Billionaire Bill Ackman Just Sold Amazon and Bought Microsoft Stock. Which Cloud Computing Behemoth Is the Better Buy Today?

Pershing Square cut its Amazon stake by over 25% in Q2 while increasing its Microsoft stake by nearly 10%, but the article argues the post-Q2 performance and fundamentals have shifted. Since Q2, Microsoft is up ~30% vs Amazon up ~10%, yet the piece highlights stronger Q2 operating momentum at Amazon (20% growth and +43% operating income, driven by AWS with 64% YoY operating income growth) alongside $220B of data-center capex in 2026. Valuation is framed as more favorable for Amazon on operating income growth, with the implication that Amazon may be the better buy versus Microsoft despite Ackman’s prior trade.

Analysis

The key market mechanism is not “which stock is cheaper,” but which hyperscaler converts incremental AI capex into incremental operating income with less lag. AMZN has more earnings torque because cloud sits deeper in the profit mix, but that also makes it the more execution-sensitive name if utilization or pricing disappoints; MSFT is the steadier compounder, yet it can still underperform on multiple expansion if investors conclude the AI capex race is becoming less capital-efficient.

Second-order winners are the infrastructure suppliers, not the platforms. Every additional dollar of data-center build-out flows first to accelerators, networking, power, cooling, and construction ecosystems; that makes NVDA the cleaner expression than either cloud owner if the build cycle stays intact. The loser is “efficient growth” software valuation more broadly: if investors normalize higher capex as the new baseline, the market may stop paying a premium for MSFT’s quality and instead reward revenue acceleration, even at the cost of lower near-term FCF.

Near term, this is mostly a momentum trade and the 13F signal is stale. Over 1-3 months, the catalyst is whether management commentary confirms booked demand behind the spending; over 6-18 months, the risk is overbuild and amortization drag, which would force margin resets and multiple compression. The thesis is falsified if AWS growth decelerates despite higher spend, or if MSFT’s capex finally translates into visibly stronger cloud monetization without a balance-sheet penalty.

Key Decisions for Investors

• Prefer NVDA over AMZN/MSFT as the cleaner 6-12 month AI capex beneficiary; buy on any 5-7% pullback, with the main risk being a slowdown in hyperscaler spend or tighter export controls.
• Do not chase MSFT after the post-earnings rerating; wait for a 5-8% pullback or a confirmatory capex step-up before adding. Falsifier: next cloud update shows acceleration with stable margins.
• Use AMZN/MSFT as a relative-value pair only on weakness: long AMZN / short MSFT if AMZN gives back part of the post-earnings move and AWS growth remains the faster comp. Stop if MSFT outperforms AMZN by ~10% from entry or Amazon capex raises FCF concerns.
• Set an alert on hyperscaler free-cash-flow margins and capex-to-revenue ratios over the next two quarters; if both compress while revenue merely holds, fade the entire cloud complex via XLK or reduce exposure to both names.