Amazon’s AI spending boom is eating into free cash flow

Amazon beat first-quarter revenue and operating income estimates, while AWS revenue rose 28% to $37.6 billion and posted its fastest growth in 15 quarters. However, free cash flow collapsed 95% to $1.2 billion over the last 12 months as capital spending jumped to $43.2 billion, mainly for AWS and generative AI infrastructure. The report is positive on growth but tempered by heavy AI-related spending and concerns that cash generation is being squeezed.

Analysis

The market is still underpricing how capital intensity can become the dominant earnings variable for hyperscalers. For AMZN, the near-term issue is not demand quality but cash conversion: when capex growth outruns revenue growth, equity holders are effectively funding a balance-sheet transfer into future GPU and networking capacity, which compresses near-dated valuation support even if headline growth stays strong. That creates a setup where the stock can keep working on sales and AWS momentum, yet multiple expansion becomes harder unless investors gain confidence that incremental spend is monetizing within 2-4 quarters rather than drifting into a longer payback cycle.

Second-order winners are the infrastructure vendors and semiconductor ecosystem, but the duration of outperformance depends on whether hyperscaler spending broadens or concentrates. If AWS is locking in large future demand now, the adjacent beneficiaries are power, networking, and specialized AI silicon suppliers; however, if the market starts questioning returns on AI capex, the second-order loser becomes the entire AI supply chain through slower order cadence and more selective procurement. That risk is most acute over the next 1-3 earnings cycles, not over years.

The contrarian read is that the cash burn may be a feature, not a bug, if it is funding scarce capacity against signed demand. In that case, the right lens is not free cash flow today but the conversion of backlog into operating leverage once deployment saturates, which could drive an inflection in 2026. The market seems split between treating this as a temporary investment trough versus a structural drag; that uncertainty is exactly what keeps implied volatility elevated even after a fundamentally solid print.

Key Decisions for Investors

• Buy AMZN on pullbacks with a 3-6 month horizon; use defined risk via call spreads rather than outright stock. Base case is that the market re-rates the capex shock as temporary if AWS growth and backlog conversion stay intact; stop out if next-quarter commentary shows no moderation in capex intensity.
• Pair trade: long AMZN / short a basket of lower-quality AI infrastructure names that trade on backlog optics alone. The trade works if investors rotate toward cash-generative hyperscalers with real end-demand and away from story stocks whose orders are more exposed to budget scrutiny.
• Overweight power and data-center enabling names for a 6-12 month window, but size conservatively. The second-order beneficiary is the electrical bottleneck, not just compute demand; if hyperscaler builds persist, grid and cooling vendors should see the most durable order tailwind.
• If already long AMZN, hedge event risk with short-dated downside puts into the next earnings cycle. The risk/reward is asymmetric because any disappointment in AWS monetization or capex guidance can compress the multiple quickly even if top-line growth remains healthy.