Andy Jassy Just Delivered Incredible News for Amazon Stock Investors

Amazon shares hit a record high after results, with AWS revenue up 37% YoY to $42.2B in Q2, driven by AI workload momentum and customer demand. AWS ended Q2 with a $496B order backlog, pointing to continued acceleration, and CEO Andy Jassy signaled AWS could reach $1T in annual revenue (vs. prior “hundreds of billions”). The near-term watch item is earnings optics: in H1 2026, about $69B of Amazon’s $120.7B pretax profit came from paper gains on Anthropic, and analysts expect earnings to decline in 2027, despite a still-undemanding valuation (P/E ~22 trailing; 27.7 forward).

Analysis

The key mechanism is not the headline growth rate; it’s the conversion rate between backlog and durable free cash flow. If AWS is already running at capacity constraints, near-term capex can lift revenue first and earnings later, but the depreciation burden arrives with a lag, so the stock can look cheap right before forward EPS estimates roll over. That makes this a 1-3 quarter story on sentiment and a 6-18 month story on whether utilization stays high enough to justify the spend.

Second-order, Amazon’s custom silicon is the more interesting competitive signal than the revenue print. If Trainium and Kiro continue to take share inside AWS workloads, the long-term loser is not just NVDA but any merchant-silicon vendor exposed to hyperscaler self-supply; the near-term winner is still the AI infrastructure stack because the buildout itself is huge. The more Amazon proves it can lower unit compute cost, the more it can pressure Azure and Google Cloud on price, but that also risks a race-to-the-bottom in cloud margins.

The contrarian miss is that a lot of the reported earnings power is non-operating and likely not repeatable at the same pace, so the market may be pricing a cleaner earnings trajectory than actually exists. If 2027 EPS starts to come down while capex remains elevated, the multiple expansion case breaks even if revenue keeps growing. The setup is most fragile if backlog conversion slows, AI capacity comes online slower than planned, or AWS margin expansion stalls despite the spend.

Key Decisions for Investors

• Long AMZN on a 3-6 month pullback; favor a risk-defined call spread over outright stock if the name keeps rerating ahead of the next cloud print. Falsify if AWS operating margin or FCF guidance disappoints while capex stays elevated.
• Reduce or hedge NVDA exposure versus AMZN over 6-12 months. AWS custom silicon is a real share-risk over time even if near-term hyperscaler capex still supports semiconductor demand.
• Treat AMZN as a backlog-conversion watch item, not a linear growth story: add only if quarterly evidence shows revenue acceleration translating into improving incremental margins. If AWS growth holds above 30% but margins compress, exit the thesis.
• No action in NFLX or NDAQ from this note; the signal is too indirect. Save risk budget for AMZN execution and the broader AI infrastructure chain.