24/7 Wall St. identifies Amazon and Oracle as hyperscalers facing the highest AI capital expenditure risks relative to cash flow generation.
Oracle (NYSE:ORCL) and Amazon (NASDAQ:AMZN) both closed their latest quarters with the same uncomfortable admission: artificial intelligence infrastructure is now consuming cash faster than either business can generate it. Oracle reported fourth-quarter results for fiscal year 2026 on June 10, 2026, followed by Amazon’s second-quarter fiscal 2026 report on July 30, 2026. When measured against peers Microsoft (NASDAQ:MSFT) and Alphabet (NASDAQ:GOOGL), both companies appear the most exposed to current spending pressures.
Oracle’s cloud infrastructure segment grew 93% year over year, while remaining performance obligations (RPO) ballooned to $638 billion, a 363% increase. The trade-off is steep: fiscal 2026 capital expenditures reached $55.66 billion against operating cash flow of just $31.98 billion, resulting in negative free cash flow of $23.69 billion. Chief Financial Officer Hilary Maxson advised analysts to expect a “net cash outlay for capital expenditures of around $70 billion” in fiscal 2027, to be financed by approximately $40 billion in debt and equity issuances. Non-current debt has already climbed to $124.7 billion. Investors have reacted accordingly: ORCL shares are down 26.88% since the earnings release.
At Amazon, AWS revenue grew 36.7% year over year to $42.2 billion—the fastest pace in 18 quarters—supported by a $496 billion backlog. However, second-quarter capital expenditures alone hit $54.21 billion, pushing first-half 2026 free cash flow into negative territory at $27 billion. CEO Andy Jassy framed the situation plainly: “we’ll spend a lot of capex and encounter free cash flow headwinds until these data centers come online.” Large-scale Trainium commitments from OpenAI (approximately 2 gigawatts starting in 2027) and Anthropic (up to 5 gigawatts) justify the buildout, but they also concentrate payback risk around two primary counterparties.
Microsoft still generated $67.99 billion of fiscal 2026 free cash flow despite recording $115.95 billion in capital expenditures. Alphabet’s second-quarter free cash flow slipped to negative $5.86 billion—an uncomfortable result that remains cushioned by its core search advertising business.
I will monitor whether Oracle’s BBB-rated balance sheet can absorb another debt cycle if its concentrated AI backlog slips even a couple of quarters. For Amazon, the central question is margin sustainability: retail operations cannot indefinitely subsidize an annual capital expenditure run rate of approximately $220 billion. All of this infrastructure requires substantial power and cooling, which is why we have compiled seven key suppliers driving the AI infrastructure buildout in a complimentary research report.
From an exposure-with-safety-net perspective, Amazon’s profile appears sturdier. AWS continues to deliver 39% operating margins, and its retail division provides reliable funding for the expansion. Oracle presents a more compelling case for turnaround investors comfortable with a 40.26% one-year drawdown and the possibility that RPO conversion materializes on schedule. A single clean quarter of positive free cash flow at Oracle would serve as a meaningful signal to track. Both companies remain ownable, but Oracle carries the thinnest margin for error among the hyperscalers.
Contact [email protected] for any questions or corrections.