Amazon Has Underperformed in 2026. Is the Stock a Bargain Now?

Amazon’s AWS grew 28% YoY to $37.6B, its fastest pace in 15 quarters, and operating margin reached a record 13.1% as total revenue rose 17% to $181.5B and operating income jumped to $23.9B. However, AI infrastructure spending surged with capex of $44.2B in Q1 2026 versus $25B a year earlier, pushing trailing-12-month free cash flow down to about $1.2B from nearly $26B. The article frames the stock as good value (~29x earnings) but emphasizes that heavy AI spend is likely to keep pressuring free cash flow for a while.

Analysis

The important read-through is not that Amazon is spending aggressively; it is that the spend is starting to show up in operating leverage faster than the market expected. That makes AMZN a rare case where AI capex is still a net positive for the equity story, while the supply chain beneficiaries are becoming more polarized: NVDA continues to capture the first wave of demand, but Amazon’s custom silicon and internal stack increasingly threaten to keep a larger share of the economics in-house over time. The second-order implication is that cloud scale is becoming a durability test, not just a growth test; the winners will be platforms that can monetize AI without permanently importing the margin structure of a hardware vendor.

The near-term risk is that investors stay anchored on free cash flow and ignore that the payback on this buildout is lumpy. Over the next 1-3 months, the stock will likely trade on whether AWS growth remains above the mid-20s and whether management keeps capex elevated enough to suppress buybacks; that is what can cap multiple expansion even if headline earnings look fine. Over 6-18 months, the falsifier is simple: if AWS reaccelerates but incremental margins do not inflect, the market will conclude the AI spend is defensive rather than accretive, and the valuation should compress rather than rerate.

The consensus is probably too focused on the FCF collapse and not enough on the fact that Amazon can self-fund a longer investment cycle without balance-sheet stress. But the bull case is also a bit complacent about timing: a 29x multiple is not cheap if the market decides the cash conversion reset is structural instead of temporary. The best expression is to own AMZN into weakness, but size it as a catalyst trade, not a forever compounder, until the next couple of quarters confirm that capex is turning into durable AWS dollar acceleration rather than just a margin bridge.

Key Decisions for Investors

• Long AMZN on 1-3 month weakness; target a re-rating if AWS growth stays >25% and operating margin holds above low-teens. Falsify the thesis if quarterly capex remains above operating cash flow for another 1-2 quarters without evidence of margin expansion.
• For relative value, consider AMZN/NVDA pair trade: long AMZN, short a smaller NVDA hedge if you think hyperscalers will increasingly internalize AI economics via custom silicon over the next 6-12 months. Best entry is after the next earnings print if AMZN guides capex up again but confirms cloud momentum.
• If you prefer defined risk, use a 6-12 month AMZN call spread rather than outright stock; the upside is a multiple re-rating from depressed sentiment, while the main risk is another quarter of flat FCF causing the stock to stay range-bound.
• Set a watch item on AWS growth falling back below 25% or capex intensity declining before revenue reacceleration is proven durable; that would be the signal to fade the move rather than buy it.