Microsoft’s proactive deployment of AI infrastructure creates upside potential for revenue forecasts.

Overview

This research report focuses on Microsoft’s monetization of its infrastructure in the early stages of the artificial intelligence (AI) cycle. Using two analytical frameworks—Revenue per MW and Capital Expenditure‑Implied Revenue—Morgan Stanley finds that Microsoft is aggressively front‑loading AI data‑center capacity, which temporarily depresses revenue per megawatt. The firm argues that this does not reflect a deterioration in monetization prospects but rather reflects the company’s strategic, proactive investments to meet future AI demand. Accordingly, current revenue forecasts may be overly conservative, leaving significant room for upward revisions. The report maintains an Overweight rating on Microsoft, with a price target of $650.

Core views

Key Point 1: Infrastructure deployment is outpacing monetization, and the decline in revenue per megawatt reflects a structural trend. According to the research report, Microsoft’s data center capacity is projected to expand rapidly from approximately 5 GW in FY24 to about 20 GW by FY28. At the same time, annualized cloud revenue per megawatt of data center capacity is expected to fall from roughly $27.5 million in FY24 to around $17.2 million by FY28. The firm notes that this decline primarily stems from the higher capital intensity and energy consumption associated with generative AI training and inference workloads compared to traditional cloud computing, rather than any weakening in Microsoft’s ability to monetize its infrastructure. Given that Microsoft has built substantial capacity well ahead of full utilization, its existing infrastructure could potentially support significantly higher revenue levels than currently anticipated. Key Point 2: A capex‑based valuation model points to substantial upside potential for Azure revenue. By constructing an income‑forecasting model anchored in capital expenditures (Capex), the report translates Microsoft’s AI‑related Capex into implied Azure AI revenue. The analysis indicates that, provided Azure AI maintains gross margins above 15%–20%, its implied revenue would far exceed current bottom‑up estimates. For instance, under a range of assumed gross margins between 20% and 60%, implied Azure AI revenue for FY29 could reach $142 billion to $284 billion, whereas current consensus forecasts hover around $115 billion to $120 billion. This suggests that the market may be underestimating the long‑term return on Microsoft’s AI investments. Key Point 3: Valuation remains attractive, with earnings growth durability not yet fully priced in. The report assigns Microsoft a target price of $650, based on a 30x price‑to‑earnings (PE) multiple applied to the FY27e earnings per share (EPS) of $21.76. While this valuation is slightly above that of its large‑cap software peers, the premium is justified given Microsoft’s dominant position in public cloud and AI, its extensive distribution network, and its expanding profit margins. The firm believes that, as the cloud‑optimization cycle winds down and commercial‑cloud revenues remain robust, Microsoft’s medium‑ to long‑term earnings growth is highly resilient. Current share prices have yet to fully reflect this durability or the value created by its AI leadership.

Analysis framework

The research report employs two primary quantitative frameworks to assess Microsoft’s AI monetization potential: 1. Revenue per MW Framework: The firm divides Microsoft’s total cloud‑ecosystem revenue—encompassing Azure, Office 365 Commercial, Dynamics 365, and other offerings—by the aggregate installed capacity of its data centers, measured in megawatts (MW). By tracking this metric over time and factoring in management guidance on capacity expansion, the firm concludes that capacity is growing at a pace far outstripping current revenue growth. This approach highlights the “capacity‑frontloading” phenomenon, where infrastructure investments are made well before corresponding revenue streams fully materialize. 2. Capex‑Implied Revenue Framework: Drawing on Microsoft’s disclosed capital‑expenditure plans, the firm isolates spending allocated to AI‑related infrastructure, such as GPU‑powered servers and data‑center construction. By assuming varying gross margin levels (20%–60%) and depreciation schedules (six years for servers, fifteen years for buildings), it retroactively estimates the revenue streams these capital outlays could theoretically generate. Comparing this implied revenue against analysts’ bottom‑up forecasts allows the firm to quantify the potential upside. This methodology is particularly well suited for evaluating the long‑term output efficiency of capital‑intensive technology companies.

