I'd Buy Amazon Stock While It Sits 12% Below Its Record

Amazon shares are down about 12% from their $284.02 Aug. 3 record to roughly $249, despite AWS Q2 revenue accelerating 37% year over year to $42.2 billion and operating income rising 64% to $16.6 billion. AWS backlog reached about $496 billion, including Anthropic's more than $100 billion, 10-year commitment, supporting the AI infrastructure investment case. The key offset is AI-driven property-and-equipment spending of roughly $169 billion over the trailing 12 months, which pushed free cash flow to a $7.6 billion outflow from an $18.2 billion inflow; an FTC and state lawsuit over ad-auction practices also creates legal risk.

Analysis

The key debate is no longer whether AWS demand exists, but whether incremental AI capacity earns above its cost of capital before depreciation and power costs accelerate. A large contracted-revenue figure should not be valued dollar-for-dollar: duration, termination rights, utilization ramp, and customer concentration matter, particularly where major AI customers are themselves capital-dependent. If cloud growth remains elevated through the next two reports while capex growth decelerates, AMZN can re-rate as the market shifts from a free-cash-flow trough narrative to an infrastructure monetization narrative.

The advertising case is more material to multiple than near-term earnings. Even without a large cash penalty, mandated auction-design changes could reduce ad take rate, impair pricing opacity, and create a precedent for marketplace scrutiny; that would favor retailers with less vertically integrated ad ecosystems, including WMT and TGT, at the margin. The market is likely underpricing the possibility that legal discovery reveals conduct that increases regulatory remedies, but overpricing an immediate P&L hit given the likely multi-year process.

Near term, AMZN remains vulnerable to any evidence that capacity is arriving ahead of revenue recognition: weaker AWS growth, lower cloud margins, or a further free-cash-flow deterioration would challenge the valuation support implied by forward earnings. Conversely, a sustained AWS growth rate materially above the prior cycle's low-20% range, coupled with stable operating-margin guidance, is the 1-3 month catalyst for the recent drawdown to close. Structurally, AMZN is a cleaner AI-demand monetization vehicle than NVDA if enterprise workloads broaden, but a worse one if AI spend compresses because its returns are delayed by the capex cycle.

Key Decisions for Investors

• Accumulate AMZN in tranches over the next 4-8 weeks rather than chase a single entry; size for a 6-12 month thesis that AWS monetization outruns capex. Add only if the next earnings release shows AWS growth above 30% and no material cloud-margin deterioration; exit/reduce if growth falls toward 20% while capital spending remains elevated.
• For defined risk into the next earnings report, use an AMZN 3-6 month call spread rather than outright calls: buy near-the-money calls and sell strikes roughly 10-15% higher. The structure targets a post-results multiple recovery while limiting exposure if free-cash-flow concerns dominate.
• Maintain a relative-value long AMZN / short NVDA basket only after confirmation that AWS demand is broadening beyond a few AI labs. The thesis is that cloud revenue and operating leverage can catch up to infrastructure spend; close the spread if hyperscaler capex guidance is cut or NVDA demand remains supply constrained.
• Set a regulatory watch item, not a short, around court rulings on discovery and remedy scope. Escalate downside hedging if the case seeks auction redesign, advertiser restitution, or restrictions that directly alter ad pricing mechanics rather than merely financial penalties.