Amazon, Alphabet, and Microsoft: Two I'm Buying and One I'm Selling

Article highlights accelerating cloud growth tied to the AI build-out: Google Cloud revenue up 82% YoY in Q2 (vs 63% in Q1) and with operating margin rising 21% to 36% YoY. AWS grew 37% in Q2 vs Azure 43% (but Q1 showed AWS 28% vs Azure 40%), with the author flagging Azure’s growth as lagging peers despite Microsoft appearing cheapest on an operating-earnings valuation view. Overall, the piece prefers Alphabet and Amazon over Microsoft for faster growth and long-term market-share capture.

Analysis

This is less a “cloud is growing” story than a supply-allocation story: in a constrained capacity market, the hyperscaler that turns capex into usable GPU-hours fastest should win the next wave of enterprise AI workloads. That favors AMZN and GOOG over MSFT near term because the market tends to reward accelerating share capture before it rewards absolute scale; if that persists for 1-2 quarters, MSFT’s cheaper multiple can stay cheap while investors rotate toward the names with the sharper revenue slope.

Second-order winners are the picks-and-shovels names that monetize the buildout regardless of which cloud wins: NVDA first, then networking, power, and cooling suppliers. The risk is that cloud gross margins get temporarily masked by depreciation and power costs, so headline operating leverage can look worse before it looks better; that usually creates a better entry point for the infrastructure complex than for the hyperscalers themselves.

The contrarian miss on MSFT is that slower Azure growth may be more of a capacity timing issue than demand loss, meaning the market could be underestimating a re-acceleration once supply comes online. The bigger falsifier for a short-MSFT/long-AMZN- or GOOG-relative view is a next-print Azure step-up or evidence that Microsoft is monetizing AI via bundled software more efficiently than cloud peers. Conversely, if GOOG/AMZN growth decelerates while capex stays elevated, the market will punish FCF quality within months, not years.

Key Decisions for Investors

• Long GOOG vs short MSFT as a 1-3 month relative-value trade; enter on any post-earnings pullback, targeting a rerating if cloud growth differentials persist, and cover if Azure re-accelerates materially next quarter.
• Overweight AMZN and GOOG on a 6-18 month basis versus MSFT for AI-capacity share gain; the risk/reward is strongest if cloud revenue acceleration continues to outrun incremental capex.
• Add NVDA on weakness as the cleaner second-order beneficiary of sustained hyperscaler capex; this is a 6-12 month trade, but trim if datacenter spend guidance rolls over or lead times normalize faster than expected.
• Watch for a short opportunity in MSFT only if the next Azure update again shows flat-to-decelerating growth while capex remains elevated; absent that confirmation, the valuation discount may already be sufficient.