Bernstein Analyst Interprets Microsoft: Azure Growth Accelerates to 45%, Copilot Seats Surpass 30 Million, Microsoft's AI Commercialization Continues to Gain Momentum

TL;DR
· Microsoft's FY26 Q4 earnings report and FY27 Q1 guidance beat expectations, with the stock price jumping about 9% in after-hours trading.
· Azure Q1 constant currency growth guidance is around 45%, with over 30 million paid Copilot seats.
· Capital expenditure reduction was mainly due to accounting changes, while AI demand and cash return remain key points of contention.

Following Microsoft's strong performance in FY26 Q4 earnings and upbeat guidance for the next quarter, Bernstein has maintained Microsoft's "Outperform" rating, slightly raising the target price from $646 to $647. The core reason behind this decision is that Azure and Copilot, the two main AI drivers, continue to accelerate.

The mere $1 increase in the target price is not the focus. More importantly, Microsoft addressed the market's two most pressing questions: whether AI demand remains strong and if AI infrastructure investment would weigh on cash flow.

Microsoft's official report shows that in FY26 Q4, total revenue was $90.07 billion, operating income was $40.603 billion, and GAAP diluted EPS was $4.81. According to AP, Microsoft's stock price surged about 9% in after-hours trading following the earnings release. The market reaction was primarily driven by the cloud business and AI commercialization figures beating expectations.

Azure Has Not Reached Saturation, Q1 Guidance Continues to Accelerate to 45%

Azure was the most direct highlight of this earnings report.

In FY26 Q4, Azure and other cloud services revenue grew 43% year-over-year in constant currency. More crucially, Microsoft's FY27 Q1 constant currency growth guidance for Azure is around 45%. Bernstein estimates that this guidance exceeds market expectations by about 400 basis points.

At least from Microsoft's disclosed short-term orders and capacity status, the cloud business has not yet entered the feared "AI saturation period" from outside observers. Management stated on the call that customer demand continues to outstrip available capacity, with capacity still being the primary constraint.

For investors, the implication of this figure is straightforward: Microsoft's massive AI data center build-out over the past few quarters has not immediately turned into idle risk. On the contrary, capacity is still insufficient, and Azure growth is still constrained by supply.

Microsoft Cloud revenue in the fourth quarter reached $59.3 billion, up 27% year-over-year, already contributing steadily over half of the total revenue. The cloud transformation, an ongoing core theme, has not ended, with AI just accelerating the growth trajectory.

Azure's constant currency revenue growth increased from the 30%-35% range to 43% in Q4, with Q1 guidance further to around 45%.

Order indicators also support this. Commercial remaining performance obligation (RPO) grew by 84% year-over-year to $67.8 billion. Even excluding the impact of OpenAI and other frontier model companies, RPO still grew by 25% year-over-year, driven by growth in enterprise customers.

This breakdown is crucial. The market has been concerned that Microsoft's AI cloud growth relies too heavily on a few large customers such as OpenAI, and if the investment pace of these customers changes, Azure revenue could be under pressure. However, excluding this portion, enterprise customers are still growing, indicating that the demand is not solely coming from a single major customer.

Copilot Seats Exceed 30 Million, Alleviating Concerns of SaaS Disruption by AI

Another outperforming figure comes from Microsoft 365 Copilot.

The Microsoft management confirmed that the paid seats for Microsoft 365 Copilot have exceeded 30 million. Bernstein stated that this number is above the market's previous expectation of "over 25 million" and significantly higher than the over 20 million level from the previous quarter.

The significance for Microsoft is not just selling another AI tool. Over the past year, one of the market's concerns about software companies is: Will AI weaken the traditional SaaS seat model? Will enterprises reduce software spend per user? The growth of Copilot seats at least shows that Microsoft can still embed AI capabilities into existing enterprise software systems and monetize through increased per-user revenue.

M365 Commercial cloud reported a fourth-quarter revenue growth of 14% in the reporting currency, with an adjusted growth of approximately 16%. Revenue improvement primarily came from E5, E7 package upgrades, and Copilot adoption. Microsoft is not selling a standalone AI plugin but rather bundling AI features with existing enterprise software packages to increase spending per seat for customers.

