Amazon: Undervalued AWS Growth

Amazon delivered a standout Q1, beating earnings and revenue estimates as AWS revenue rose 28% year over year to $37.6B. AWS operating margins improved to 23%, while new hyperscaler agreements and a $20B AI chip revenue run-rate underscore accelerating cloud and AI demand. The report is likely to support AMZN shares and reinforce the company’s leadership in enterprise cloud and data center infrastructure.

Analysis

AMZN’s surprise strength is less about a one-quarter beat and more about a re-rating of its two highest-multiple profit pools: cloud and custom silicon. The second-order read-through is that enterprise AI spend is not normalizing downward; it is broadening from experimentation into multi-year infrastructure commitments, which should keep AWS growth and mix tailwinds intact even if retail margins remain noisy. That also pressures the rest of the cloud stack: hyperscaler peers and legacy infrastructure vendors will likely face a tougher sell cycle because Amazon is proving it can monetize AI demand without giving up economics.

The custom chip ramp matters because it changes the bargaining power inside the AI supply chain. If AMZN can internalize more inference/training workload economics, it reduces dependence on third-party accelerators and should improve capex efficiency over time, while subtly compressing addressable demand for external AI silicon at the margin. The likely beneficiaries are datacenter buildout enablers and power/cooling vendors, but the relative losers are GPU- and networking-heavy vendors whose pricing power depends on sustained scarcity.

The main risk is that investors extrapolate this into a straight-line cloud acceleration story and ignore that capex intensity can compress free cash flow before it translates into durable EPS upside. Over the next 1-3 months, the key test is whether management guidance forces the market to treat the AI build as accretive rather than merely strategic; over 6-12 months, the question is whether competitive response from other hyperscalers closes the gap. A reversal would likely come from enterprise optimization cycles, a deceleration in cloud spending, or any sign that the AI chip economics are less scalable outside Amazon’s own workload base.

Consensus is probably still underestimating how much operating leverage exists if AWS growth remains in the high-20s while the AI chip business scales. What the market may be missing is that the winner is not just AMZN equity beta, but also the adjacent ecosystem tied to power delivery, liquid cooling, and datacenter capacity in regions where Amazon is expanding fastest. That makes the current setup more durable than a simple earnings beat trade, but also more crowded if investors chase the headline before the next capex print.

Key Decisions for Investors

• Add AMZN on any post-earnings pullback over the next 1-2 weeks; target a 3-6 month hold, with upside driven by upward estimate revisions and AWS multiple expansion rather than retail margin improvement.
• Pair trade: long AMZN / short a basket of slower-growth cloud incumbents over 1-3 months to capture relative re-rating as enterprise AI budgets consolidate around the strongest platform.
• Express the AI infrastructure beneficiary angle with a selective long in datacenter power/cooling names for 6-12 months; risk/reward improves if Amazon’s capex guidance stays elevated without a demand deceleration.
• Hedge the ‘capex eats FCF’ risk by buying short-dated downside on AMZN into the next guidance event if the stock runs hard; this is a good expression if the market starts pricing perfection.
• Avoid chasing third-party AI accelerator exposure at current levels if the market is already assuming perpetual scarcity; Amazon’s internal chip progress raises the odds of margin pressure for that segment over 6-18 months.