Deutsche Bank Maintains Buy on Microsoft: Azure AI Margins Under Pressure but the Value Creation Thesis Remains Intact

Overview

This report is part of Deutsche Bank's research series on the economics of AI cloud infrastructure, focusing on Microsoft's Azure and Intelligent Cloud businesses. The core conclusion is that although Azure gross margin has come under clear pressure in recent years due to changes in the AI revenue mix, pre-revenue capacity expansion costs, and fading depreciation accounting benefits, the current AI cloud business still has positive economic returns and a sustainable commercial foundation. Deutsche Bank believes Microsoft's highly profitable core businesses, strong operating leverage, and sustained ability to generate capital returns enable it to maintain and gradually expand company-level operating margins even during a heavy capital expenditure cycle.

Core views

Deutsche Bank maintains a Buy rating and a $550 target price on Microsoft. The key views are: first, Azure gross margin may bottom out and stabilize in the low-50% range in FY27/28, and could recover to the mid-50% range if AI service efficiency improves faster; second, Azure AI Services gross margin is estimated in the mid-to-high 20% range for FY26 and could rise to the low-30% range in FY27 as scale, engineering efficiency, the pricing environment, revenue-sharing changes, and GitHub Copilot pricing adjustments take effect; third, Intelligent Cloud operating margin remains resilient, with expense growth significantly below revenue growth, offsetting most gross margin pressure; fourth, although the company's return on incremental invested capital is below historical levels, it has still remained in the mid-20% range in recent years, supporting the view that value creation continues through the AI-driven investment cycle.

Analysis framework

The report uses a segment breakdown and bottom-up margin estimation approach. Because Microsoft does not directly disclose Azure costs and margins, Deutsche Bank starts with disclosed Intelligent Cloud revenue and costs, assumes Server Products gross margin of 94.5%, Enterprise Services gross margin of 27.5%, and Other gross margin of 30.0%, and then backs out Azure revenue, gross profit, and gross margin. It also cross-checks the results using Microsoft Cloud's overall revenue and gross margin, deriving an ex-Azure cloud gross margin of about 80%, which is consistent with SaaS businesses such as M365. The report further breaks Azure gross margin changes into the effects of AI revenue mix, pre-revenue capacity expansion costs, and changes in depreciation useful lives, and adjusts Intelligent Cloud operating profit for capitalization and amortization of pre-revenue capacity expansion costs.

Methodology notes

• Segment Margin DecompositionBottom-up Intelligent Cloud Gross Margin Estimation

  Using disclosed Intelligent Cloud revenue and costs, plus gross margin assumptions for Server Products, Enterprise Services, and Other, to back out Azure gross margin.

  Deutsche Bank assumes Server Products gross margin of 94.5%, Enterprise Services at 27.5%, and Other at 30.0%, and based on this estimates Azure gross margin declining from about 59% in FY24 to about 52% in FY26 and about 51% in FY27.

• Cross-checkBack-solving Microsoft Cloud ex-Azure Gross Margin

  Using Microsoft Cloud overall revenue and gross profit minus estimated Azure figures to verify whether the remaining cloud business gross margin is reasonable.

  The back-solved result shows Microsoft Cloud ex-Azure gross margin at about 80%, consistent with M365 Commercial Cloud and large SaaS businesses with gross margins in the low 80% range, thereby increasing confidence in the Azure gross margin estimate.

• Operating Profit NormalizationCapitalization and Amortization Adjustment for Pre-revenue Capacity Expansion Costs

  Treating expansion costs such as rent and electricity incurred before new capacity goes live and generates revenue as economic costs that should be amortized over the six-year useful life of cloud computing equipment.

  On top of Intelligent Cloud operating profit, Deutsche Bank adds back current-period pre-revenue capacity expansion costs, amortizes them on a straight-line basis over six years, and excludes depreciation benefits from changes in useful lives for servers and networking equipment to observe a more realistic underlying earnings profile.

• Efficiency FrameworkMargin Accountability Program and Compound S-Curve

  Continuously improving the unit economics of AI infrastructure through data center design, GPU/CPU resource substitutability, and token output per dollar per watt.

  Deutsche Bank believes Microsoft has already improved gross margins over time in its non-GPU Azure business through similar engineering efficiency gains, providing a reference for future narrowing of the gross margin gap between Azure AI and non-AI businesses.

Asset mapping & comparison

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

• MICROSOFT CORP / MSFT.US

  Core covered name; the report focuses on the profitability of Azure, Azure AI Services, and Intelligent Cloud.

  Strengths

  Highly profitable core businesses, Azure scale advantage, strong operating leverage, robust AI demand, a full Azure AI technology stack, continued OpenAI-related revenue sharing, and still-healthy company-level ROIC.

  Weaknesses

  A rising share of AI revenue depresses Azure gross margin, pre-revenue capacity expansion costs are increasing, and depreciation benefits from earlier useful-life changes for servers and networking equipment are fading.

  Comparison

  Compared with AI infrastructure participants such as CoreWeave and Oracle, Deutsche Bank believes Azure AI Services' mid-to-high 20% gross margin is above its estimate of mid-to-high teens AI gross margins for CoreWeave and Oracle over the last twelve months, mainly due to Microsoft's scale economies and richer mix of upper-layer services.

  Risks

  A sharp increase in R&D spending, expanded subsidies for first-party AI applications, rising component prices, delays in capacity delivery, and slower-than-expected narrowing of the gross margin gap between AI and non-AI businesses.

• COREWEAVE INC / CRWV.US

  A comparable company and prior research subject for AI cloud infrastructure economics.

