in Q3. Property and equipment additions alone were $30.9 billion, compared with $16.7 billion a year ago.
That means Microsoft is pouring huge cash into data centres, GPUs, CPUs and AI infrastructure. This is not optional spending. Management said AI demand continues to exceed available capacity, which means Microsoft still needs to build more.
Here is the simple explanation: Microsoft is building the AI highway before all the toll collections have fully started. If traffic explodes, this becomes a smart land grab. If monetisation takes longer, investors will keep questioning the payback.
What Investors Should Not Miss in Microsoft Earnings
Microsoft’s cost of revenue grew 22%, faster than revenue growth of 18%. Microsoft Cloud gross margin also fell to 66%. Management directly linked this to AI infrastructure investment and higher AI product usage. That matters because Microsoft has historically been loved for high-margin software economics. AI changes the equation.
Every Copilot query, AI workload and model inference has a cost behind it. So investors are now asking: can Microsoft price AI high enough to protect margins? This is the real debate after MSFT earnings.
What Analysts Are Saying About MSFT Stock
Analysts are not calling Microsoft weak. Azure’s 40% growth was only slightly above the previous quarter’s 39%. That is good, but investors wanted more acceleration given the size of Microsoft’s AI spending. The comparison also hurt because Google Cloud recently posted much faster growth, making Azure look less explosive.
Rebecca Wettemann of Valoir said the market “wanted to be wowed” given questions around AI infrastructure spending, while Melissa Otto of S&P Global Visible Alpha said it was not the “blow-away quarter” needed to move the stock higher.
What The Earnings Call Was Really Hinting At
The earnings call gave investors two important signals”
- Microsoft expects to remain capacity constrained at least through 2026. That means demand is strong, but it also means capex may stay elevated.
- CFO Amy Hood said Azure growth should show modest acceleration in the second half of calendar 2026.
There was another important hint. Microsoft is moving AI monetisation from just a licence model to a mix of licence plus consumption. In plain English, Microsoft wants companies to pay not only for Copilot access, but also for how much AI they actually use. That could be powerful, but only if usage keeps rising.
Microsoft Copilot Is The Real Test
Microsoft now has over 20 million paid Microsoft 365 Copilot seats, up from 15 million in January. Seat additions grew 250% year-on-year, and customers with more than 50,000 seats quadrupled. Accenture alone has over 740,000 seats.
That is strong momentum. But Microsoft has hundreds of millions of Office users. So 20 million paid seats is still early. The key question is whether Copilot can become a daily habit like Outlook or Teams. If that happens, Microsoft can defend premium pricing. If not, Copilot may remain a promising but expensive add-on.
This is the analyst debate in simple terms:
| Bull View | Bear View |
| Azure still growing around 40% at Microsoft’s scale is powerful | Growth is not accelerating enough versus AI capex |
| $627 billion commercial RPO shows demand visibility | Backlog does not automatically equal near-term cash flow |
| Copilot adoption is improving | 20 million paid seats is still early relative to Microsoft’s user base |
| AI revenue run rate crossed $37 billion | Investors need clearer margin and free cash flow payback |
What Should Investors Do Now?
For investors evaluating Microsoft stock, the decision should not be based only on whether Microsoft earnings beat estimates. The better framework is:
| Investor Question | What To Watch |
| Is AI demand real? | Azure growth, AI revenue run rate, RPO growth |
| Is AI profitable? | Cloud gross margin, depreciation, free cash flow |
| Is Copilot working? | Paid seats, usage per user, large enterprise rollouts |
| Is spending disciplined? | Capex as a share of operating cash flow |
| Is competition rising? | Google Cloud, AWS, model marketplace dynamics |
Microsoft remains one of the strongest AI-cloud compounders globally. But after this quarter, MSFT is no longer being valued just as a software company with high margins. It is increasingly being evaluated as a software-plus-infrastructure company where data centre spending, GPU costs and utilisation rates matter.