I've Been Wrong About Amazon Stock for Almost 10 Years. Here's Why I'm Finally Changing My Mind.

Amazon’s profit mix is shifting toward higher-quality engines: AWS generated roughly mid-double-digit billions in 2025 operating income, far exceeding retail, with software-like margins. Advertising is also scaling, with annual ad revenue disclosed at over $60B, up ~20% year over year and outpacing the core business. The article argues Amazon’s logistics network has evolved from a cost center into an operating moat, supporting a more durable long-term business outlook even as risks (capex/regulatory pressure) remain.

Analysis

AMZN’s investment case is increasingly a cash-flow durability story, not a retail margin story. That matters because the market should be willing to underwrite a higher multiple when two less-cyclical engines — cloud and retail media — finance the lower-margin distribution layer; the second-order winner is AMZN’s own pricing power, while the losers are merchants and brands that become more dependent on Amazon’s closed-loop funnel and pay-up ad auctions to defend share.

The bigger spillover is competitive, not accounting: ad dollars that migrate to commerce-intent surfaces are incremental for AMZN but largely displacement for GOOGL and, to a lesser extent, META. On the infrastructure side, persistent AWS capex is indirectly supportive for NVDA, but only if spend keeps converting into attachable compute demand rather than idle capacity. If that conversion weakens, the market will quickly stop paying for “quality mix shift” and revert to scrutinizing free-cash-flow yield.

Catalysts are mostly in the next 1-2 earnings prints: AWS growth, retail-media growth, and commentary on capex intensity versus operating leverage. The contrarian miss is that the bullish re-rating may already be partly in the price; this is a better business than the old Amazon, but not automatically a better stock if margins plateau and capex stays elevated. Falsify the thesis if AWS growth decelerates into the low-teens, ad growth normalizes materially, or management signals that incremental capex is no longer translating into margin expansion.

Key Decisions for Investors

• Buy AMZN on post-earnings or broad-tech pullbacks as a 6-12 month core long; prefer entries when the stock de-risks to the lower end of its 3-month range. Risk/reward is attractive only if the market is not already pricing sustained AWS and ad reacceleration.
• Use a 3-6 month AMZN call spread instead of stock if you want upside exposure with defined risk; thesis only works if the next earnings cycle confirms operating leverage, not just narrative improvement.
• Pair trade: long AMZN / short GOOGL for 1-3 months to express commerce-ad share gain versus general search ad exposure. This is a relative-value trade, not a macro bet; cover if GOOGL’s AI/search monetization holds up better than expected.
• Add NVDA on weakness only as a secondary beneficiary of sustained AWS capex; this is a lower-conviction proxy and should be sized smaller than AMZN because the linkage is indirect and capex can decelerate quickly.
• Set a watch trigger on the next earnings report: if AWS growth or retail-media growth slows materially while capex stays elevated, reduce exposure immediately; that would invalidate the ‘quality mix shift’ rerating thesis.