Amazon Stock Could Soar Over the Next Few Years -- and It's Coming From an Unlikely Source

Amazon’s AWS unit generated 21% of Q1 revenue but 59% of operating profit, underscoring that the investment case is driven more by cloud profitability than retail. The company is committing $200 billion in capex this year to expand data centers and is seeing triple-digit growth in custom AI chips, which could support faster-than-28% overall growth if adoption continues. The article is bullish on AWS and AI-related upside, but it is primarily an analyst opinion piece rather than new company guidance.

Analysis

The market is still underappreciating that the real margin lever in Amazon is not retail monetization but compute supply control. If AWS keeps filling capacity faster than peers can build it, Amazon effectively turns capex into a flywheel: more infrastructure lowers unit costs, which improves price competitiveness, which then pulls in more committed workloads and makes the next round of capex easier to underwrite. The second-order winner is the broader AI infrastructure stack around power, networking, and data-center equipment; the loser is any hyperscaler or chip vendor that depends on keeping model training tied to generic GPU demand.

The custom-chip angle matters less as a product story than as a bargaining chip against Nvidia’s pricing power. If enterprise buyers increasingly view Trainium-style economics as a credible alternative for inference and some training workloads, the market should expect a gradual mix shift rather than an abrupt displacement; that means Nvidia revenue risk is more of a 12-24 month compression story than a near-term collapse. Intel is only relevant here as a proof point that Amazon can win workload share when it is willing to subsidize adoption and optimize around total cost of ownership.

The biggest contrarian issue is that consensus may be overestimating how quickly AI demand converts into durable AWS operating leverage. Heavy capex can compress free cash flow for several quarters even if revenue growth stays strong, so the stock can stall if investors start focusing on cash yield instead of top-line momentum. In other words, this is a multi-quarter execution trade, not a quick re-rate; the catalyst path is sustained AWS backlog conversion and evidence that custom silicon is taking attach share without cannibalizing overall cloud growth.

Key Decisions for Investors

• Long AMZN on 6-12 month horizon; favor entry on any post-capex or post-earnings drawdown, with thesis that the market will re-rate AWS quality once capacity converts into operating profit rather than just revenue growth.
• Pair trade: long AMZN / short NVDA for a 3-6 month horizon if AI capex remains elevated; thesis is that incremental AI spend increasingly benefits the platform owner with pricing control, while GPU economics face gradual mix pressure.
• Use a bullish call spread on AMZN 9-12 months out rather than outright stock if you want defined risk; upside case is multiple expansion from AWS margin inflection, while downside is prolonged free-cash-flow drag from capex.
• Avoid chasing INTC on this headline alone; the chip-design reference is strategic validation for Amazon, not a direct Intel operating improvement, so any sympathy bid in INTC should be treated as fadeable.
• If you want a higher-beta expression on the same theme, pair AMZN long against a basket of AI infrastructure beneficiaries that are most exposed to deployment slippage, but keep position size modest because the trade can be crowded and factor-sensitive.