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

Amazon’s AWS reaccelerated with Q1 2026 revenue up 28% YoY to $37.6B (fastest in 15 quarters) and AWS operating margin at 37.7% ($14.2B operating income). However, the AI infrastructure build pushed capex to $44.2B in the quarter (from $25B a year earlier), driving trailing-12M free cash flow down to about $1.2B versus nearly $26B. Operating margin reached a record 13.1% (operating income $23.9B on $181.5B revenue), but investors remain focused on the continued FCF pressure from heavy spending.

Analysis

The market is treating AMZN like a cash-flow story, but the real mechanism is balance-sheet optionality: management is swapping near-term FCF for a larger, more durable share of AI infrastructure spend. That should keep earnings estimates supported while the multiple stays capped until investors believe the capex curve is peaking; in the meantime, the equity is more vulnerable to “show me” skepticism than to any real credit concern.

Second-order, the cleanest near-term beneficiary is NVDA, because Amazon’s spend still largely validates the GPU cycle even as AWS custom silicon grows. But that custom silicon roadmap is the quiet long-term threat to NVDA: every incremental internal workload migrated to Trainium/Inferentia is a dollar of third-party accelerator demand that never fully arrives. Over 6-18 months, the question is not whether AI spend stays high, but whether Amazon captures enough of the stack to earn a higher share of the economics than the market currently discounts.

The contrarian view is that the stock may be less cheap than it looks if free cash flow remains near zero for several quarters; however, the bearish thesis breaks if AWS growth stays above ~25% and capex moderates without re-accelerating. For the next 1-3 months, the stock likely trades on capex optics and analyst model cuts; for 6-12 months, the catalyst is FCF inflection. A miss on AWS growth or another step-up in capex is the main falsifier for any long thesis.

Key Decisions for Investors

• Buy AMZN on pullbacks over the next 2-4 weeks, sized as a medium-term core long; target a 12-18 month re-rating if AWS growth holds above 25% and capex normalizes, with downside capped if the multiple compresses on FCF optics alone.
• Use NVDA strength as a tactical hedge, not a conviction short: if Amazon commentary implies custom silicon is taking share faster than expected, fade NVDA on rallies over the next 1-3 months; otherwise avoid fighting the GPU cycle.
• Set a watch item on AMZN for the next two quarters: if trailing FCF stays near zero while capex remains above ~$35-40B/quarter, delay adding and wait for evidence of operating leverage before increasing exposure.
• No action in AAPL or NFLX from this note; the signal is idiosyncratic to AMZN and does not justify a broader megacap AI basket trade.