Amazon CEO Andy Jassy Just Made a Shocking Trillion-Dollar Prediction

Amazon shares are up ~18% after AWS posted 37% growth and helped drive AWS to ~60% of operating profit in Q2. Andy Jassy projects AWS could reach $1T in annual revenue and notes Amazon expects a computing-capacity shortage in 2026 (and into 2027), with early demand already appearing for 2028. Amazon also plans 2026 capex of $220B, indicating sustained AI-driven demand that underpins the bullish multi-year outlook for operating income and valuation.

Analysis

The market is likely underpricing the second derivative of AWS scarcity: when a hyperscaler cannot satisfy demand, the near-term stock response is not just faster growth, but pricing power and customer lock-in for the firms that secure capacity early. That helps AMZN, but the more immediate relative winner is NVDA, because every additional tranche of data-center spend ultimately needs accelerators, networking, and power infrastructure before it becomes revenue. The loser set is the rest of cloud: MSFT and GOOGL would face a tougher sell if customers believe reserved capacity is the binding constraint rather than software differentiation.

The main risk is that investors extrapolate revenue while ignoring the capital intensity required to get there. A multiyear capex wave can produce a period where reported growth stays strong but free cash flow and buyback capacity flatten, which typically caps multiple expansion after the first move higher. For AMZN, the key question over the next 1-3 quarters is whether incremental compute dollars translate into sustained operating leverage, or whether depreciation, power, and network costs dilute the margin profile as supply finally comes online.

My contrarian take is that the bullish narrative is directionally right but probably too linear. A stated 2028 demand queue is not the same as contracted revenue; if enterprise AI spend pauses, the backlog can evaporate faster than the market expects, especially if pricing on inference workloads falls. Falsifier: AWS growth must stay above the high-20s/low-30s and margin should not slip materially below the high-30s; if either happens, the market will likely de-rate the long-duration AWS comp and treat the capex program as less accretive than advertised.

Key Decisions for Investors

• Stay long AMZN, but add on pullbacks rather than chasing; use a 3-6 month horizon and require AWS growth to hold above ~30% for the thesis to stay intact.
• Long NVDA vs. short a cloud basket proxy (e.g., AMZN/QQQ beta-neutral overlay) over the next 1-3 months: capex monetization should flow first to semis and networking before it fully accrues to platform margins.
• If already long AMZN, consider a collar into the next earnings cycle: buy downside protection around the post-rally gap to guard against any capex/margin disappointment that would compress the multiple quickly.
• Watch MSFT and GOOGL for relative weakness on any sign that AWS capacity remains the bottleneck; if their cloud growth does not accelerate, the market may have to reprice share gains for AMZN upward over 6-12 months.
• Set a falsifier alert on AMZN if operating margin trends below the high-30s or if free cash flow inflects down despite continued revenue growth; that would argue the capex supercycle is becoming a valuation headwind.