AI Hyperscaler Capital Spending and Performance 2026
Yardeni Research reports that the 2026 AI‑related capital spending plans of the four major hyperscalers—Amazon, Alphabet, Microsoft and Oracle—total roughly $600 billion, up from $309 billion in fiscal 2025. Much of the capacity build‑out scheduled for this year is being deferred to 2027 and beyond because of delays linked to power supply, equipment availability and local permitting. Yardeni characterises the trend as a "delay, not retreat," noting that Sightline Climate has recorded only nine cancellations among the 777 large data‑center projects it tracks.
For fiscal 2026, the remaining performance obligations (RPOs) that customers have contracted but not yet received exceed $2.3 trillion across the four firms. Microsoft leads the backlog with $678 billion, followed by Oracle at $638 billion. Yardeni flags customer concentration risk, especially the ability of OpenAI and Anthropic to meet their contracted compute commitments. In July, S&P downgraded Oracle’s credit rating to BBB‑, citing exposure to OpenAI as a key risk factor.
The financing mix for the build‑out has shifted. Historically, hyperscalers funded capital spending primarily from internal cash flow, but spending needs have outstripped free cash flow. Alphabet recorded its first‑ever quarter of negative free cash flow in Q2 2026, while Amazon’s trailing‑12‑month free cash flow swung to a $7.6 billion outflow. Consequently, external capital—mainly debt—covered 26% of hyperscaler capex in 2025, a share projected to rise to 33% in 2026 and roughly 35% by 2027, according to Goldman Sachs estimates.
Cloud revenue growth remains robust. Microsoft Cloud generated $214.3 billion in fiscal 2026, up 27% year‑over‑year. Amazon Web Services (AWS) reported $134.7 billion over the four quarters through June, a 50% increase. Google Cloud posted $77.7 billion, up 60%, and Oracle Cloud delivered $34.0 billion, up 39%.
Investors are increasingly differentiating among the hyperscalers rather than treating them as a monolithic group, evaluating each company’s AI spending plans, contracted backlog and actual cloud revenue delivery. Market reactions reflect these nuances: Oracle shares are down 28% year‑to‑date on concerns over negative free cash flow and OpenAI reliance; Microsoft shares fell 19% before rebounding on strong cloud results; Alphabet is up 20% year‑to‑date; and Amazon is up roughly 18% following its AWS earnings report.
"The AI trade no longer rests on blind faith, as in the initial days of AI euphoria, but on results," Yardeni concluded.