Microsoft

Microsoft delivered a textbook beat on April 29, 2026 — Q3 FY2026 revenue of $82.9 billion (+18% reported, +15% constant currency), diluted EPS of $4.27, and operating income of $38.4 billion (+20%) — and the stock fell anyway. The reason: CFO Amy Hood's calendar-year 2026 capex disclosure of $190 billion (up 61% YoY), with $25 billion of that increase attributed not to additional GPU buys but to higher component pricing pressure from constrained supply chains. Mustafa Suleyman, CEO of Microsoft AI since March 2024, simultaneously announced his shift from Copilot operations to leading the company's superintelligence research effort, with former Snap exec Charles Lavin elevated to run consumer Copilot.

What makes Microsoft uniquely positioned in 2026 is the Azure-OpenAI commercial restructuring. Azure revenue grew 40% YoY in Q3 — beating analyst expectations of 39.3% and the buy-side whisper of 38.8% — driven by AI workloads now running at a $37 billion annualized run-rate (+123% YoY). This figure includes both third-party model providers (OpenAI, Anthropic via Bedrock-style integrations) and Microsoft's own Copilot family. The new OpenAI agreement, signed Q4 2025, gives Microsoft a 27% economic interest in OpenAI's for-profit entity through 2032 plus capped revenue-share economics on AI infrastructure spend — meaning every dollar OpenAI ships to Azure increasingly flows back through that economic interest line.

Microsoft's smart-money holders we track include Terry Smith of Fundsmith Equity Fund (London-based, $25B AUM), who has held MSFT continuously since 2014 and has been adding modestly through 2025-2026. Terry Smith's thesis is the simplest of the BMI-tracked managers: he treats Microsoft as a Buffett-grade compounder where regulatory risk is materially lower than for Alphabet or Apple, and where capex spending — even at $190 billion — is being directly absorbed by Azure revenue growth that exceeds 40%. Smith trimmed slightly into the post-earnings dip but did not exit any portion of the position.

Christopher Hohn's TCI Fund Management increased its MSFT position by approximately 8% during Q1 2026 according to the Q1 13F filing. TCI's thesis emphasizes the Activision integration value-creation (paid down via $69B acquisition closing 2023), with operating leverage now visible in the Gaming segment's 35% operating margin in Q3 FY2026. Insider activity at Microsoft has been very routine — Satya Nadella's last 10b5-1 sale was December 2025 at $432, and Amy Hood's last sale was January 2026 at $416. CTO Kevin Scott has not transacted since promotion to that role. The most-watched smart-money event remains Mustafa Suleyman's compensation disclosure in the next proxy filing, which will reveal how Microsoft has structured incentives around the superintelligence program.

Microsoft trades at $432 per share, roughly 32.8× consensus FY2026 EPS of $13.15 — a meaningful premium to Alphabet (24.5×) and Amazon (38× but with lower margin profile). The post-earnings -7% drop has compressed the multiple from a recent high of 36×, but Microsoft still commands the highest enterprise software multiple among mega-caps. Wall Street consensus across 56 covering firms averages $510 (Morgan Stanley $540, JPMorgan $525, Goldman $515, Bernstein the bear at $440), implying ~18% upside. The premium to Alphabet is justified by Azure's higher-margin profile, the OpenAI economic interest, and lower antitrust regulatory exposure — but compressing if 40%+ Azure growth fails to sustain through fiscal 2027.

1. July 30, 2026 (estimated): Q4 FY2026 earnings — first quarter to test whether the post-Q3 capex shock has been fully digested by guidance; CFO Amy Hood will provide initial fiscal 2027 capex framework which markets will read as the central valuation input
2. Mid-2026: MAI-2 frontier model launch (Microsoft's own family, succeeding April 2026's MAI-1 mid-class) — Suleyman's first chance to demonstrate independent frontier capability and reduce OpenAI partnership risk
3. H2 2026: Vera Rubin deployment ramp at Fairwater AI superfactories — first quarters of revenue contribution from Microsoft's largest single-purpose AI infrastructure investment, validating or invalidating the $190B capex thesis

Microsoft is the most expensive of the four mega-cap AI bets I track, and the post-earnings -7% reaction tells you the market is starting to price the limits of capex-funded growth. I'm patient on this name — the OpenAI economic-interest structure is genuinely valuable, Azure's 40% growth is sustainable in my view, and Suleyman's superintelligence pivot reduces a structural risk that's been overhanging the stock since 2023. But at 32.8× forward, you're paying for execution that still has to be delivered. My add-trigger is a clear demonstration that Q4 FY2026 capex is not accelerating beyond the $190B run-rate — that would signal Microsoft has found the upper bound of its AI infrastructure commitment. I would buy more aggressively in the $400-410 range, where the multiple compresses below 30× and where Terry Smith's average cost remains underwater. The thesis breaks if Azure deceleration drops below 30% growth before fiscal 2027.