Amazon Doesn't Pay a Dividend and Constantly Dilutes Shareholders. Here's Why I'd Still Buy and Hold It Forever.

The article argues Amazon’s Prime “moat” is worth roughly $1,430/year versus a $139 price, citing nearly $105B in annual fast-delivery savings and JPMorgan’s updated valuation of Prime membership. It highlights Amazon’s AI strategy with Anthropic committing $100B+ over 10 years to AWS, with Amazon investing $5B upfront and potentially $20B more (total up to $13B), plus ~1 GW of Trainium2/3 capacity coming online by year-end. It also frames the lack of a dividend as reinvestment-led growth (Prime network build-out expansion from 70 to 200 rural delivery stations and more bundled services), positioning Amazon for continued compounding rather than cash payouts.

Analysis

AMZN’s real edge is not retail share gain; it is the compounding of fixed-cost infrastructure into a higher-ARPU subscription and enterprise compute stack. That makes the earnings power less cyclical than the market typically prices, but it also means the incremental upside depends on management keeping reinvestment yields high enough to justify continued capex and equity dilution. The stock should outperform on any evidence that customer retention stays sticky while logistics density and AI compute monetization improve, because that combination supports both multiple expansion and a longer runway for free-cash-flow growth.

The second-order winner is AMZN’s own silicon and cloud ecosystem, not just the consumer business. If Trainium adoption scales, the market may need to haircut some of the most optimistic share assumptions for NVDA in hyperscale inference/training, but the timing matters: this is a 6-18 month share-shift story, not a next-quarter revenue event. The more immediate competitive pressure falls on retailers and delivery-native platforms—WMT, TGT, and DoorDash/Instacart-style convenience networks—because faster fulfillment raises the hurdle rate for any competitor trying to defend baskets without matching Amazon’s logistics spend.

Contrarianly, the consensus may be overestimating how much of this is already in the stock. Prime economics are increasingly mature, so the near-term catalyst is not membership count; it is whether management can show that AI and logistics capital intensity are still producing rising incremental returns. If AWS growth or operating margin guidance softens while capex stays elevated, the narrative shifts from “toll booth” to “expensive empire maintenance.” Watch for any sign that Trainium volume ramps slip or that hyperscaler AI workloads remain Nvidia-heavy; that would blunt the most underappreciated part of the thesis.

Key Decisions for Investors

• Long AMZN on any post-headline consolidation over the next 1-2 weeks; target 8-12% upside into the next earnings/guidance reset if margin resilience and AI capex monetization hold. Falsify if operating margin guidance is cut or AWS growth decelerates materially.
• Pair trade: long AMZN / short WMT or XRT for 1-3 months. Thesis is that fulfillment-speed differentiation protects AMZN’s share gains while the margin burden of matching it falls on broader retail peers. Exit if Walmart shows materially better online fulfillment economics or AMZN capex reacceleration compresses margins.
• Tactical hedge on AI supply-chain spillover: trim NVDA exposure versus AMZN if you believe custom silicon adoption is real; the clean expression is long AMZN, short a partial basket of NVDA over 6-12 months. Invalidated if Trainium deployment slips or Anthropic/Hyperscaler workloads stay Nvidia-centric.
• Avoid chasing AMZN with short-dated calls after a bullish narrative piece; the better entry is on any pullback tied to capex fear. Use 3-6 month horizon, because the operating leverage from logistics density and AI will show up in guidance before it shows up in reported revenue.
• Watch item, not a trade yet: if Prime pricing is raised again without churn, that would be the strongest proof the subscription moat is still under-monetized and could justify a higher multiple. If churn rises, the stock likely de-rates faster than fundamentals fall.