Microsoft Azure Surges 45% as Copilot Hits 30M Subscriptions

Woofun AI reports that Bernstein has maintained its "outperform" rating on Microsoft, adjusting the target price from $646 to $647, driven by sustained momentum in Azure and Copilot. This valuation adjustment underscores a critical market shift: the firm’s two primary AI initiatives are demonstrating robust growth trajectories that directly address prevailing investor anxieties regarding artificial intelligence demand sustainability and cash flow strain from infrastructure build-outs. The modest $1 increase in the target price is less significant than the underlying thesis that Microsoft has successfully decoupled heavy capital expenditure from immediate liquidity risks, thereby validating the commercial viability of its AI strategy.

The financial performance for the fourth fiscal quarter of FY26 provided the empirical basis for this bullish stance, with total revenue reaching $90.007 billion and operating revenue climbing to $40.603 billion. GAAP diluted earnings per share settled at $4.81, figures that collectively triggered a 9% surge in share price during after-hours trading. This market reaction was not merely a response to top-line growth but a direct acknowledgment that both cloud business metrics and AI-related commercialization data exceeded consensus expectations. The strength of these results signals that Microsoft’s operational execution remains tightly aligned with its strategic pivot toward AI-integrated services, reinforcing investor confidence in the company’s ability to monetize its technological advantages.

Azure emerged as the central catalyst for this positive sentiment, recording 43% constant currency growth in the fourth fiscal quarter of FY26. More critically, Microsoft projected Azure revenue to expand by approximately 45% on a constant currency basis in the first fiscal quarter of FY27. Bernstein estimated this projection to be 400 basis points higher than market expectations, effectively dispelling fears that the cloud business had entered an "AI digestion phase" characterized by slowing adoption.

Management confirmed during the earnings call that customer demand continues to outstrip available capacity, identifying supply constraints rather than demand weakness as the primary bottleneck. This dynamic implies that the massive data center construction undertaken in recent quarters has not resulted in idle assets; instead, Azure’s growth remains artificially capped by infrastructure availability, suggesting a prolonged period of supply-constrained expansion.

Woofun AI data shows Microsoft Cloud revenue reached $59.3 billion in the fourth fiscal quarter, representing 27% year-on-year growth and accounting for more than half of the company’s total revenue. This metric reinforces the structural trend of cloud transformation, now accelerated by AI integration. Supporting this view, commercial remaining performance obligations (RPO) surged 84% year-on-year to $678 billion. Even after excluding the impact of OpenAI and other leading model companies, RPO still grew by 25% year-on-year, driven by broader enterprise customer adoption. This segmentation is vital, as it alleviates concerns that Azure’s growth is overly dependent on a few key clients; the diversified enterprise demand indicates a resilient revenue base that is not vulnerable to shifts in investment patterns by single major partners.

The adoption of Microsoft 365 Copilot further validated the company’s software monetization strategy, with paid subscriptions surpassing 30 million. This figure exceeded Bernstein’s earlier market estimate of "over 25 million" and marked a substantial increase from the 20 million subscriptions recorded in the previous quarter. For Microsoft, this growth signifies more than just the sale of an additional AI tool; it demonstrates the successful integration of AI capabilities into existing enterprise software systems. The continued rise in subscriptions counters the narrative that AI would undermine the traditional SaaS subscription model or lead enterprises to reduce per-employee software spending. Instead, Microsoft is leveraging AI to justify higher per-user fees, thereby strengthening the economic foundation of its SaaS business model.

Within the M365 Commercial cloud segment, reported growth stood at 14% in the fourth fiscal quarter, expanding to approximately 16% on an adjusted basis. This improvement was primarily driven by upgrades to E5 and E7 plans and the widespread adoption of Copilot features. Microsoft is not selling AI as an isolated plugin but is embedding these functions into existing enterprise software packages to increase spending per subscription.

Concurrently, GitHub Copilot expanded its user base to 50 million, gradually shifting toward a hybrid pricing model that combines fixed subscriptions with usage-based charges. This evolution indicates that revenue from developer-focused AI tools is no longer solely dependent on static subscription fees but is increasingly tied to actual usage levels, creating a more dynamic and scalable revenue stream.

Capital expenditure figures, often prone to misinterpretation, revealed a nuanced picture of Microsoft’s investment strategy. The company revised its CY26 capital expenditure forecast to around $175 billion, down from the previous estimate of $190 billion.

However, Microsoft clarified that this reduction stems from accounting changes rather than a slowdown in physical construction. Starting in FY27, the estimated useful life of data centers and office buildings will extend from 15 years to 25 years, leading to a reclassification of many data center leases from finance leases to operating leases. Consequently, the downward revision reflects a change in leasing classification and useful life assumptions, not a reduction in the actual scale of AI data center construction or hardware procurement.

Looking ahead, the first fiscal quarter of FY27 guidance indicates that capital expenditure will remain above $50 billion, following a fourth fiscal quarter total of approximately $41 billion. Of this amount, $35.8 billion was spent on purchasing property, plant, and equipment (PPE) in cash, with roughly two-thirds allocated to short-lived assets such as GPUs and CPUs. Despite the intensity of these investments, Microsoft’s management expects free cash flow to remain positive in FY27. This outlook contrasts sharply with some hyperscale cloud providers that experienced negative free cash flow during peak AI investment periods, highlighting Microsoft’s superior ability to balance aggressive infrastructure build-outs with liquidity preservation.

Operational efficiency metrics further support this optimistic cash flow projection. Microsoft Cloud gross margin stood at 65%, down 3 percentage points year-on-year due to AI infrastructure investments, yet Azure’s efficiency continues to improve. The throughput of Copilot workloads increased fourfold over the course of the year, helping to reduce unit delivery costs. As of the fourth fiscal quarter, Microsoft’s platform supports over 11,000 AI models, including offerings from leading research labs, open source models, and Microsoft’s own proprietary models. This scale and efficiency gain position Microsoft to withstand fluctuations in the AI infrastructure cycle better than most competitors, provided that AI-related revenue growth remains strong and free cash flow stays positive.

Valuation metrics reflect these growth expectations, with Bernstein estimating FY27 revenue at $392.63 billion and FY28 revenue at $466.335 billion. These projections result in adjusted P/E ratios of 19.6 times and 16.3 times, respectively. While these valuations are not exaggerated, they assume continued strong growth in Azure, ongoing Copilot adoption, and no significant erosion of cash flows due to AI investments. The primary risk remains that AI demand might fall short of expectations, potentially extending the payback period for data center and hardware investments.

Furthermore, the true flexibility of capital expenditure is uncertain, as GPUs, CPUs, electricity, and data center leases often require advance commitments. The real test lies in whether these investments in hash rate will continue to translate into corporate payments, cloud revenue, and sustainable free cash flow, rather than fueling an "AI bubble" that eventually bursts.