Amazon to report Q3 earnings after the bell with AI spend under scrutiny

Ahead of Thursday’s Q3 report, analysts expect Amazon adjusted EPS of $1.96 and revenue of $202.31 billion, including AWS revenue of $45.48 billion and advertising revenue of $20.82 billion; AWS growth is forecast at 38%. Investors will weigh cloud demand and AI opportunities against rising investment: capex is forecast by StreetAccount to exceed $280 billion in 2027, while FactSet estimates a $9.9 billion current-quarter free-cash-flow outflow. Amazon has cut more than 57,000 staffers since 2022 and laid off fewer than 250 employees in its device software and services team on Wednesday; the article reports no Q3 results or share-price reaction.

Analysis

The key earnings question is whether AWS’s capacity scarcity converts into durable returns or simply pulls forward a larger, lower-near-term cash burden. Capacity commitments improve demand visibility, but are not equivalent to realized revenue or attractive incremental returns: watch utilization, pricing, AWS operating performance, and the timing of data-center and power costs. If capex rises again without evidence of conversion, depreciation and cash-flow pressure can outlast the current AI enthusiasm and weigh on the multiple.

The immediate bar is high: AWS growth is expected to accelerate again, so meeting the estimate may not be enough if guidance fails to extend that trajectory. The 1–3 month catalyst is management’s explanation of capex, capacity delivery, and cash-flow recovery; the 6–18 month test is whether AI workloads produce monetizable usage rather than stranded capacity. Peer FCF pressure at Alphabet and Meta suggests sector-wide investment intensity, but does not remove Amazon-specific execution risk; Microsoft’s positive, though weaker, cash flow may make it a relative-quality beneficiary if investors favor funding resilience.

Agentic commerce is a longer-dated threat to sponsored-product economics, not yet proof of near-term ad displacement. Amazon’s first-party shopping assistant and control of checkout give it leverage to retain monetization, but blocking agents could also push discovery elsewhere. Usage claims alone do not establish incremental purchases or ad yield. The contrarian risk is that investors treat sold-out capacity and assistant engagement as earnings visibility before either is demonstrated in returns.

Key Decisions for Investors

• Avoid a blind pre-print directional position: AWS acceleration and heavy investment are both visible, while the surprise likely lies in forward capex, AWS profitability, and the cash-flow path.
• After results, consider a conditional long AMZN only if AWS growth meets or exceeds the 38% expectation and management links spending to deliverable capacity and monetization; reduce or exit if AWS growth disappoints or cash-flow recovery is deferred without stronger demand evidence.
• If capex guidance increases without improved evidence on utilization or returns, favor a relative-quality pair of short AMZN / long MSFT, sized modestly. Falsify the setup if Amazon demonstrates stronger cash conversion or Microsoft’s cash-flow resilience deteriorates materially.
• Track ad revenue and evidence of purchases or ad monetization from Alexa for Shopping over the next several quarters. Rising assistant usage without measurable commerce economics is not a bullish signal; accelerating ad weakness alongside external-agent adoption would strengthen the disintermediation concern.