Oracle delivered record fourth-quarter and full-year results this week, driven by surging demand for cloud infrastructure and AI workloads. Yet despite one of the strongest quarters in its history, shares fell as much as 10% in Wednesday’s after-hours trading, according to MarketWatch. The reason was not the results themselves but the cost of sustaining them.
Enterprise AI Demand Is Still Outpacing Supply
Oracle Cloud Infrastructure revenue nearly doubled year-on-year in the quarter ended May 2026, reflecting a market where supply continues to lag well behind what enterprises need. In the company’s Q4 2026 earnings call, Co-Chief Executive Officer, Clayton Magouyrk, observed: “there is still a massively higher demand than there is supply.” GPU utilisation across Oracle’s global data centres ran at 97.5% during the quarter, and most capacity that came up for renewal was absorbed by other customers in the same period.
Oracle’s contracted backlog suggests customers are also making longer-term commitments. It reached $638 billion at the end of the quarter, a figure nearly ten times its full-year revenue, after adding approximately $85 billion of new contracts in a single quarter. CFO Hilary Maxson explained that the backlog is expected to convert into revenue faster over coming quarters, suggesting organisations are not simply signing agreements but actively preparing to deploy AI at scale.
Customers Are Helping Fund the Build-Out
One of the more telling signals from Oracle’s results is how far some customers are willing to go to secure AI capacity. Oracle signed $67 billion in AI infrastructure contracts during the quarter, many structured as prepaid arrangements or bring-your-own-hardware deals in which the customer supplies the computing hardware directly. The total value of such arrangements now stands at $75 billion.
Magouyrk said the dynamic speaks for itself: “Customers are showing they chose OCI to deliver their infrastructure even when they are bringing the capital themselves.” Maxson also confirmed that margins on these structures are at or better than Oracle’s standard contracts.
Why the Strong Results Still Spooked Investors
The market’s negative reaction was clearly not about demand given these numbers. The concern was tied to the unexpected costs of meeting this demand. Oracle revealed plans for roughly $70 billion in capital expenditure next fiscal year, with around $40 billion in additional financing to fund its data centre build-out. This is on top of the company’s $162 billion debt, which saw $50 billion added in the past year alone, as reported by The Motley Fool.
Will the Demand Last?
Co-CEO Michael Sicilia shared his view of where enterprise AI stands: “Our customers have moved past the experiment stage with AI. They are ready to implement enterprise-ready, complete agentic solutions to help run their businesses.” For CX leaders, that shift is already visible in the capabilities now being deployed at scale. Agentic AI is emerging as the dominant model in enterprise CX, spanning customer service automation, personalisation, sales copilots, and journey orchestration.
Oracle’s results are a strong signal that this demand for AI infrastructure is real and growing. The backlog, the utilisation figures, and the willingness of customers to co-fund infrastructure all point in the same direction. Oracle is clearly betting the wave continues, and is spending accordingly.
Whether that bet pays off is less certain. The central question for investors is how long the current wave of AI investment holds, and whether Oracle’s aggressive build-out positions it well or leaves it exposed if enterprise demand plateaus before the returns arrive. For businesses weighing their own AI investments, they must face a different version of the same question: how to ensure spending translates into measurable returns rather than sunk cost. The fates of CX AI vendors like Oracle are interwoven with the outcomes of these enterprise customers.