Amazon.com, Inc. (AMZN)

Twin Engines Behind the Capex Cloud: AWS's Chip-Driven Re-Rating and a Resilient E-Commerce Core

Sector: Big Tech / Cloud Computing / E-Commerce / AI Infrastructure

Ticker: NASDAQ: AMZN

Rating: Buy

Report Date: July 6, 2026

1. Key Takeaways

• AMZN shares have underperformed the S&P 500 year-to-date, driven primarily by market anxiety over surging hyperscaler capex and its long-term return profile — not by any deterioration in the underlying business.
• AWS revenue growth continues to accelerate alongside expanding operating margins. Management has indicated that the majority of new 2026 capacity additions are backed by committed customer contracts, providing visibility into the return path on incremental capex.
• The in-house Trainium chip program is a standout: Trainium2 delivers roughly 30% better price-performance than comparable GPUs and is essentially sold out. Multi-year, multi-gigawatt commitments from Anthropic, OpenAI, and Uber have pushed Trainium's revenue backlog above $225 billion, with the business already running at a $20+ billion annualized revenue pace.
• Free cash flow is expected to turn negative in FY2026, but this reflects deliberate, demand-backed capacity investment rather than deteriorating operating quality. Consensus expects FCF to turn positive again in 2027, with FCF margin recovering to roughly 8% by FY2028.
• The core e-commerce franchise remains robust — U.S. online retail spend during this year's Prime Day rose 9.3% year-over-year — reinforcing the flywheel between retail and AWS and the company's broader ecosystem moat.
• Near-term risks remain concentrated in softening consumer confidence and broad market/systematic risk: AMZN's beta remains meaningfully above 1, implying elevated sensitivity to a market-wide risk-off move at a time when valuations across equities remain elevated.

2. Share Price Performance and Market Sentiment

Despite solid fundamentals, AMZN shares have lagged the broader market for most of 2026. The original bullish thesis — attractive relative valuation and strong expected AWS demand — remains intact, and AWS has indeed benefited from strong growth momentum. Yet the stock has essentially gone nowhere year-to-date, underperforming the S&P 500 on both an absolute and risk-adjusted basis. In June, shares broke back below a key prior breakout level near $245, leaving investors who had bought the April breakout underwater on a mark-to-market basis, with the one-year total return falling to roughly 7%.

This divergence between strong operating results and weak share performance stems largely from broad-based investor concern over the future returns on large-scale AI infrastructure build-outs — a concern that has weighed disproportionately on infrastructure-as-a-service (IaaS) providers, even as semiconductor stocks have been the primary beneficiaries of the same capex super-cycle and have re-rated sharply higher. In other words, the market appears to have already priced in upside for the semiconductor layer of the AI stack, while applying an arguably excessive discount to the cloud infrastructure layer's return path.

3. AWS: Accelerating Growth and Improving Profitability

Over the last two reporting periods, AWS revenue growth has continued to accelerate while operating margins have expanded in tandem — a sign that scale economics are starting to show through. On the Q1 2026 earnings call, management stated explicitly that the majority of AWS's 2026 capacity additions are already backed by customer commitments, indicating that the aggressive capex plan reflects visible, contracted demand rather than speculative overbuild.

At the same time, Amazon's backlog — reportedly now exceeding $460 billion — provides multi-quarter revenue visibility. The fact that peers such as Alphabet have reportedly needed to seek external financing to fund their AI infrastructure ambitions is, if anything, a further signal of how tight capacity remains industry-wide — a dynamic in which Amazon's scale and first-mover advantage leave it relatively well positioned in the current capacity race.

4. Capex and Free Cash Flow: Near-Term Pressure, Manageable Path

The market's central concern with Amazon centers on the sharp escalation in capex and the resulting compression in free cash flow margin. As Microsoft, Alphabet, Oracle, and Meta all ramp capital spending in parallel, the hyperscaler cohort's aggregate FCF margin is expected to decline meaningfully. For Amazon specifically, this dynamic could push FY2026 free cash flow into negative territory.

It is worth noting this would not be Amazon's first period of negative free cash flow: FY2021–FY2022 saw a similar dynamic tied to a prior investment cycle, though that decline was compounded by rising operating costs and inventory levels in Asia. By contrast, operating cash flow in the most recent quarter improved meaningfully year-over-year even during a seasonally soft Q1 — suggesting the current FCF pressure is driven primarily by a deliberate, front-loaded capex program rather than any underlying deterioration in operating quality.

