TIP813: Microsoft (MSFT): Is Microsoft a Misunderstood AI Opportunity? w/ Daniel Mahncke & Shawn O’Malley

What It Covers

Daniel Mahncke and Shawn O'Malley analyze whether Microsoft, trading at a forward PE of 20x after a 35% six-month selloff, represents a misunderstood AI opportunity similar to Alphabet in early 2025. They examine all three business segments, the OpenAI partnership dynamics, Azure's competitive positioning, CapEx implications, and a base-case valuation of approximately $500 per share versus a bear-case of $280.

Key Questions Answered

• •Valuation Reset: Microsoft's forward PE compressed from 40x to 20x in six months despite reporting 17% revenue growth, 21% operating income growth, and 60% earnings growth on a $350B revenue base. The selloff was triggered by Azure growing 39% instead of the expected 40%. Investors applying a simple framework here: premium multiples require zero visible risk, and any plausible bear case causes dramatic multiple compression regardless of underlying business performance.
• •The $70B Profit Pool Risk: Microsoft's productivity software segment generates roughly $70B in annual operating profit at ~60% margins from 450 million commercial Office 365 seats at $25/user/month. The AI disruption threat is not competitor switching — it is seat count stagnation. As LLMs handle cognitive tasks previously requiring multiple human workers, enterprises resist price increases on tools used less frequently, compressing average revenue per user without mass cancellations.
• •Christensen's Dilemma Reframed: Google Docs failed to disrupt Microsoft Office over 20 years because competing products face identical structural barriers: file format standards embedded in legal contracts, enterprise power-user requirements that Google Sheets cannot meet, and cross-company network effects. AI disruption operates differently — it reduces the human labor intensity of tasks rather than offering an alternative tool, making the $70B profit pool vulnerable in a way no prior competitor achieved.
• •Azure's Four-Layer Stack Problem: Microsoft's ideal AI economics require owning all four value chain layers simultaneously: software output layer (Office), platform orchestration layer (Azure AI Foundry), AI model layer (Copilot/MAI-1), and compute infrastructure (Maya 200 chips). Currently, Copilot runs primarily on OpenAI models, meaning Microsoft pays OpenAI a cut on every response. Enterprises using Anthropic's Claude route compute through AWS, not Azure, eliminating Microsoft's infrastructure revenue from those workloads entirely.
• •OpenAI Partnership Deterioration: Microsoft committed $13B to OpenAI in early 2023 for 49% economic interest and exclusive Azure cloud rights, but much of that investment consisted of redeemable compute vouchers rather than cash. OpenAI subsequently signed a $50B enterprise deal routing workloads through AWS via a product called Frontier, arguing it falls outside the exclusivity agreement. Anthropic's $30B+ revenue run rate now exceeds OpenAI's, and OpenAI's nonprofit-to-for-profit conversion creates uncertainty about Microsoft's equity stake conversion.
• •CapEx Math and Margin Risk: Microsoft's CapEx reached $70B in just the first half of fiscal 2026, with full-year guidance of $120-150B versus $28B total in fiscal 2023. Two-thirds targets short-lived GPU and CPU assets with 3-5 year useful lives, creating guaranteed future depreciation charges. At $140B annual spend, Azure must generate $17-20B in incremental annual revenue just to clear a 12-15% cost of capital hurdle. Azure's current ~$28B annual revenue addition clears that bar only if 40% growth rates hold.
• •GitHub as Structural Moat: Microsoft acquired GitHub in 2018 for $7.5B when it had 28 million users; it now has 100 million registered developers and is the primary channel through which enterprises make cloud infrastructure decisions. GitHub Copilot demonstrates measurable ROI by accelerating code shipping velocity 20-30x. Because developers building on GitHub and Visual Studio default to Azure deployment, this acquisition functions as a developer-relationship capture mechanism that structurally advantages Azure over AWS and Google Cloud in enterprise onboarding.