Oracle Q4 Shows AI Buildout Reaching Escape Velocity

As cloud infrastructure revenue nearly doubles and contracted backlog hits $638 billion, Oracle is rewriting its identity as the enterprise backbone of the AI economy — at a price investors are still weighing.

Key Takeaways

• Oracle’s cloud infrastructure revenue surged 93% in Q4 to $5.8 billion, with full-year IaaS growth of 77%, driven by AI workload demand that has materially reordered the company’s revenue mix.
• A record $638 billion in remaining performance obligations, up $85 billion sequentially, signals that enterprise AI contract momentum is structural, not cyclical, anchored by $75 billion in customer prepayments.
• Despite record operating cash flow of $32 billion, Oracle’s $55.7 billion capital expenditure program pushed free cash flow to negative $23.7 billion, raising legitimate questions about dilution from a planned $40 billion capital raise in fiscal 2027.

A Company Recast

Oracle has spent two decades being underestimated. For much of that time, the company occupied a peculiar position in the technology landscape: indispensable to its customers, yet perpetually overshadowed by the hyperscalers racing to redefine enterprise computing. Fiscal year 2026 may mark the moment that characterization becomes untenable.

The numbers Oracle reported for the twelve months ended May 31 are not merely strong. They represent a structural reordering of what the company is. Total revenue reached $67.4 billion, up 17 percent for the year, but the headline obscures the more important story unfolding beneath it. Oracle Cloud Infrastructure revenue grew 77 percent annually to $18.1 billion. In the fourth quarter alone, IaaS expanded 93 percent to $5.8 billion. Combined cloud revenue for the year hit $34 billion, growing 39 percent. These are not the metrics of a software incumbent managing a gradual transition. They are the metrics of a company accelerating into a generational infrastructure cycle.

The AI Architecture Underneath

What is driving that acceleration is not simply Oracle selling more of what it has always sold. The company’s Multicloud AI Database grew 404 percent in the fourth quarter, its fastest-growing product on record. Enterprises embedding vector search, generative capabilities, and autonomous operations into their existing technology estates are finding that Oracle’s database layer sits precisely at the intersection of what AI applications require: reliable, secure, high-performance data infrastructure that integrates natively across cloud environments.

That is a meaningful competitive position. The hyperscalers — Amazon, Microsoft, Google — dominate public cloud broadly, but Oracle’s deep integration into enterprise application stacks, its multicloud flexibility, and its sovereign cloud capabilities address workloads that hyperscalers are structurally less suited to handle. Healthcare systems, financial institutions, and government agencies managing sensitive data at scale are not easily migrated onto a single public cloud provider. Oracle’s architecture is built for precisely that complexity, and AI adoption is making that distinction more consequential, not less.

Oracle Health, the company’s reconstituted Cerner business, is advancing an AI-enabled platform that management expects to deliver double-digit growth in fiscal 2027. The clinical rationale is substantive: reducing administrative burden on clinicians, accelerating drug discovery through molecular design modeling, and compressing the regulatory timeline for clinical trials. These are not distant aspirations. They are areas where AI is already producing measurable productivity gains, and Oracle’s installed base across health systems gives it a direct and defensible channel to deploy them at scale.

The Backlog That Changes the Conversation

Perhaps the most revealing number in Oracle’s fiscal 2026 results is not a revenue or margin figure. It is the $638 billion in remaining performance obligations — contracted, committed revenue that has not yet been recognized. That figure increased $85 billion in a single quarter and rose 363 percent year over year. The sequential jump alone exceeds Oracle’s total annual revenue from just three years ago.

The composition of that backlog matters as much as its size. A substantial portion reflects large-scale AI infrastructure contracts from enterprises and governments committing to Oracle’s cloud at multi-year horizons. Embedded within those arrangements are $75 billion in customer prepayments and customer-supplied GPU deals that materially reduce the net capital Oracle must deploy to bring capacity online. It is a financing mechanism easy to overlook in the headline capital expenditure figures, but it fundamentally alters the risk profile of Oracle’s expansion.

The backlog trajectory also answers a question that regularly attaches to high-growth infrastructure stories: is the demand real? In Oracle’s case, the answer is demonstrably yes. These are signed contracts, not pipeline projections. Converting that backlog into recognized revenue over the coming years is simultaneously the central operational challenge and the central investment thesis.

Where Capital Intensity Meets Investor Discipline

Oracle generated $32 billion in operating cash flow for fiscal 2026, a record representing 54 percent growth year over year. That figure reflects the enduring cash-generative power of its installed base and the early monetization of its cloud buildout. But the company also spent $55.7 billion in capital expenditures over the trailing four quarters, producing free cash flow of negative $23.7 billion. To fund that expansion, it raised $43 billion in debt and $5 billion in equity during the year and has signaled plans to raise approximately $40 billion more in fiscal 2027, including a $20 billion at-the-market equity program.

The market’s response on June 10 was instructive. Oracle shares fell more than 10 percent despite clear beats on revenue and earnings. The reaction was not a rejection of the business model. It was a measured repricing of execution risk. Investors absorbing the scale of planned capital outlays are weighing whether GPU procurement can be executed at pace, whether power infrastructure can be secured in time, and whether the $638 billion backlog converts on the schedules embedded in Oracle’s guidance. Equity issuance at this magnitude also raises dilution concerns that no amount of backlog arithmetic fully resolves in the near term.

None of this is unfamiliar territory in periods of large-scale technological transition. Infrastructure cycles demand capital commitments that stretch well ahead of the revenue they ultimately generate. The question investors are pricing is not whether Oracle’s opportunity is real. It is whether the company can execute at the speed and discipline the opportunity demands.

The Long Arc

For senior investors evaluating Oracle’s trajectory, fiscal 2026 presents a company in purposeful transition rather than one coasting on existing advantage. Guidance for fiscal 2027 calls for total revenue of $90 billion, first-quarter cloud revenue growth of 57 to 63 percent, and non-GAAP EPS of $8.05, representing 18 percent growth after adjusting for one-time items. These targets are ambitious but not disconnected from the contracted backlog supporting them.

Oracle’s enterprise relationships, multicloud positioning, and sovereign cloud capabilities are durable differentiators in an AI infrastructure race that is far from settled. The company has built, over decades, the kind of institutional trust that is difficult to replicate and costly to displace. The financial architecture assembled for this expansion — prepayments that blunt net capital intensity, operating leverage that continues to deliver record margins, and a backlog that dwarfs current revenue — is internally coherent and strategically defensible.

The tension the market is pricing is real. Executing a hyperscale buildout while managing a $40 billion capital raise, converting an extraordinary backlog into predictable revenue, and sustaining margin discipline across a rapidly shifting cost structure is a genuinely difficult undertaking. Oracle has earned the credibility to attempt it. Whether it delivers will define the company’s standing in the AI economy for the decade ahead.