Amazon stock: Wolfe sees AWS growth underappreciated By Investing.com

Wolfe projects ~30% annual AWS revenue growth over the next three years (vs. ~25% expected by Street), driven by AI demand and partnerships, and estimates Anthropic could generate about $15.2B in AWS-related revenue in 2026. The firm expects Amazon to add ~6 GW of compute capacity per year in 2026-27 and to invest roughly $1 trillion in capex from 2024-2030 (which may pressure near-term free cash flow but should improve returns by ~2029); Wolfe's price target is $255 (~25x 2027 EPS) vs the stock trading at ~21x those earnings.

Analysis

AWS-driven AI demand creates a multi-stage market where hardware/service vendors capture value before the cloud operator does. Expect server OEMs, GPU suppliers and data-center infra names to enjoy near-term volume and pricing power as customers accelerate deployments; that supply-side revenue will likely re-rate faster than AWS’s cash-flow profile because order flows are visible and less encumbered by long monetization tails.

Large-scale capex programs impose a meaningful short-to-medium-term earnings cadence risk for the cloud owner: higher depreciation and cash burn can compress free cash flow and force more conservative capital allocation commentary even if long-term returns are attractive. Simultaneously, customer concentration on a few model providers and multi-cloud strategies create inventory and pricing cyclicality — hardware demand can spike and then retrench quickly, amplifying quarters of outperformance followed by pronounced downside.

The path to realizing durable economics is event-driven and multi-year: meaningful upside requires clear proof points (sustained per-customer spend on hosted model inference, stickier revenue-sharing contracts, and reduced commissioning lead times for new datacenters). Near-term catalysts to watch are supplier backlog indicators, AMZN gross margin progression excluding retail, and large third-party AI contract rollouts; any mismatch between capex cadence and revenue recognition will be the fastest way to reverse current optimism.

Contrarian tension: consensus underweights both the speed at which hardware winners capture incremental margin and the depth of near-term FCF drag at the cloud operator. That sets up asymmetric trades where concentrated hardware/execution winners can outperform materially into the next 6–18 months while the cloud owner’s multiple re-rates only after several quarters of realized monetization.

Key Decisions for Investors

• 1) Long SMCI equity (size 1–3% portfolio) for 6–12 months to capture server/edge box order flow; set a stop at 20% drawdown. Rationale: direct exposure to OEM demand with asymmetric upside if AI deployments sustain; reward skew ~2:1 vs downside given order visibility.
• 2) Pair trade — Long SMCI (0.75–1.5% portfolio) / Short AMZN (0.5–1% portfolio) over 3–9 months. Rationale: capture hardware margin cycle vs potential FCF/earnings compression at the cloud owner; target 3–5% net portfolio gain if hardware outperforms cloud by ~10–15%.
• 3) Options play on AMZN: buy a 9–12 month call spread to limit premium exposure while keeping upside if monetization accelerates. Risk/reward: limited premium loss (defined) vs 2–4x upside if cloud monetization catalysts hit; exit on two sequential quarters of AWS revenue upside surprises.
• 4) Tactical long on data-center REITs (EQIX/DLR-sized exposure via sector ETF) for 6–18 months to ride pricing power in colocations. Risk: demand retrenchment if inventory cycles; target total return 10–20% with dividend cushion and tighten stop if utilization falls two consecutive quarters.