  Strengths

  Benefits from strong demand for AI infrastructure, with stable contract profitability mentioned as being in the mid-20% range.

  Weaknesses

  Deutsche Bank estimates its AI gross margin over the last twelve months is below Azure AI Services.

  Comparison

  Azure AI Services is estimated to have a gross margin in the mid-to-high 20% range, above the estimated AI gross margins for CoreWeave and Oracle in the mid-to-high teens.

  Risks

  GPU depreciation, financing costs, contract concentration, and execution risk in capacity buildout.

• ORACLE CORP / ORCL.US

  A comparable company and prior research subject in AI cloud infrastructure.

  Strengths

  Management targets suggest steady-state AI OCI gross margin could reach 30%-40%.

  Weaknesses

  Current AI business gross margin is still estimated to be below Azure AI Services.

  Comparison

  Deutsche Bank uses Oracle AI OCI and Azure AI Services as horizontal benchmarks for cloud AI economics.

  Risks

  Capital spending on large-scale AI contracts, customer concentration, buildout and delivery, and component price risks.

• ALPHABET INC / GOOGL.US

  A comparable large-cap technology company in the cloud and AI ecosystem.

  Strengths

  Has scale advantages in AI infrastructure and cloud businesses.

  Weaknesses

  The report references it only for industry background and management commentary, without a formal rating conclusion.

  Comparison

  Like Microsoft, it is a large provider of cloud and AI infrastructure and faces the challenge of balancing AI investment with margins.

  Risks

  Returns on AI capex, cloud competition, and pricing pressure.

Key data

• RatingBuyDeutsche Bank maintains a Buy rating on Microsoft.
• Target PriceUSD 550.00The report's cover page discloses a target price of $550.
• Current PriceUSD 365.46The price reference date is June 24, 2026.
• Implied Upsideapproximately 50.5%Estimated based on the $550 target price and the current price of $365.46.
• FY24 Azure Gross Margin Estimateapproximately 59.4%Deutsche Bank bottom-up estimate.
• FY26E Azure Gross Margin Estimateapproximately 52.0%Azure gross margin is pressured by the AI mix and expansion costs.
• FY27E Azure Gross Margin Estimateapproximately 50.6%The report believes gross margin may stabilize in the low-50% range.
• FY26 Azure AI Services Gross Margin Estimatemid-to-high 20% rangeDeutsche Bank estimates this based on non-AI Azure gross margin, company disclosures, management commentary, and industry data.
• FY27 Azure AI Services Gross Margin Outlooklow-30% rangeDriven by scale, efficiency, pricing, and revenue-sharing adjustments.
• FY26 Pre-revenue Capacity Expansion Cost as % of Revenueapproximately 2.1%Deutsche Bank estimates this headwind peaks in FY26 and falls to about 1.8% in FY28.
• FY26 Normalized Intelligent Cloud Adjusted Operating Marginapproximately 42.1%Adjusted for pre-revenue capacity expansion costs and useful-life accounting effects.
• Intelligent Cloud FY24-FY28 Adjusted Operating Marginapproximately 42.1%-42.7%Deutsche Bank expects underlying operating margin to remain broadly stable.

Impact & implications

For investors, the report weakens the bearish narrative around whether Microsoft's AI cloud contracts generate positive returns. Deutsche Bank argues that the near-term decline in gross margin is more a result of accounting and temporary pressure from revenue mix and capacity expansion timing, rather than a failure of the business model. If Azure AI gross margin rises from the mid-to-high 20% range to the low-30% range as expected and gradually converges toward non-AI Azure, Microsoft's AI infrastructure investments can continue to translate into revenue growth, operating leverage, and capital returns. The key investment implication is that support for MSFT's valuation comes not only from AI revenue growth, but also from its ability to sustain margins and ROIC during a high-capex cycle.

Risks

• If Microsoft significantly increases R&D spending to develop frontier models in-house, it could weigh on mid-term operating margin improvement.
• If subsidies for first-party AI applications expand, Intelligent Cloud and overall company margins could come under pressure.
• Rising prices for AI hardware components could compress margins on the existing backlog and new contracts.
• AI capacity buildout is complex, and delivery delays could affect revenue recognition and cost absorption.
• If the gross margin gap between Azure AI and non-AI businesses narrows more slowly than expected, the stabilization and recovery of Azure gross margin could lag Deutsche Bank's base case.
• Although pre-revenue capacity expansion costs are expected to peak as a share of revenue, the absolute dollar amount may continue to rise.
• Investors cannot precisely isolate AI-specific returns from public disclosures, and the report's conclusions rely on multiple estimation assumptions.

What to watch

• Whether Azure AI Services revenue exceeds about one-third of total Azure revenue in FY27.
• Whether Azure overall gross margin stabilizes in the low-50% range in FY27/28, and whether there are signs of recovery toward the mid-50% range.
• Whether Intelligent Cloud expense growth can remain in the mid-single-digit range and continue to trail revenue growth.
• Whether GitHub Copilot's shift to usage-based pricing can ease the margin pressure previously caused by fixed-seat pricing.
• Whether after revisions to the OpenAI partnership, Microsoft's cessation of certain Azure OpenAI Service revenue-sharing payments lifts Azure AI Services gross margin by about 1 percentage point as expected.
• Whether tight AI infrastructure supply-demand conditions and pricing on new contracts continue to support better unit economics.
• Whether growth rates of short-lived assets such as servers, networking equipment, and GPUs decline after peaking in FY26.
• Whether Microsoft increases investment in frontier model training, causing R&D spending and AI application subsidies to exceed expectations.