Why Is Everyone Talking About Amazon Stock? | AMZN Stock Deep Dive Part 2

Parkev Tatevosian, CFA•Aug 4, 2026

Summary

Parkev highlights Amazon's strategic shift and massive acceleration within its AWS and artificial intelligence segments. The thesis centers on Amazon's ability to develop its own hardware, specifically the Trainium and Graviton chips, which offer cost-effective alternatives to high-end Nvidia GPUs. Parkev notes that 98% of Amazon's top 1,000 customers already use Graviton, and revenue commitments for these chips have tripled in a single quarter. This hardware integration allows Amazon to capture a wide market of companies looking for performance without the cutting-edge price tag of competitors.

Parkev also analyzes the significant increase in capital expenditure, which Amazon recently raised to $220 billion for 2026. This increase is attributed to rising memory costs from suppliers like Micron and SK Hynix. However, Parkev emphasizes that the long-term economics of these data centers are favorable. The 'shell' of a data center lasts over 30 years, while the internal servers break even in under three years and have a useful life of five to six years, leading to several years of high-margin free cash flow. Parkev explains that the demand for these services is 'striking' and expected to outstrip supply through 2028.

Amazon (AMZN): Parkev argues that Amazon is a strong investment due to the 'striking' demand for its AI services that will likely outstrip supply through at least 2027. Parkev states that the company's proprietary Trainium and Graviton chips provide a significant cost advantage over competitors, leading to a tripling of revenue commitments in a single quarter. Parkev explains that the break-even point for AI hardware is under three years, providing several years of high-margin free cash flow thereafter.

Nvidia (NVDA): Parkev acknowledges Nvidia as the industry leader for cutting-edge, high-performance GPUs. Parkev states that while Nvidia remains the top choice for the most performant chips, Amazon is successfully attracting customers who prioritize cost-efficiency with its own proprietary chips. Parkev views Nvidia as the benchmark that Amazon is complementing with its own cost-efficient hardware alternatives.

Micron (MU): Parkev highlights Micron as a key beneficiary of the AI boom, noting that its memory products have seen price increases of 200-300% year-over-year. Parkev explains that these rising costs are the primary reason Amazon increased its capital expenditure forecast by $20 billion. Parkev views the massive growth in average selling prices for Micron as evidence of the immense value being captured by memory suppliers in the current AI cycle.

Mentioned Stocks

Sentiment: BUY

Reasoning: Parkev notes that Micron is seeing massive growth, with 200-300% year-over-year increases in selling prices for certain memory products. Parkev explains that these rising prices are significant enough to impact major players like Amazon, who had to increase their capital expenditure estimates by $20 billion due to memory costs.

Loading chart...

AMZN

Sentiment: BUYAction: RECOMMENDED

Reasoning: Parkev argues that Amazon's AWS segment is seeing incredible adoption, with proprietary chips like Graviton being used by 98% of top customers. Parkev states that the 'striking' demand for AI will likely outstrip supply through 2028, and the favorable economics of their data centers (3-year break-even vs. 6-year life) will drive significant free cash flow.

Loading chart...

NVDA

Sentiment: HOLD

Reasoning: Parkev states that Nvidia remains the leader for the most cutting-edge and performant chips. However, Parkev explains that Amazon is gaining traction by offering proprietary chips that serve as cost-saving alternatives for companies that do not require Nvidia's peak performance levels.

Loading chart...