Equity Research: Microsoft Corporation (MSFT)

Recent News & Earnings

Business Summary

Microsoft is a $281.7B revenue business built on three segments: enterprise software (Office 365, LinkedIn, Dynamics), cloud infrastructure (Azure, Server products, GitHub), and personal computing (Windows, Xbox, Bing). Founded in 1975, the company has evolved from a desktop software monopoly into the world's second-largest cloud provider and, in my view, the best-positioned AI infrastructure company in the public market.

Productivity & Business Processes (~32% of Revenue)

This segment covers Microsoft 365 (Office apps, Teams, Exchange, SharePoint), LinkedIn, and Dynamics 365. Microsoft 365 Commercial seat growth has slowed but revenue continues growing as Copilot AI add-ons drive average revenue per user (ARPU) higher. LinkedIn generated over $17B in annual revenue in FY2025 with growing recruitment and B2B advertising demand. I see this segment as the distribution engine for Copilot monetization.

Intelligent Cloud (~48% of Revenue)

Azure is the growth engine. Azure grew 31% year over year in Q2 FY2026 and has become the primary host for enterprise AI workloads through the exclusive Azure OpenAI partnership. GitHub, now part of this segment, has 1.8M+ paid Copilot subscribers with 77% seat growth year over year. Server products (Windows Server, SQL Server) still generate meaningful revenue, though growth is slower as workloads migrate to Azure.

More Personal Computing (~20% of Revenue)

Windows OEM licensing, Devices (Surface), Xbox gaming, and Bing Search sit here. This is the slowest-growth segment. Xbox Game Pass subscription revenue is growing, but device sales have declined as the PC market normalizes after COVID demand. I do not build significant growth assumptions into this segment for my 3-year model.

AI & Cloud Platform Analysis

Microsoft's AI strategy is not a roadmap — it is a running revenue line. Azure OpenAI is live, Copilot is billing at $30/user/month, GitHub Copilot has 1.8M+ paid users, and Microsoft Cloud crossed $51.5B in a single quarter. I believe MSFT has the deepest AI monetization of any hyperscaler today, backed by exclusive access to OpenAI's frontier models and 365 million existing Office users as a distribution channel.

Azure OpenAI Service

Microsoft holds exclusive rights to host OpenAI models on Azure, including GPT-4o, the o1 reasoning series, and Phi small language models. Enterprise customers accessing these models through Azure do not need to send data to OpenAI directly — they use Microsoft's infrastructure with their own data security terms. This is a significant compliance advantage for regulated industries. Azure AI capacity is the primary driver behind Microsoft's $37.5B quarterly capex run rate.

Microsoft 365 Copilot

Copilot is a $30/user/month add-on to Microsoft 365 that embeds AI into Word, Excel, Outlook, Teams, and PowerPoint. Seat counts grew 3x year over year as of Q2 FY2026. Even at 5% penetration of the 365 million commercial Office users, Copilot would generate over $65B in additional annual revenue. Penetration is currently well below that. I see Copilot as the largest per-seat monetization event in enterprise software history if it scales.

GitHub Copilot & Developer Platform

GitHub Copilot had 77% year-over-year seat growth as of Q2 FY2026, with over 1.8 million paid subscribers. Developers who use Copilot for code completion are significantly more productive, and GitHub's data shows strong net revenue retention. The moat here is network: GitHub hosts over 420 million repositories, meaning the training data and the deployment surface both belong to Microsoft.

Commercial Remaining Performance Obligation (RPO)

Microsoft's commercial RPO stood at $625B as of Q2 FY2026. This is contracted future revenue that has not yet been recognized. $281B of that total (45%) comes from OpenAI contracts. The remaining $344B grew 28%, driven by broad enterprise Azure and annuity commitments. RPO at 2.2x annualized revenue gives me high confidence in near-term revenue visibility.

AI Bubble Risk Assessment

The most common concern I hear about MSFT is that it is an AI bubble stock. I disagree with that framing, but I do think the capex risk is underappreciated. Here is how I think about whether MSFT belongs in the "bubble" category and what the real risks are.

The Capex Surge Is Real

Microsoft spent $28.1B on capital expenditures in FY2023. In Q2 FY2026 alone, it spent $37.5B. That is a 3x increase in spending compressed into two fiscal years. Two-thirds of Q2 FY2026 capex went to short-lived assets (GPUs and CPUs), which depreciate quickly and must be replaced. If Azure AI demand meets projections, this investment earns a strong return. If demand disappoints, there is no secondary market for custom AI hardware at scale.