M365 Commercial Cloud Growth Improves, Copilot Paid Seats Increase from 15M, 20M to over 30M.

GitHub Copilot is also expanding its user base, reaching 50 million users, and gradually transitioning to a "subscribe + consume" hybrid pricing model. The revenue of the developer AI tool now not only relies on fixed seats but also begins to be tied to usage.

However, Copilot's numbers still have boundaries. Over 30 million paid seats demonstrate adoption is accelerating, but it cannot directly prove that all enterprise customers will indefinitely expand their AI software budgets. Whether it can continue to contribute revenue depends on renewal rates, usage frequency, and the scale of internal enterprise deployments.

Capital Expenditure Appears to Decrease, but Real AI Expansion Has Not Slowed Down

One of the most easily misinterpreted numbers in the financial report is capital expenditure.

Microsoft has adjusted its CY26 capital expenditure expectation to around $175 billion, lower than the previous estimate of around $190 billion. On the surface, this seems like a slowdown in AI investment. However, Microsoft explains that starting from FY27, the company will extend the estimated useful life of its data centers and office buildings from 15 years to 25 years, leading to more data center leases transitioning from finance leases to operating leases. Apart from the useful life impact, the CY26 capital expenditure investment expectation remains unchanged.

In other words, the decrease is mainly due to accounting and leasing classification changes and does not mean that Microsoft is cutting back on AI data center construction.

A more telling indication of the real investment intensity is the next quarter's guidance: FY27 Q1 capital expenditure is still expected to exceed $50 billion. Total capital expenditure for Q4 is around $41 billion, with cash spent on PPE at $35.8 billion, with about two-thirds used for short-lived assets such as GPUs and CPUs.

Q4 total capital expenditure is around $41 billion, with cash spent on PPE approximately $35.8 billion, with about two-thirds directed towards short-lived assets such as GPUs and CPUs.

Capital expenditure has not slowed down, and the market will naturally inquire about cash flow. The Microsoft management team expects FY27 free cash flow to remain positive, contrasting with some hyperscale cloud providers that have shifted to negative free cash flow during the peak of AI investment.

Microsoft Cloud gross margin is 65%, decreasing by 3 percentage points year-over-year due to AI infrastructure investments, but Azure efficiency is still improving. Copilot's workload throughput has quadrupled during the year, helping to reduce unit delivery costs. As of the fourth quarter, the Microsoft platform has supported over 11,000 AI models, covering leading labs, open-source models, and proprietary models.

Microsoft Cloud Q4 Revenue of $59.3 billion, up 27% year-over-year, accounting for over half of total revenue.

This is also a key reason Bernstein remains bullish on Microsoft: if AI revenue continues to grow, while free cash flow remains positive, Microsoft is better positioned than most peers to withstand the fluctuations of the AI infrastructure cycle.

AI Demand Continues to Rise, Capital Return Still Under Scrutiny

This earnings report alleviated market anxiety but did not eliminate all controversies.

The most immediate risk is still AI demand falling below expectations. The current signal from Microsoft is that demand far exceeds supply, with strong RPO growth and increasing Copilot adoption. However, if enterprise AI adoption lags behind expectations, or spending from leading AI model companies slows down, the payback period for data center and hardware commitments will be prolonged.

Another risk comes from the true flexibility of capital expenditures. Microsoft emphasizes that there is some room for adjustment in data center and hardware commitments, but in the AI infrastructure construction cycle, GPUs, CPUs, power, and data center leases often need to be locked in advance. If demand falls short of expectations, there is still uncertainty about whether the flexibility is sufficient to protect capital returns.

Valuation is also not entirely without pressure. Bernstein estimates that Microsoft's revenue for FY27 and FY28 is expected to be $392.63 billion and $466.335 billion, respectively, with adjusted P/E ratios of approximately 19.6 times and 16.3 times. This valuation is not unreasonable, but the premise is that Azure continues high growth, Copilot continues to penetrate, and AI investment does not significantly erode cash flow.

The answer provided by this earnings report is more like this: Microsoft's AI narrative is still gaining momentum and has not been derailed by capital expenditures. However, it is not yet the definitive proof of the "AI bubble bursting." The real test is whether these computing power investments can continuously translate into enterprise payments, cloud revenue, and free cash flow.