Amazon, Alphabet, and Microsoft: 2 Cloud Giants I Am Buying in September and 1 I Am Avoiding

Alphabet's Google Cloud Q2 revenue grew 82% year over year with a 36% operating margin, while Amazon Web Services revenue increased 37%, accelerating from low-20% growth last year. Microsoft Azure grew 43%, but its pace was only modestly above the prior two quarters' 40% and 39%, leading the author to favor Alphabet and Amazon over Microsoft. The article argues that AI-driven demand and large-scale data-center spending—about $200 billion by Alphabet this year—should sustain cloud growth, although this is opinion-based analysis rather than new company guidance.

Analysis

The relative-growth framing is insufficient to infer durable share loss for MSFT. Azure’s enterprise installed base, bundled Copilot distribution, and hybrid/regulated-workload positioning can monetize AI through software ARPU even if reported cloud growth does not reaccelerate; AMZN and GOOG carry greater risk that incremental AI capacity is lower-return infrastructure revenue. The critical variable is not headline growth but incremental cloud gross margin, utilization of new capacity, and whether enterprise AI workloads convert from experimentation into recurring consumption.

The near-term beneficiary of hyperscaler capacity additions is NVDA, but the second-order readthrough is more nuanced: sustained capex favors networking, power and cooling suppliers while raising the risk of a 2027 digestion phase if capacity comes online ahead of monetizable demand. For AMZN and GOOG, consensus may be underpricing depreciation and power costs that can lag revenue growth by several quarters; a strong revenue print without operating-margin expansion would challenge the bullish interpretation. Conversely, MSFT’s slower infrastructure growth could reflect supply allocation, customer mix, or deliberate prioritization of higher-value services rather than competitive slippage.

This article alone is not a stand-alone directional catalyst; the relevant 1-3 month events are each company’s AI contribution to backlog/RPO, capex guidance, cloud margin, and management commentary on GPU availability. Over 6-18 months, the likely winner is the provider demonstrating that AI revenue exceeds the combined burden of depreciation, energy, and customer incentives. A sustained deterioration in Azure growth relative to peers alongside declining remaining performance obligations would validate the bearish relative thesis; absent that, shorting MSFT on growth-rate comparisons is vulnerable to valuation and buyback support.

Key Decisions for Investors

• Maintain a market-neutral watch pair: long GOOG / short MSFT only after the next earnings cycle confirms a widening gap in cloud growth and incremental operating margin. Target 10-15% relative return over 3-6 months; stop if Azure reaccelerates by at least 300bps or GOOG’s cloud margin contracts despite revenue growth.
• Prefer AMZN over GOOG for a long hyperscaler allocation if AWS growth converts into consolidated operating-income upside while retail margins remain stable. Size after confirmation that incremental AWS margin holds; the principal risk is capex/depreciation absorbing the cloud revenue upside, which would make the position vulnerable to a 10%+ multiple reset.
• Do not initiate an outright MSFT short from this signal. Instead, set an alert around Azure growth, commercial RPO, and Copilot monetization: a reacceleration in any two metrics would invalidate the relative-underperformance narrative and could support covering the MSFT leg of a pair rapidly.
• Use NVDA as the cleaner 1-2 quarter capacity-build beneficiary only if hyperscaler capex guidance is raised or maintained. Reduce exposure if aggregate capex is maintained but management commentary shifts from supply constraint to utilization concern, the earliest sign that infrastructure demand is normalizing.