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

Google Cloud is leading hyperscaler acceleration, with Q2 revenue up 82% y/y and operating margin expanding 2100bps to 36%, after Q1 growth of 63% (noting the potential to reach triple-digit growth as more capacity comes online). AWS is also accelerating (Q2 37% vs Azure’s 43%, but Q1 AWS 28% vs Azure 40%), while the article flags a potential red flag in Microsoft if Azure growth doesn’t re-accelerate. Despite Microsoft being described as the cheapest on an operating-valuation basis, the piece favors Amazon and Alphabet over Microsoft for long-term AI cloud market-share gains.

Analysis

The market is likely to keep rewarding the name that can turn incremental AI demand into the fastest near-term capacity monetization, but that does not automatically mean the cheapest stock wins. In the next 1-3 months, relative performance should be driven less by headline valuation and more by whether each platform can show accelerating backlog conversion; if one platform is merely “keeping up,” its multiple can stay stuck even if the absolute business is still excellent.

The second-order winner from continued cloud acceleration is the AI infrastructure stack, not just the hyperscalers. Faster utilization and earlier capacity openings argue for persistent upside in GPU demand, networking, and power/cooling supply chains; that is supportive for NVDA on the margin and for data-center-linked real assets over 6-18 months, because every surprise to demand tends to get reinvested into capex rather than harvested as margin.

The contrarian point is that investors may be overpaying for growth-rate momentum while underestimating the base effect: once cloud growth crosses a high hurdle, the market starts requiring proof of durable share gains rather than temporary capacity catch-up. That makes the cheapest name potentially a value trap if its growth trajectory does not inflect; conversely, the faster growers can still disappoint if capex expands faster than revenue recognition, pressuring near-term FCF and keeping the stock volatile around earnings. A break in growth acceleration next quarter would be the clearest falsifier for the pro-growth thesis.

Key Decisions for Investors

• Relative-value trade: long GOOGL/AMZN vs short MSFT, sized market-neutral, for the next 1-2 earnings cycles. Thesis is that accelerating cloud growth should keep premium multiples attached to the faster growers while MSFT’s lower valuation only matters if Azure reaccelerates meaningfully; cover if Azure growth inflects or if the spread narrows by ~15-20%.
• If already long mega-cap AI, rotate part of the exposure from MSFT into AMZN and GOOGL rather than adding gross risk. Better risk/reward is in the names with visible acceleration; the cheaper name offers less upside if the market is simply repricing lower growth expectations.
• Tactical long NVDA on any broad AI pullback over the next 2-6 weeks. Cloud acceleration implies capex persistence, which tends to extend the GPU order cycle; risk is that hyperscalers stretch depreciation schedules or delay deployments, which would show up first as weaker guidance from the next cloud earnings prints.
• Watchlist alert: if the next cloud update shows no further acceleration, fade the “AI capex wave” trade via QQQ or XLK hedges. The consensus is currently extrapolating capacity additions into growth acceleration; if that bridge fails, multiple compression can happen quickly over 1-2 months.