Oracle has disclosed extensive risks tied to its massive bet on AI infrastructure, as capital expenditures jumped to $55.7 billion（約8.9兆円） in fiscal 2026 and are planned to reach $90 billion（約14兆円） to $95 billion（約15兆円） in fiscal 2027. The company warned investors of potential pitfalls including construction delays, GPU shortages, customer defaults, and regulatory constraints. Oracle's detailed warning—filed as part of its annual report—reflects the company's high-stakes role in the AI buildout and suggests that even large tech players face significant uncertainty in delivering the capacity the industry demands.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Oracle disclosed in its annual report that it faces numerous risks in its AI data center buildout, including construction delays, GPU shortages, customer defaults, and potential regulatory constraints. The company's capital expenditures rose to $55.7 billion（約8.9兆円） in fiscal 2026 (ended May), up from $21.2 billion（約3.4兆円） the previous year, with plans for $90 billion（約14兆円） to $95 billion（約15兆円） in fiscal 2027.
Why it matters
Oracle is betting heavily to become a foundational player in the AI boom, with major deals with OpenAI and Meta and participation in Stargate, a project that could invest up to $500 billion（約80兆円） in data centers over the coming years. However, the company's exhaustive risk disclosures—citing overbuilding, credit risk, power shortages, permitting delays, and regulatory uncertainty—highlight that the massive infrastructure buildout required to support AI is far from guaranteed to pay off. The filing suggests real exposure for investors and the broader industry.
What to watch
Oracle shares fell 40% over the past month, and other AI stocks like Nvidia have also tumbled, signaling that investors are growing more cautious about the cost of the AI infrastructure buildout. The company's risks center on whether it can build capacity fast enough while managing costs, supply chain disruptions, and the credit risk posed by customers like OpenAI and Anthropic that are still spending far more than they earn.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