FCF Compression Is the Real Concern

Revenue grew 14.9% in FY2025. FCF fell from $74.1B to $71.6B. That is a deterioration in capital efficiency, not a bubble signal, but it is a signal that investors should watch. FCF margin dropped from 30.2% in FY2024 to 25.4% in FY2025. If capex stays at current levels and revenue growth decelerates, FCF per share could stay flat for 2 to 3 years. My model assumes capex begins to moderate relative to revenue starting in FY2027 as short-lived GPU assets cycle.

Is MSFT Bubble-Priced?

No. At $408 per share, MSFT trades at 24.1x forward earnings (FY2026E EPS of $16.92). The 5-year average forward P/E for MSFT is approximately 32x. The stock is trading at a 25% discount to its own historical average multiple. A bubble implies multiple expansion on speculative revenue. MSFT is the opposite: the multiple has contracted while real AI revenue is accelerating. A PEG ratio of 1.63x on 24% EPS growth is not bubble math.

OpenAI Dependency Is a Structural Risk

Microsoft's AI differentiation depends partly on its exclusive partnership with OpenAI. $281B of its $625B commercial RPO comes from OpenAI contracts. If OpenAI's models lose their frontier lead, or if the relationship fractures (as the November 2023 Sam Altman governance crisis briefly threatened), Microsoft's AI competitive advantage narrows. I treat this as a tail risk, not a base case, but it warrants position sizing discipline.

My Conclusion

MSFT is not a bubble. It is a proven cash machine front-loading capital investment to capture an AI infrastructure buildout that I believe is real and durable. The risk is that capex ROI arrives slower than expected, compressing near-term FCF. That is a return timing risk, not a fundamental impairment. At 24x forward earnings, the market is not pricing in perfection. There is a margin of safety.

Industry Context

Cloud infrastructure is a winner-take-most market with three dominant players: AWS (33% share), Azure (25%), and Google Cloud (12%). The total cloud infrastructure market is expected to reach $1.5T by 2030, growing at approximately 20% annually. Microsoft is the fastest-growing large hyperscaler in percentage terms. What I find most compelling is that enterprise migration to cloud is still early — a large portion of global enterprise workloads still run on-premises. As those workloads migrate, Azure benefits from Microsoft's existing enterprise relationships without needing to win new customers from scratch.

AI is accelerating the cloud growth cycle. Enterprise AI workloads require GPU compute at scale, and the three hyperscalers are the only credible providers of that infrastructure. Microsoft's OpenAI partnership gives it a differentiated model portfolio that AWS and Google cannot replicate without training their own frontier models. Anthropic's partnership with AWS and Google's Gemini are the closest analogues, but Microsoft got there first and at greater scale. I expect cloud AI to drive above-market growth for Azure through at least FY2029.

Enterprise software is the other key market. Microsoft 365 has over 365 million commercial seats. Salesforce, ServiceNow, and SAP compete on adjacent workflow layers, but none has Microsoft's breadth across productivity, communication, CRM, and cloud infrastructure. The network effect of Teams, SharePoint, and Office together creates a switching cost that I view as durable for the 3-year forecast horizon.

Competitive Advantages & Moat

1. Enterprise Switching Costs

Microsoft 365 is embedded in day-to-day workflows at the majority of Fortune 500 companies. Teams, Outlook, SharePoint, and Excel are not just products — they are how organizations communicate and store institutional knowledge. Ripping out the Microsoft stack requires retraining hundreds of thousands of employees, migrating terabytes of data, and replacing workflow integrations built over decades. The switching cost is so high that most enterprises do not attempt it.

2. Azure + OpenAI Exclusivity

Microsoft's $13B investment in OpenAI gives it exclusive rights to deploy OpenAI's models commercially on Azure. No other cloud can offer GPT-4o or o1 as a managed service. Enterprise customers who want the best frontier models must use Azure. This is a moat that compounds as OpenAI releases more capable models — each new release strengthens Azure's position relative to AWS and Google Cloud.

3. GitHub Network Effects

GitHub hosts over 420 million repositories and is the default version control platform for professional developers worldwide. This creates a compounding data advantage for GitHub Copilot: the more code on GitHub, the better the AI training data, the better the Copilot suggestions, the more developers use GitHub. Competitors like GitLab and Bitbucket are distant alternatives with meaningfully smaller developer communities.

