The Better AI Cloud Stock: Microsoft or Amazon?

Microsoft's Azure grew 40% year over year in fiscal Q3 2026, ahead of AWS's 28% growth, while Microsoft also posted $54.5 billion in cloud revenue, a 46.3% operating margin, and $10.2 billion returned to shareholders. Amazon's AWS reaccelerated to its fastest growth in 15 quarters, but Amazon trades at a higher P/E of about 29 versus Microsoft's roughly 23 and has seen free cash flow fall to about $1 billion over the last year. The article favors Microsoft on the combination of faster cloud growth and cheaper valuation, though both companies face heavy AI capex and margin pressure.

Analysis

The market is beginning to separate “AI beneficiaries” into two distinct cohorts: those with operating leverage from installed enterprise demand and those still funding growth with near-zero free cash flow. In that framework, Microsoft screens as the cleaner monetization vehicle because it is converting AI/capacity spend into durable contractual backlog faster than it is diluting margins, while Amazon is still in the phase where growth is outrunning cash generation. That matters because the next leg of re-rating will likely be driven less by headline cloud growth and more by proof that each incremental dollar of AI capex produces a return above the cost of capital.

The second-order winner is the semiconductor and networking stack, but only selectively. If both hyperscalers keep spending near current levels, the bottleneck shifts from raw compute to power delivery, cooling, optics, and high-end memory; that argues for a broader infrastructure basket rather than chasing the cloud names alone. The loser, in relative terms, is any enterprise software vendor whose AI story depends on Azure/AWS pass-through without controlling enough of the workflow to capture pricing power.

The main contrarian risk is that the market is underestimating how fast AI capex can compress reported free cash flow and cap ex intensity for longer than expected. Over the next 2-4 quarters, a single quarter of slowing backlog conversion or any evidence that customers are not consuming capacity as quickly as it is built could trigger multiple compression even if revenue growth stays strong. Conversely, if AWS closes the growth gap while Microsoft’s margins stabilize, the valuation discount on AMZN could narrow faster than expected, making this a relative-value, not absolute, call.

Key Decisions for Investors

• Long MSFT / short AMZN on a 6-12 month horizon; thesis is that Microsoft deserves a higher multiple for superior cash conversion and dividend support while both spend heavily. Target 8-12% relative outperformance, with risk that sustained AWS reacceleration closes the gap.
• If initiating fresh exposure to hyperscalers, buy MSFT on weakness rather than chase strength; the recent drawdown creates a better entry point for a stock with lower forward P/E and better capital returns. Use a 3-6 month window and size for modest multiple expansion rather than explosive upside.
• Pair long NVDA with short a basket of cloud names only if you want to express capex intensity; the better AI trade is still picks-and-shovels over direct cloud monetization. Risk is that cloud margins surprise to the upside and the basket underperforms less than expected.
• Avoid chasing AMZN until free cash flow inflects meaningfully; the market is already rewarding the acceleration, so the asymmetry is poorer than MSFT. Consider entry only on a pullback or if AWS growth re-accelerates again without further FCF deterioration.