This Famous Investor Just Sold All Of His Alphabet Stock and Loaded Up On Microsoft Stock. Should You?

Bill Ackman’s Pershing Square fully exited Alphabet and built a 5.65 million-share Microsoft stake, funded by the GOOG sale. The article argues Microsoft’s valuation is reasonable but its Azure growth has stabilized in the high-30% range, while Alphabet’s Google Cloud grew 63% and AWS grew 28%, potentially offering better risk-reward. Microsoft also faces higher 2026 capex of about $190 billion and a looser OpenAI moat after revised partnership terms.

Analysis

The important signal is not the headline switch from one mega-cap to another; it is that AI infrastructure spending is still in an acceleration phase, but the market is starting to reward the “next layer” of beneficiaries more than the platform incumbent. That favors the names with higher marginal demand surprise and more operating leverage to AI-capex cycles — currently the cloud back ends rather than the software front ends. In that setup, Microsoft can remain fundamentally strong while still lagging on multiple compression if investors conclude its AI monetization is more gradual than its capex burden.

The second-order issue is competitive positioning in cloud. If hyperscaler growth re-rates on acceleration rather than absolute scale, the market will likely pay more for companies that can show rising backlog conversion and faster incremental revenue per unit of AI spend. That is structurally more constructive for GOOGL and AMZN than MSFT over the next 2-4 quarters, especially if Azure growth plateaus while Google Cloud and AWS keep inflecting. The consequence is that capex-heavy leaders may face a temporary valuation tax until the market sees proof that spend is converting into durable free-cash-flow expansion.

A more subtle risk is that the OpenAI partnership reset reduces the exclusivity premium around Microsoft’s AI ecosystem just as the company is increasing capital intensity. That does not break the thesis, but it lowers the scarcity value of MSFT’s AI distribution moat and shifts bargaining power toward model providers and competing clouds. Meanwhile, the possibility that AI copilots cannibalize parts of enterprise software is probably a years-long rather than days-long risk, but it matters because it caps how much multiple expansion Microsoft 365 can support absent clear productivity monetization.

The consensus may be overestimating how much a high-quality compounder can rerate when earnings are already good and underestimating how fast relative winners rotate inside AI. For the next 3-6 months, the more likely trade is not "short Microsoft" but "long the faster-accelerating cloud beneficiaries versus the one making the largest AI investment commitment." If Microsoft’s next few quarters show any deceleration in Azure or weaker AI payback evidence, the stock could underperform again despite solid fundamentals.

Key Decisions for Investors

• Go long GOOGL / short MSFT as a 3-6 month relative-value trade: thesis is faster cloud acceleration and better AI optionality at a lower capex burden; stop if MSFT re-accelerates Azure or if GOOGL cloud growth stalls.
• Add AMZN vs MSFT pair on a 2-4 quarter horizon: AWS growth is still catching up from a lower base and offers more upside if incremental AI workloads migrate across clouds; risk is Amazon margin noise from retail.
• Buy MSFT only on a pullback, not strength: prefer waiting for evidence that AI capex is translating into backlog conversion and free-cash-flow resilience; upside is quality compounding, but near-term risk/reward is capped by valuation and spending intensity.
• Use GOOGL call spreads or AMZN call spreads into the next earnings cycle to express cloud-acceleration upside with defined risk; these are better than outright equity if the market keeps rotating among mega-cap AI winners.
• Avoid shorting MSFT outright unless Azure growth clearly rolls over for two consecutive quarters; the better expression is a pair trade, since the downside is more likely to be relative underperformance than absolute de-rating.