Prediction: Owning 100 Shares of Amazon Stock Will Turn $25,000 Into $50,000 by 2030

Amazon is projected to compound earnings at just over 20% annually, which could support a doubling of its $251.35 share price to roughly $500 by 2030 if its 20x forward P/E multiple holds. Q2 2026 AWS revenue rose 37% year over year to $42 billion, while AWS generated 60% of Amazon's operating income and held a $496 billion backlog. Higher-margin non-retail revenue grew 24% to $124 billion, and Amazon's Trainium and Graviton chips are generating more than $25 billion of annualized revenue, supporting potential margin expansion despite elevated data-center and transportation spending.

Analysis

The investable issue is not AWS demand but incremental return on AI infrastructure. If internally designed silicon shifts a meaningful portion of inference from merchant GPUs, Amazon captures both the cloud-service margin and the hardware-cost saving; the resulting margin expansion could justify a premium multiple despite elevated capex. This is a more durable earnings lever than retail efficiency, but the stated EPS growth rate should be normalized for investment-related gains before underwriting any valuation target.

Near term (days to 1-3 months), AMZN is vulnerable to a "capex versus FCF" debate: backlog is not revenue until capacity, power, and customer deployment convert. The key catalyst is evidence that AWS growth sustains while capex intensity stabilizes, alongside disclosure of Trainium/Graviton adoption and gross-margin benefit. A miss on AWS growth, a rising depreciation burden, or weaker retail contribution would expose the stock to multiple compression because the upside case already relies on retaining its current earnings multiple.

Second-order, greater custom-chip penetration is modestly negative for NVDA's AWS-specific inference opportunity, but not necessarily NVDA's aggregate demand while training remains GPU-heavy and industry compute supply is constrained. The more direct relative beneficiary is AMZN versus cloud peers dependent on external accelerators; over 6-18 months, proprietary silicon can support lower customer pricing without sacrificing margin, raising switching costs and pressuring Microsoft and Google cloud economics at the margin.

Consensus likely overweights the headline revenue opportunity and underweights execution friction: power availability, depreciation, and customer migration can delay the margin payoff by several quarters. Conversely, a visible inference-cost advantage could cause estimates to lag, since modest incremental AWS margin has outsized consolidated earnings impact.

Key Decisions for Investors

• Accumulate AMZN only on post-results confirmation that AWS growth remains above 30% and consolidated operating-margin guidance is stable; use a 6-12 month horizon. Target 15-20% upside from earnings-estimate revisions, with thesis invalidated by two consecutive quarters of decelerating AWS growth or materially higher capex without FCF conversion.
• Prefer a relative-value long AMZN / short a basket of MSFT and GOOGL in equal cloud-exposure weights over 6-18 months, sized modestly. The trade expresses proprietary-infrastructure margin advantage while reducing broad AI-demand beta; exit if competitors demonstrate equivalent custom-silicon cost economics or AMZN's AWS margin fails to expand.
• Do not short NVDA solely on Trainium adoption. Instead, monitor AWS disclosures for custom-silicon share of inference workloads and NVDA data-center guidance; only consider an NVDA hedge if merchant-GPU demand shows sequential weakness rather than merely mix shift.
• Set an alert around next earnings for capex, depreciation, and AWS operating-margin disclosure. If capex accelerates faster than cloud revenue for two quarters, defer new AMZN exposure: the near-term valuation risk is FCF dilution, not demand failure.