Here's How Much You'd Have Today if You Invested $5,000 Each in Amazon and Costco 3 Years Ago. The Gains May Surprise You.

Article argues Amazon should outperform Costco over the next three years, highlighting AWS as a primary driver with 28% market share (vs. Microsoft/Azure at 21%) and $128.7B revenue in 2025. It cites Goldman Sachs: global cloud computing sales could reach $2T by 2030, with 10%-15% tied to generative AI, plus Amazon’s custom AI accelerator chip development. Despite noting Amazon’s valuation is lower (forward P/E 29.2 vs. Costco 41.1), it frames AI infrastructure capex as the key swing factor for future margins and long-term returns.

Analysis

The market is likely still underpricing how much of AMZN’s current equity story depends on a credible path from AI capex to operating leverage, not just top-line growth. In the next 1-3 months, any evidence that AWS spend is converting into higher utilization, better pricing, or faster attach of higher-margin services should matter more than raw cloud share headlines; if that conversion does not show up, the stock can de-rate quickly because the market is already paying for future scale.

COST is the cleaner balance-sheet/consumer-quality compounder, but that quality is already embedded in a premium multiple. The second-order issue is that if rates stay elevated or consumer trade-down intensifies, the market may rotate away from paying up for defensive growth and toward names with self-funded optionality; that creates a relative tailwind for AMZN versus COST over 6-18 months.

The bigger ecosystem winner from AI infrastructure spend is still the supplier stack, especially NVDA and, to a lesser extent, the broader semiconductor/capex cohort. The risk is that hyperscaler capex becomes a crowded narrative: if one or two quarters of spend fail to translate into margin or revenue acceleration, the whole group could see multiple compression even if absolute revenue keeps growing. The consensus is probably missing that the burden of proof is now on monetization, not adoption.

Key Decisions for Investors

• Long AMZN / short COST as a 3-6 month relative-value trade: AMZN has more credible upside to re-rating if AWS monetization improves, while COST’s premium valuation offers less room for error. Use any post-earnings weakness in AMZN to enter; thesis fails if AWS growth or segment margin decelerates for 2 consecutive quarters.
• Watch NVDA as a second-order beneficiary of AWS capex intensity; if hyperscaler capex guidance re-accelerates, NVDA remains the cleanest lever. Prefer a small long on pullbacks rather than chasing strength; invalidate if cloud capex commentary turns to efficiency-led slowdown.
• Avoid initiating a fresh long in COST at current multiples unless consumer traffic and membership renewal data remain exceptionally firm. It can still compound, but the risk/reward is poor if the market rotates toward lower-duration cash flow stories.
• If you want optionality on the AI monetization debate, buy AMZN call spreads around the next earnings window rather than stock. This caps downside if capex disappoints while keeping exposure to a positive AWS margin surprise.