5. Trainium: An Underappreciated Second Growth Curve

Progress on Amazon's in-house Trainium AI accelerator is arguably the most under-appreciated storyline in the recent print. Management disclosed multi-year, multi-gigawatt Trainium purchase commitments from two leading AI labs — Anthropic and OpenAI — alongside a growing roster of enterprise customers such as Uber, pushing Trainium's revenue backlog above $225 billion. Trainium2 delivers roughly 30% better price-performance than comparable GPUs and is essentially sold out.

On an operating basis, the chip business grew nearly 40% sequentially in Q1, with annualized revenue surpassing $20 billion and triple-digit year-over-year growth — a scale that management noted actually understates the business's true size. Management estimated that if the chip business were broken out as a standalone unit and sold to AWS and third parties the way leading merchant chip vendors do, its annualized revenue run-rate would be closer to $50 billion. That potential scale suggests the market may still be underestimating the strategic value of Amazon's silicon strategy, positioning the company as a credible alternative to Nvidia in a capacity-constrained AI accelerator market.

6. Token-as-a-Service: Converting Capacity Advantage into Margin Advantage

Ample compute supply not only cements Amazon's role as an infrastructure provider to leading AI labs — more importantly, it lays the groundwork for a higher-margin "Token-as-a-Service" layer above raw infrastructure. As Bedrock-related revenue continues to grow as a share of AWS's overall revenue, Amazon is gradually shifting from simply renting out compute toward monetizing model inference at the token level, a structurally higher-margin business.

Anthropic is a case in point: capacity-constrained and actively seeking to expand, it stands to benefit meaningfully from Amazon's infrastructure scale, while Amazon in turn captures upside from the token economy as its partner scales usage. As long as frontier labs keep shipping smarter models that meaningfully improve enterprise workflow productivity, underlying token demand is likely to prove more resilient than the market's current, more cautious assumptions imply.

7. E-Commerce: The Ballast Beneath the AI Narrative

Beyond the AI infrastructure story, Amazon's North American e-commerce business continues to grow steadily and remains a critical pillar of overall profitability. During this year's four-day Prime Day event, total U.S. online retail spend across all retailers exceeded $26 billion, up 9.3% year-over-year. On a gross merchandise value (GMV) basis, Amazon has now surpassed Walmart as the largest retailer in the United States, and the reinforcing relationship between retail and AWS — each a major internal customer of the other — continues to strengthen the company's long-term competitive moat.

8. Risk Factors

• Uncertain return on capex: If industry-wide AI infrastructure investment fails to deliver expected returns, valuation multiples for hyperscale cloud operators could face further pressure.
• Softening consumer confidence: A prolonged weakening in the macro consumer backdrop would directly weigh on the growth trajectory of the e-commerce segment.
• Broad market/systematic risk: With overall market valuations elevated and AMZN's beta meaningfully above 1, the stock's sensitivity to a broad market drawdown is elevated, raising the risk of outsized declines if risk appetite reverses.
• Competitive and pricing risk: With Google, Microsoft, and Oracle all ramping capex in parallel, the possibility of compute pricing competition emerging cannot be ruled out, which could compress industry-wide margins.

9. Valuation and Investment Recommendation

On valuation, AMZN now trades below 30x earnings, below the levels seen at both the start of 2026 and the April 2026 highs, and closer to the trough valuation reached during the April 2026 pullback. Given the company's multiple growth vectors — cloud capacity, proprietary silicon, and the emerging token economy — layered on top of a stable cash-generative e-commerce core, the current pullback looks more like a long-term accumulation opportunity than the start of a trend reversal.

On balance, market concerns over hyperscaler capex and declining free cash flow margins are likely to keep weighing on the stock's valuation multiple in the near term, but this appears to reflect an overreaction in sentiment rather than any genuine deterioration in fundamentals. We maintain a Buy rating on AMZN, and would recommend investors accumulate on weakness — while managing position sizing and remaining mindful of consumer-confidence and broad market risk — in order to capture the medium-to-long-term re-rating potential from the combined momentum of cloud infrastructure, AI silicon, and the e-commerce ecosystem.

Disclaimer: This report was compiled from publicly available market information and third-party analyst commentary for informational and discussion purposes only. It does not constitute investment advice or a solicitation to buy or sell any security. Investing involves risk, including possible loss of principal.