4. Copilot Distribution Flywheel

Microsoft can distribute Copilot to 365 million existing Microsoft 365 commercial users at near-zero incremental acquisition cost. No AI startup can match that distribution. When Microsoft decides to include Copilot in a higher-tier M365 plan, the feature is instantly available to tens of millions of enterprise users. This is why I believe the $30/user/month Copilot add-on can become one of the most significant ARPU expansion events in enterprise software history.

5. Financial Fortress

$94.6B in cash and short-term investments, $34B net cash position, Aaa-rated debt, and $136B in operating cash flow in FY2025. Microsoft can sustain $80B in annual capex while returning $42B+ to shareholders through buybacks and dividends without straining its balance sheet. This financial strength lets it outlast competitors in the AI infrastructure arms race.

Revenue Growth & Profitability

Earnings Per Share

The FY2023 EPS flat year (+0.3%) reflects a $16B goodwill impairment charge related to gaming. The underlying business continued growing. FY2026 consensus EPS growth of 24% is strong, driven by Azure mix shift and Copilot ARPU expansion. My FY2028E of $22.50 implies modest moderation from consensus growth, which I consider appropriate given capex pressure on FCF and the uncertainty around AI product penetration timing.

Balance Sheet

Microsoft's balance sheet is one of the cleanest among large-cap technology companies. Net cash of $34B, Aaa credit rating, and $343B in shareholders' equity give it room to sustain $80B+ in annual capex without cutting shareholder returns. PP&E (net) grew from $70.8B in FY2021 to $229.8B in FY2025, reflecting the AI infrastructure buildout. Goodwill stands at approximately $67B, largely from the Activision Blizzard acquisition. I do not consider Activision goodwill a material risk given strong Xbox Game Pass growth.

3-Year Forecast

I built three scenarios around the core question: how fast does Azure AI demand translate into revenue, and does capex moderate enough to let FCF recover? The base case assumes consensus FY2026 and FY2027 estimates hold, and FY2028 grows at a slightly slower 14% rate as the law of large numbers kicks in. I use 25x P/E on FY2028E EPS of $22.50 to arrive at a $563 base target.

Valuation

I ran a discounted cash flow model using a WACC of 9.0%, consistent with Microsoft's Aaa credit rating, low leverage, and a beta slightly below 1.0. I used a terminal growth rate of 2.5%, the project default. The DCF assumes FCF recovers as capex moderates relative to revenue from FY2027 onward. The resulting intrinsic value of approximately $520 per share is my floor estimate. My 3-year base price target of $563 is anchored to 25x my FY2028E EPS of $22.50, consistent with a quality enterprise software comps multiple.

Sensitivity Analysis

I stress-tested my DCF across a WACC range of 8.0% to 10.0% and a terminal growth rate range of 2.0% to 3.0%. The base case at 9.0% WACC and 2.5% TGR yields ~$520. Even the most conservative scenario (10.0% WACC, 2.0% TGR) produces $412, roughly in line with the current price. This gives me confidence that the downside is limited relative to the upside across a wide range of assumptions.

Key Risks

1. Capex ROI Delay

$37.5B in a single quarter is an extraordinary level of capital spending. If Azure AI demand ramps slower than Microsoft's infrastructure buildout, capex will exceed the revenue it generates for longer than my model assumes. FCF could stay compressed through FY2027, limiting the stock's re-rating potential.

2. OpenAI Partnership Fragility

$281B of Microsoft's $625B commercial RPO comes from OpenAI contracts. If OpenAI loses its frontier model lead to Google DeepMind, Anthropic, or xAI, or if the partnership dissolves, Microsoft's AI differentiation narrows sharply. The November 2023 governance crisis was a warning sign that this relationship is not risk-free.

3. AI Commoditization

Foundation model capabilities are compressing fast. If GPT-4o-level intelligence becomes a commodity available from open-source models, the pricing power for Azure OpenAI tokens erodes. Microsoft's moat would shift entirely to distribution and integration, which is still strong but commands a lower margin.

4. Copilot Penetration Disappointment

My bull case depends on Copilot penetrating a meaningful fraction of the 365 million Microsoft 365 commercial seats. If enterprise adoption stalls at early-majority adoption (5 to 10% penetration), the ARPU uplift falls short. Early data suggests strong retention among Copilot users, but the total addressable upsell is still largely untested at scale.

5. Antitrust and Regulatory Risk

Microsoft's OpenAI partnership has attracted regulatory attention in the EU, UK, and US. The FTC has scrutinized the deal's competitive implications. A forced divestiture or licensing requirement on Azure OpenAI would be a material negative, though I consider this a low-probability outcome over the 3-year horizon.

Final Verdict