Methodology notes

• Industry/ Sector Analysis FrameworkVolume-price decomposition

  Revenue per MW Analysis

  We decompose total revenue into “capacity (volume)” and “revenue per unit of capacity (price/efficiency).” In the AI era, given the extremely high power consumption of AI chips, the denominator—capacity in MW—has expanded rapidly, making revenue per unit appear to decline. However, this is a typical phenomenon during a phase of technological iteration and should be assessed in conjunction with overall revenue growth.

• Company Fundamentals and Financial FrameworkFree cash flow analysis

  Capital Expenditure–Implied Revenue Model

  By tracking the allocation of large‑scale capital expenditures—such as GPU purchases and data‑center construction—and applying assumptions about gross margins and depreciation periods, one can retroactively estimate the future revenue these assets are expected to generate. This serves as a cross‑validation approach to assess whether revenue forecasts are conservative.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Microsoft Corp (MSFT.US)

  Beneficiary: As a global leader in cloud services and AI platform provision, Microsoft has established a competitive edge in AI infrastructure by proactively expanding its data center capacity.

  Strengths

  Strong public cloud positioning (Azure), an extensive customer base, a diversified AI product portfolio (including Copilot and Azure AI), and a robust balance sheet that underpins sustained capital expenditures.

  Comparison

  Compared with other hyperscale cloud providers, Microsoft enjoys a distinct advantage in enterprise‑level software integration—M365 plus Azure—which helps enhance the stickiness of its AI services and improve monetization efficiency.

  Risks

  Macroeconomic weakness is weighing on IT spending; AI adoption has fallen short of expectations; and intensifying competition is putting pressure on profit margins.

Key data

• Data Center Installed Capacity Forecast~5GW (FY24) → ~20GW (FY28)It is expected to quadruple within four years, reflecting an aggressive expansion of AI infrastructure.
• Revenue per megawatt of cloud capacity forecast~$27.5M (FY24) → ~$17.2M (FY28)The decline was primarily attributable to the capital-intensive nature of AI workloads and a deterioration in non‑monetizable factors.
• Target price$650.00Based on CY27e EPS of $21.76 and a 30x P/E multiple, this implies approximately 56% upside potential.
• Upside potential in Azure AI’s implied revenueSignificantly higher than the current forecastIf the gross margin exceeds 15%–20%, the Capex model projects FY29 revenue that significantly surpasses bottom-up consensus estimates.

Impact & implications

For Microsoft, this means that its current substantial capital expenditures are not merely a “cost burden,” but rather a necessary investment to prepare for the explosive growth in AI demand over the coming years. Once the adoption of AI applications—such as Copilot and Azure AI services—accelerates or inference workloads scale up, its existing massive computing infrastructure will quickly translate into robust revenue streams, without requiring commensurate additional infrastructure spending. This, in turn, will enhance operating leverage and expand profit margins. From a market perspective, research reports caution investors against turning bearish on Microsoft simply because of a short-term decline in the “revenue per megawatt” metric; instead, they should focus on the company’s leading position in infrastructure deployment. Should AI adoption outpace expectations, Microsoft’s revenue and earnings forecasts face significant upside risk—meaning actual results could exceed consensus estimates—which could further propel its stock price higher.

Risks

• Weaker macroeconomic conditions have led to a reduction in corporate IT spending.
• The Substitution Effect of On-Premises Deployment on Cloud Services
• Excessive investment has constrained the expansion of profit margins.
• AI adoption has fallen short of expectations, failing to effectively monetize infrastructure investments.

What to watch

• Azure revenue growth rate and the proportion of AI’s contribution
• The actual utilization rate of data center capacity has been trending upward.
• The paid conversion rate of Copilot and other AI applications
• Execution Progress and Adjustments to Capital Expenditure Plans